Podcast Episode 30

Episode 030 30 min

Dominik Schürmann, Co-Founder and CEO of heylogin

with Dominik Schürmann · heylogin

0:00 / 30:00

Also on Spotify, Apple Podcasts, YouTube, Podbean, Player FM

Episode 30 of The Security Podcast of Silicon Valley, published July 3, 2023. Host Jon McLachlan, co-founder of YSecurity, talks with Dominik Schürmann, co-founder and CEO of heylogin, a password manager with no master password. Dominik explains how COTECH became heylogin, how he got from Visual Basic for Kids to cryptography, and why his research on usable security, including a USENIX Security study of key fingerprints with over 1,000 people, taught him that security people trained to think like attackers make extreme choices, especially by committee. They cover German ID cards, OpenKeychain, the next PGP version, why password managers need new thinking, 110 SaaS apps and single sign-on, master passwords written on paper, the two-factor option most people never turn on, a design with no master password built on the phone's security chip and a PIN, face or fingerprint, swipe to login, browser keys that expire every day, a free private plan and a B2B business, a review that called heylogin too easy to feel secure, Tesla's race car, fear marketing, a failed first pitch, a crash course with a consultant, a conservative finance plan, nine pregnant women, Mozilla as an investor, and why fail fast is only partly right. About 30 minutes.

Episode notes

In this episode of a YSecurity.io production, we sit down with Dominik Schürmann, the Co-Founder and CEO of heylogin, who shares an incredible journey from childhood, and research, toward more usable security. How are usable security values different from "formal" or "provable" security? Why does it matter? How does that change where you spend your time and energy? And how can we integrate with human nature and our organic tendencies to build more secure software? These questions, and many more, guide Dominik to start heylogin, as he shares his outside-the-box vision for a better and more secure future.

Hosted by Jon McLachlan — co-founder of YSecurity.

The deep dive

Security that prevents adoption doesn't help

Dominik Schürmann, co-founder and CEO of heylogin, with Jon on a pivot from COTECH, Visual Basic for Kids, a study of key fingerprints with over 1,000 people, why security people think like attackers, death by committee, German ID cards, OpenKeychain, the next PGP version, why password managers need new thinking, 110 SaaS apps and single sign-on, master passwords on paper, the two-factor option most people never turn on, a password manager with no master password, a PIN bound to a security chip, swipe to login, keys that expire every day, a free plan and a B2B business, a review that called it too easy to feel secure, Tesla's race car, fear marketing, a failed first pitch, a crash course, cash flow, nine pregnant women, Mozilla and Sequoia, and why fail fast is only partly right.

Episode 30 of The Security Podcast of Silicon Valley aired on July 3, 2023, with Jon McLachlan hosting. The guest was Dominik Schürmann, co-founder and CEO of heylogin, a password manager with no master password (Podbean). Jon’s introduction started with the job and worked back.

“So Dominik is the founder and CEO of heylogin.”

Jon McLachlan, at 0:14

Listen · Jon McLachlan on episode 30, from 0:34
About 25 seconds: Jon introduces Dominik: freelance cryptography, TU Braunschweig, end-to-end encrypted communications and COTECH. The clip stops on its own. Read the full transcript.

Dominik’s LinkedIn fills in the dates. He has been CEO and co-founder of heylogin since March 2020. Before that he co-founded COTECH in May 2018 and led it until November 2021, building a Hardware Security SDK, a developer framework for authentication and encryption with security keys and smartcards on Android. From January 2014 to March 2018 he was a research associate at TU Braunschweig’s Institute for Operating Systems and Computer Networks, where he supervised more than 15 theses and published more than 15 papers, and from 2015 to 2018 he analyzed and developed cryptographic protocols as a freelancer for Germany’s Federal Office for Information Security (BSI) and for ENISA, the European Union Agency for Cybersecurity. He was a guest researcher at Aalto University twice, in 2012 and in 2017, the second time on a DAAD scholarship (LinkedIn).

For readers outside Germany, a few of the names on that list. TU Braunschweig was founded in 1745 as the Collegium Carolinum and is a member of TU9 (Wikipedia). Aalto University, in Espoo, Finland, was formed in 2010 from a merger of three universities (Wikipedia). The BSI is the German federal agency for information security, founded in 1991 in Bonn (Wikipedia), and the DAAD, the German Academic Exchange Service, is the joint organization of German universities, founded in 1925 (Wikipedia). heylogin’s about page adds that he holds a PhD in IT security (heylogin).

The tape doesn’t give a recording date. It fits the months before the July 2023 air date: Mozilla had already invested in heylogin, which the company dates to fall 2022 (heylogin).

This page is the rest of it: a section for each part of the conversation with the sources beside it, sixteen figures, nineteen clips from the tape, the full transcript, and five tools: a quiz, a scorecard for whether your security gets adopted, a calculator for the logins that live outside your single sign-on, a timeline from 1975 to 2026, and a to-do list.

A note before the sections. Dominik speaks English as a second language, and a few words on the tape are hard to hear. Where the decodes disagree, the transcript marks its best guess in brackets, and the page doesn’t quote those lines. Numbers he gives from memory, like OpenKeychain’s user count and the 110 SaaS apps, are his account, and the page says so.

We started as COTECH in 2018

Dominik’s first answer was a correction of the timeline. heylogin and COTECH are one company.

“So actually, we started as COTECH in 2018.”

Dominik Schürmann, at 1:13

“And I always say 2018 and 2019 were kind of the years of finding an actual problem that we should solve.”

Dominik Schürmann, at 1:18

Same company, he said, renamed.

Listen · Dominik Schürmann on episode 30, from 1:11
About 25 seconds: Dominik on COTECH, two years of looking for the right problem, and the rename to heylogin. The clip stops on its own. Read the full transcript.

The problem they found was password management. COTECH’s product, on Dominik’s LinkedIn, was the Hardware Security SDK for security keys and smartcards on Android, and the heylogin role starts in March 2020, the month he later calls the big pivot (LinkedIn). Jon’s verdict on the pivot was that it was clean.

“So it’s such a good pivot that they look like two distinct companies.”

Jon McLachlan, at 1:40

A timeline of Dominik's path. 2012, Aalto University: guest researcher on delay-tolerant networks, May to August. 2014 to 2018, TU Braunschweig: research associate at the Institute for Operating Systems and Computer Networks. 2015 to 2018, freelance: cryptographic protocols for the BSI and ENISA. 2016, USENIX Security: the key fingerprint study with 1,047 participants. 2017, Aalto again: a DAAD scholarship, for research on gait-based pairing of mobile devices. May 2018, COTECH, filled: co-founded, with a Hardware Security SDK for security keys and smartcards on Android. March 2020, heylogin, filled: the big pivot to password management, and then the rename, by his account. Dominik's line at 1:13: So actually, we started as COTECH in 2018.
From guest researcher to CEO: the path on Dominik's LinkedIn, and the pivot in March 2020.

Visual Basic for Kids

Jon asked what shaped him. Dominik started with a book.

“I started with Visual Basic for Kids. It was a great book.”

Dominik Schürmann, at 2:18

“I think I started at 14 years doing some programming using Visual Basic.”

Dominik Schürmann, at 2:26

“And then I was deep down into modding of a computer game and stuff like that.”

Dominik Schürmann, at 2:32

Visual Basic is Microsoft’s name for a family of programming languages, and the original version was supported from 1991 to 2008 (Wikipedia). Modding is the alteration of a video game by its players or fans (Wikipedia).

Political involvement sometimes leads to cryptography

Cryptography came later, at university, and by a route he named on the tape.

“But when I really got started with cryptography, it was during my university time.”

Dominik Schürmann, at 2:38

“And political involvement sometimes leads to cryptography, I would say.”

Dominik Schürmann, at 2:53

Listen · Dominik Schürmann on episode 30, from 2:38
About 20 seconds: Dominik on how he got into cryptography at university. The clip stops on its own. Read the full transcript.

The movement he named was the Pirate Party, a German party founded in September 2006 as part of the international movement of pirate parties, which calls itself a party of the information society (Wikipedia).

“So that got me involved into cryptography.”

Dominik Schürmann, at 3:04

From there it was cryptography through his bachelor’s and master’s, a PhD, and his work as scientific staff.

Trade-offs between usability and security

His research had one theme, and it runs through the whole episode.

“My research was always like finding trade-offs between usability and security.”

Dominik Schürmann, at 3:21

“So it’s not that easy, but it’s an interesting way to think about.”

Dominik Schürmann, at 3:28

The field has a name. Usable security is the part of computer science, human-computer interaction and cybersecurity concerned with the user interface of security systems, and it emphasizes the human side of security rather than the technical (Wikipedia). The idea is older than most of the tools it studies. Jerry Saltzer and Michael Schroeder listed “psychological acceptability” among their design principles in 1975, and the view that security and usability are at odds persisted anyway (Wikipedia).

“We did, for example, studies with over 1,000 people to find which kind of key fingerprints compare best.”

Dominik Schürmann, at 3:35

Which kind of key fingerprints compare best

Jon asked about the results. Dominik called it one of his best papers.

“This was one of the USENIX Security papers I did with a colleague of mine.”

Dominik Schürmann, at 3:57

Listen · Dominik Schürmann on episode 30, from 3:57
About 40 seconds: Dominik on comparing key fingerprints: hexadecimal, numeric and phrases. The clip stops on its own. Read the full transcript.

A key fingerprint is a short sequence of bytes, made by hashing a public key, that people compare to check they have the right key (Wikipedia). Dominik listed the three common ways to show one: hexadecimal, as in PGP, numbers, as in Signal, and words, like the backup phrases for Bitcoin wallets.

“So a lot of different ways, but they all have some upsides and downsides.”

Dominik Schürmann, at 4:22

Phrases, for instance, only work in English, he said, which is why Signal went with numbers.

The paper is “An Empirical Study of Textual Key-Fingerprint Representations,” presented at the 25th USENIX Security Symposium in Austin in August 2016 by Sergej Dechand, Dominik Schürmann, Karoline Busse, Yasemin Acar, Sascha Fahl and Matthew Smith (USENIX). It tested six textual representations with 1,047 participants. Hexadecimal, the widely used format, was more vulnerable to partial preimage attacks, a sentence-based format had “the highest attack detection rate and best usability perception,” and where a language-based format won’t work, plain numbers still beat hexadecimal (USENIX). Signal asks users to compare fingerprints out of band, and one study found 21 of 28 participants failed to compare them correctly (Wikipedia). USENIX, which runs the symposium, is a nonprofit founded in 1975 in Berkeley (Wikipedia).

Sergej Dechand, the paper’s first author, was the guest on episode 21 of this show, as co-founder and CEO of Code Intelligence.

Three boxes for ways to write a key fingerprint: hexadecimal, letters and digits, as in PGP. Numbers, digits only, as in Signal. Words, filled, like a wallet's backup phrase. Three numbers from the study: 1,047 participants, 6 textual formats tested, 2016, the 25th USENIX Security, Austin. Findings: hexadecimal, the widely used format, was more vulnerable to partial preimage attacks. A sentence-based format had the highest attack detection rate and the best usability perception. Where a language-based format won't work, plain numbers still did better than hexadecimal. Phrases only work in English, Dominik said, which is why Signal went with numbers. Dominik's line at 3:57: This was one of the USENIX Security papers I did with a colleague of mine.
Three ways to show the same key, and what the 2016 study of 1,047 people found.

Security people are trained to think like an attacker

Jon asked whether a truth had emerged from those years that most people hadn’t noticed. Dominik had thought about the question in advance.

“Security people are trained to think like an attacker, and that means they are always trying to find corner cases.”

Dominik Schürmann, at 5:26

“Like, this is a very specific attack, and this means the system is not secure.”

Dominik Schürmann, at 5:34

“So all these, like, these sorts of processes lead to extreme choices in security protocols, especially if you design security protocols by a committee.”

Dominik Schürmann, at 5:39

Listen · Dominik Schürmann on episode 30, from 5:26
About 25 seconds: Dominik on the attacker mindset and the extreme choices it leads to. The clip stops on its own. Read the full transcript.

That training shows up in threat modeling, the process of identifying and listing potential threats and prioritizing the countermeasures (Wikipedia). Dominik’s point is about what happens next. If every corner case has to be closed, the protocol gets harder to build and harder to use.

Death by committee

Jon’s reaction to the word committee was two short groans.

“Oh no, by a committee, oh no.”

Jon McLachlan, at 5:55

“That’s why some people call it death by committee.”

Dominik Schürmann, at 5:59

Design by committee is the pejorative term for a project with many designers and no unifying plan or vision (Wikipedia). His example was close to home.

“In Germany, there’s a lot of examples, like the ID cards in Germany.”

Dominik Schürmann, at 6:04

“It’s some very interesting cryptography involved, but it’s so difficult to implement that these ID cards don’t work with most of Android phones, and they don’t work with most card readers.”

Dominik Schürmann, at 6:09

Germany’s electronic ID card dates to November 2010. It carries a contactless chip, and using it to log in online takes a six-digit PIN, with a PUK to unlock it after three wrong tries (Wikipedia). Which phones and readers work with it is Dominik’s account at the time of the recording.

“So I think this is a bad way to go at security protocols.”

Dominik Schürmann, at 6:22

Four boxes joined by arrows: attacker mindset, how security is trained. Corner cases, each one fails the system. Extreme choices, worse by committee. No adoption, nobody uses it, filled. Notes: German ID cards have interesting cryptography, but are so hard to implement that most Android phones and card readers don't work with them, by Dominik's account at the recording. The card's online login uses a six-digit PIN, with a PUK after three wrong tries. Design by committee: a project with many designers and no unifying plan or vision. Dominik's line at 5:59: That's why some people call it death by committee.
How thinking like an attacker turns into security nobody can use, in Dominik's telling.

OpenKeychain, which still has 300,000 users

The second example was his own. For several years he maintained a PGP app for Android.

“It’s called OpenKeychain, which still has 300,000 users.”

Dominik Schürmann, at 6:41

He came at it with the same mindset, he said: take care of every attack scenario. That led to no adoption. The insight he took from it is to find a trade-off, because security measures that prevent adoption don’t help anyone.

Listen · Dominik Schürmann on episode 30, from 6:41
About 30 seconds: Dominik on OpenKeychain, attack scenarios and why security that prevents adoption doesn't help. The clip stops on its own. Read the full transcript.

OpenKeychain is a free and open-source Android app for encryption compatible with the OpenPGP standard, used to encrypt, decrypt, sign and verify text, email and files (Wikipedia). It began in March 2012 as a fork of Android Privacy Guard, released version 2.0 in January 2013, and was listed on the OpenPGP homepage and recommended by the Guardian Project. Since August 17, 2021 it has been in maintenance mode, because its maintainers concluded that offering its features as an app rather than as a library had been a misguided decision (Wikipedia). The 300,000 users are Dominik’s count at the time.

Two columns of three boxes joined by arrows. Left, dashed: every attack scenario, all of them covered. Hard to use, for most people. No adoption, so nobody is protected. Right: a trade-off, found on purpose. Easy to use, for most people. Adoption, so people are protected, filled. A number: 300,000 OpenKeychain users at the time, by Dominik's count. The app is a free and open-source OpenPGP app for Android. Jon's line at 7:13: You have the most secure piece of software or hardware in the world, and if nobody uses it, are we really making a difference in the world?
Security that people don't adopt protects nobody. The trade-off Dominik argues for.

If nobody uses it

Jon put the argument in one line.

“You have the most secure piece of software or hardware in the world, and if nobody uses it, are we really making a difference in the world?”

Jon McLachlan, at 7:13

Dominik agreed, and went straight to an example.

The next PGP version

“PGP is one of the worst examples, and it’s not getting better.”

Dominik Schürmann, at 7:26

“I’m a little bit following the current standardization process of the next PGP version”

Dominik Schürmann, at 7:32

His verdict on it, from a distance, was “it’s a disaster.”

Phil Zimmermann developed PGP in 1991 for signing, encrypting and decrypting text, email and files (Wikipedia). PGP itself follows the OpenPGP standard in RFC 4880, a version Wikipedia now calls historical, and doesn’t support the updated formats in RFC 9580 (Wikipedia). Jon gave the standards people their due.

“I know it’s very easy to get sucked into rabbit holes.”

Jon McLachlan, at 7:59

Dominik knows most of them personally, he said, from meetings of the people who build PGP into their products.

“We had a lot of meetings of all the PGP implementation people, like the Thunderbird guys, then the Proton Mail people.”

Dominik Schürmann, at 8:08

“Just my point of view now is they’re prioritizing the wrong stuff.”

Dominik Schürmann, at 8:23

Thunderbird is the free and open-source email client made by MZLA Technologies, a subsidiary of the Mozilla Foundation (Wikipedia), and Proton Mail is the end-to-end encrypted email service launched in 2014 in Geneva (Wikipedia).

It’s not fixable

Jon asked whether this was what nudged him toward heylogin. It was, and it’s also where he met his co-founder.

“My co-founder then also got involved in this project.”

Dominik Schürmann, at 8:46

“And we always try to innovate on this very broken standard.”

Dominik Schürmann, at 8:53

“And at one point, we just thought, okay, it’s not fixable.”

Dominik Schürmann, at 8:58

“Let’s do something else where we can design everything from the ground up.”

Dominik Schürmann, at 9:03

Listen · Dominik Schürmann on episode 30, from 8:38
About 30 seconds: Dominik on years as a PGP maintainer, a broken standard, and starting from the ground up. The clip stops on its own. Read the full transcript.

heylogin’s about page names the co-founder: Vincent Breitmoser, the CTO, who led end-to-end encryption projects for K-9 Mail and XING chat, ran keys.openpgp.org, and co-founded Stratum0, the Braunschweig hackerspace (heylogin). K-9 Mail is a free and open-source email client for Android (Wikipedia). The same page says the company grew out of their encryption research and their work as security consultants, where they saw customers keep passwords on sticky notes and in spreadsheets (heylogin).

A timeline. 1991, PGP: Phil Zimmermann develops it to sign and encrypt text, email and files. March 2012, OpenKeychain: the Android app starts as a fork of Android Privacy Guard. January 2013, version 2.0: later listed on the OpenPGP homepage and recommended by the Guardian Project. March 2020, heylogin, filled: design everything from the ground up, on a different problem. August 17, 2021, maintenance mode: OpenKeychain's maintainers decide that an app, rather than a library, was the wrong call. On the tape, dashed, the next PGP version: Dominik follows the standardization from a distance, and calls it a disaster. Dominik's line at 8:58: And at one point, we just thought, okay, it's not fixable.
PGP, OpenPGP and OpenKeychain: the standard Dominik and his co-founder decided they couldn't fix.

1Password is 16 years old now

Password managers, Dominik said, are the same kind of problem as PGP: a concept that hasn’t changed much since it started.

“Like, 1Password is 16 years old now.”

Dominik Schürmann, at 9:21

“So in the IT world, that’s a long time.”

Dominik Schürmann, at 9:26

A password manager stores usernames and passwords in a vault, and the vault’s encryption key is commonly derived from a master password (Wikipedia). 1Password, from the Canadian company AgileBits, locks its vault with a master password and by default keeps it on 1Password’s servers for a monthly fee (Wikipedia). That master password is the thing heylogin set out to remove.

110 software as a service platforms

The business case was the number of logins. Dominik had just seen a statistic, he said, that B2B companies use 110 software as a service platforms on average. He didn’t name the source, so the number is his account.

Any of them that can’t be connected to single sign-on, he said, needs a password manager.

The market is growing, he said, and yet the tools are still built for the IT department.

“the tools are still targeted to an IT audience.”

Dominik Schürmann, at 10:08

“heylogin focuses on usability, and that’s why we think differently in a lot of aspects of password management.”

Dominik Schürmann, at 10:17

Single sign-on lets a person log in once, with one ID, to several related but independent systems (Wikipedia). Software as a service began around 2000 and by 2023 was the main way software was deployed (Wikipedia). Every app that can’t be connected to single sign-on is another password someone has to keep. The calculator below counts yours, starting from his 110.

A grid of 110 small squares, 44 filled and 66 outlined. 110 SaaS platforms at the average B2B company, by the statistic Dominik cited. 66 still need their own password if 40% sit behind single sign-on, our example. Dominik's line at 10:08: the tools are still targeted to an IT audience.
110 apps by Dominik's statistic: the ones behind single sign-on, and the ones that still need a password.

Security through usability

Jon summed up the pitch in his own words.

“So heylogin is a password manager that is obsessed with the usability of the experience and trying to aim for security through usability.”

Jon McLachlan, at 10:26

“Because once you get the usability, you get the adoption, you get the ease of use, you are actually taking the burden off the shoulder of the user.”

Jon McLachlan, at 10:41

The argument comes back on this show. Taher Elgamal, who designed SSL, explained on episode 61 why the login page ended up with a password, and Riad Wahby built a company on hardware-backed keys around the idea that security versus speed is a false choice, on episode 69.

The master password on a piece of paper

Dominik’s evidence came from consulting work in 2018 and 2019, before heylogin.

“So one of our experience during our time as consultants in the early years, like 2018, 2019, was that even if companies deploy a password manager, their employees which are forced to use it just write down the master password on a piece of paper and put it on their desk.”

Dominik Schürmann, at 10:51

It’s the same problem as having no password manager, he said, moved to a different tool.

Listen · Dominik Schürmann on episode 30, from 10:51
About 35 seconds: Dominik on employees who write the master password on a piece of paper. The clip stops on its own. Read the full transcript.

Jon named what the paper means in security terms.

“and now in security terms, you have a single point of failure for all your passwords.”

Jon McLachlan, at 11:28

Three boxes joined by arrows: a password manager, rolled out by the company. Required for everyone, employees are forced to use it. The master password, on paper, on the desk, dashed. An arrow down to a filled box: one note opens every password, a single point of failure, in Jon's words, and still only one factor. Notes: seen in consulting work in 2018 and 2019, before heylogin, by Dominik's account. heylogin's about page: customers kept passwords on sticky notes and in spreadsheets. Jon's line at 11:28: and now in security terms, you have a single point of failure for all your passwords.
A password manager nobody wants to use: the master password moves to a sticky note, and one note opens everything.

Most people don’t enable it

The second problem, Dominik said, is that a master password is one factor.

“And it’s still a one-factor solution.”

Dominik Schürmann, at 11:34

“If you just use the master password, it’s just one factor, something you know, so knowledge.”

Dominik Schürmann, at 11:38

1Password and its competitors offer two-factor security as an option, he said,

“but most people don’t enable it.”

Dominik Schürmann, at 11:49

“So it’s not like there is a huge adoption of these options.”

Dominik Schürmann, at 11:52

Multi-factor authentication grants access only after a person presents two or more distinct types of evidence, which protects an account from someone who has discovered a single password (Wikipedia). An option that most people leave off doesn’t do that. The same gap shows up in security questionnaires, where “MFA enabled” and “MFA enforced” are different answers, as our post Enabled Isn’t Enforced explains.

No master password, no password at all

So heylogin dropped the master password.

“And that’s where we are trying to think differently.”

Dominik Schürmann, at 12:05

“In heylogin, there is no master password, no password at all.”

Dominik Schürmann, at 12:09

“So it’s two-factor secure by default, and even easier, because you have just the factor of possession, the security chip in your smartphone, and the second factor is the PIN or face unlock or fingerprint.”

Dominik Schürmann, at 12:19

Listen · Dominik Schürmann on episode 30, from 12:04
About 40 seconds: Dominik on a password manager with no master password, built on the phone's security chip. The clip stops on its own. Read the full transcript.

heylogin’s security architecture page describes the design. Vaults are encrypted with the security chips of the user’s devices. When the browser extension needs a vault, it asks the phone, the chip is unlocked locally with a fingerprint, face or PIN, and the device’s seed reaches the extension end-to-end encrypted. The extension derives its key pairs and discards the seed, and the cloud only stores the encrypted data and, in heylogin’s words, “has no ability to decrypt the vaults” (heylogin). The page names XSalsa20 and Poly1305 for encryption and Curve25519 for key exchange, and says vaults are also encrypted to administrators’ chips, so a company can recover access when a device is lost without admins reading personal vaults (heylogin).

A secure cryptoprocessor is a dedicated chip for cryptographic operations, packaged with physical security measures that make it tamper resistant, and it doesn’t hand decrypted data to an environment it can’t trust (Wikipedia). Our post on device-bound credentials explains why secrets that never leave that kind of hardware are hard to steal.

Left, a dashed box: master password, something you know, one factor. Notes under it: two-factor is an option, and most people don't enable it. Written down, it opens everything. Right, two boxes: the phone's security chip, something you have, filled. PIN, face or fingerprint, unlocks the chip. Notes: no master password, no password at all: everything is end-to-end encrypted using the phone. Two-factor secure by default, and easier to remember and use than a master password. Vaults are also encrypted to admins' chips, so access can be recovered when a device is lost, per heylogin's security page. Dominik's line at 12:09: In heylogin, there is no master password, no password at all.
One factor or two: the master password, and the phone's chip plus a PIN, face or fingerprint.

The PIN is bound to your security chip

Jon liked the convenience of a face or fingerprint. Dominik made the case for the PIN too: it’s more secure than a master password, he said, because of where it lives.

“the PIN is bound to your security chip in your phone, and there is a limited number of tries.”

Dominik Schürmann, at 13:04

A PIN is short. What makes it work is the hardware around it. The chip that holds the key also counts the tries, so a thief can’t keep guessing, and the key can’t be copied off the phone to guess against somewhere else. Germany’s ID card uses the same pattern, with a six-digit PIN and a PUK after three wrong tries (Wikipedia).

Swipe to login

Jon asked the practical question. How does a 100-character password get from the phone into Safari on a laptop? Dominik warned that a full answer would be a deep dive, and Jon asked for just a tease.

“There are several cryptographic keys involved here, but let me just say the most important key is the one that is generated at first-time use inside the security chip of your phone, and this key cannot be extracted from your phone.”

Dominik Schürmann, at 13:39

“from this key, we derive several other cryptographic keys for different purposes.”

Dominik Schürmann, at 13:56

Then the flow. You click a button on the site in your laptop’s browser, and a notification comes up on your phone.

“On your phone, you can just swipe to login.”

Dominik Schürmann, at 14:15

That’s what they call it, swipe to login.

“And then you’re logged in on the desktop.”

Dominik Schürmann, at 14:24

Listen · Dominik Schürmann on episode 30, from 14:01
About 30 seconds: Dominik walks through swipe to login, from a click in the browser to a swipe on the phone. The clip stops on its own. Read the full transcript.
Five numbered steps. 1: at first use, a key is generated inside the phone's security chip, and it can't be extracted from the phone. 2: other keys are derived from it for different purposes. 3: on a website in Safari on the laptop, you click a button. 4: a notification comes up on the phone, and you swipe to login. 5, filled: the laptop is logged in, and that browser gets a temporary key for the day. Dominik's line at 14:15: On your phone, you can just swipe to login.
Swipe to login, step by step, as Dominik described it.

Once per day

The swipe isn’t needed for every login.

“And when you’re doing this, you just need to do it once per day.”

Dominik Schürmann, at 14:28

After the swipe, he explained, a temporary key is generated for that browser for that day, and the browser keeps access until the phone revokes the key.

“So we don’t need to transmit all passwords from the phone to the browser.”

Dominik Schürmann, at 14:51

“It’s a little bit more complicated, but that allows us better responsiveness.”

Dominik Schürmann, at 14:56

Jon liked the expiry.

“So it sounds like we’ve got a very high security bar here, and things are expiring on a 24-hour basis, which is amazing.”

Jon McLachlan, at 15:03

“I do like that, because then now my computer is much more secure, in the sense that any sensitive data is going to have a TTL, essentially.”

Jon McLachlan, at 15:18

“You can always also open the app on your phone and revoke at any time you like, if you don’t like to follow the 24-hour protocol.”

Dominik Schürmann, at 15:32

Time to live is a mechanism that limits the lifespan of data in a computer or network. Once the set time has passed, the data is discarded or revalidated (Wikipedia). A browser key that dies at the end of the day limits what a stolen laptop is worth to a thief.

A filled bar from swipe to 24 hours: a temporary key for this browser, for this day. A dashed mark partway along: or revoked from the phone at any time. Notes: once per day, the swipe creates a temporary key for that browser for that day. Not every password travels from the phone to the browser, which keeps it responsive. Open the app on the phone and revoke it whenever you like. Jon's read: sensitive data on the laptop gets a time to live. Dominik's line at 14:28: And when you're doing this, you just need to do it once per day.
A key for one browser for one day, and a revoke button on the phone.

Bottom-up

Jon asked about go-to-market. B2C or B2B, top-down or bottom-up?

“Bottom-up.”

Dominik Schürmann, at 15:50

“So, I mean, 1Password, LastPass, Dashlane, they all have a private user plan.”

Dominik Schürmann, at 15:50

“We currently don’t have a private user plan, because our private user plan is free.”

Dominik Schürmann, at 15:56

“I don’t think there is a lot of market left for B2C in this space.”

Dominik Schürmann, at 16:05

Jon called it saturated. The people who pay for a password manager already have one, he said, and they won’t easily switch to another paid one.

“They don’t easily switch to another paid password manager, but they may switch to a free plan and a better password manager.”

Dominik Schürmann, at 16:24

“So it’s kind of like the typical product-led growth strategy, where you have a freemium plan and then do an upsell.”

Dominik Schürmann, at 16:36

Listen · Dominik Schürmann on episode 30, from 16:12
About 30 seconds: Dominik on a free private plan and a paid B2B plan. The clip stops on its own. Read the full transcript.

The money is in the B2B plan. Freemium is the pricing strategy where a basic product is free and money is charged for features that expand it, used in software since the 1980s (Wikipedia). The competitors he named are older. LastPass was founded in 2008, bought by LogMeIn, now GoTo, for $110 million in 2015, and breached in late 2022 (Wikipedia). Dashlane was founded in Paris on July 6, 2009 (Wikipedia).

Three boxes joined by arrows: a free private plan, anyone can use it. People switch, to a free, better manager. The B2B plan, the business model, filled. Notes: 1Password, LastPass and Dashlane all sell private plans. People who pay already have a password manager and don't easily switch to another paid one. Little market left for B2C, in Dominik's view, and Jon called it saturated. Product-led growth: a freemium plan, then an upsell. Dominik's line at 16:36: So it's kind of like the typical product-led growth strategy, where you have a freemium plan and then do an upsell.
Bottom-up: a free plan for people, and a paid plan for their companies.

So easy to use that it doesn’t feel so secure

Jon asked for the best day so far. Dominik picked a week.

“Yeah, so when we got the first reviews of our B2B features, it was quite a great week.”

Dominik Schürmann, at 17:04

“So that kind of proved our strategy and our usability.”

Dominik Schürmann, at 17:25

They printed one review and hung it on the wall. The reviewer loved the product.

“He said it’s like the best user experience ever.”

Dominik Schürmann, at 17:47

And then the last line of the review.

“it’s so easy to use that it doesn’t feel so secure.”

Dominik Schürmann, at 17:53

“So it’s quite funny, because it’s a compliment at the same time.”

Dominik Schürmann, at 18:00

Listen · Dominik Schürmann on episode 30, from 17:36
About 25 seconds: Dominik on the review they printed and hung on the wall. The clip stops on its own. Read the full transcript.

Jon’s read was that people have been trained to expect security to hurt.

“It’s going to be an impediment to the flow of your day, a roadblock, an obstacle, or something.”

Jon McLachlan, at 18:19

Tesla’s first car was basically a race car

Jon’s comparison was a car.

“You know, I imagine this is why Tesla, the very first car that they built, was basically a race car.”

Jon McLachlan, at 19:05

“You know, they’re trying to break the stereotype: electric cars are slow, less of a car, whatever.”

Jon McLachlan, at 19:12

Listen · Jon McLachlan on episode 30, from 19:05
About 25 seconds: Jon on Tesla's race car, and breaking the stereotype that security has to be slow. The clip stops on its own. Read the full transcript.

Break the stereotype with a very fast car, he said, then build the sedan, and do the same for security in password management. The first-generation Tesla Roadster was a battery electric sports car based on the Lotus Elise chassis, produced from 2008 to 2012, and it went from 0 to 60 miles per hour in 3.7 or 3.9 seconds (Wikipedia). The four-door Model S, the company’s second vehicle, went into series manufacture in June 2012 (Wikipedia).

Two boxes joined by an arrow: the stereotype, security gets in the way, dashed. Break it, make it fast and easy first, filled. A timeline: 2008 to 2012, Tesla Roadster: a battery electric sports car on the Lotus Elise chassis, 0 to 60 mph in 3.7 or 3.9 seconds. June 2012, Model S, filled: series manufacture of the four-door, the company's second vehicle. Dominik's lines from the review: at 17:47, He said it's like the best user experience ever. At 17:53: it's so easy to use that it doesn't feel so secure.
The review on the wall, and Jon's race car: break the stereotype first.

Fear marketing

Dominik added a second side to the same idea.

“There is even the side of marketing.”

Dominik Schürmann, at 20:00

“So IT security products are often sold with what I call fear marketing.”

Dominik Schürmann, at 20:02

“People should fear data leaks.”

Dominik Schürmann, at 20:09

“Like, the most costly ad Dashlane ever ran was a Super Bowl ad.”

Dominik Schürmann, at 20:18

Listen · Dominik Schürmann on episode 30, from 20:01
About 20 seconds: Dominik on fear marketing in security, and a Super Bowl ad. The clip stops on its own. Read the full transcript.

In the ad as he remembers it, a man dies, arrives in heaven, and is asked for his password. The ad and its cost are his account. heylogin’s answer is a different pitch.

“So that’s why we try to go the way of showing how heylogin provides more productivity for your employees.”

Dominik Schürmann, at 20:52

Jon agreed.

“I’ve never really seen it pay off in any long-term success.”

Jon McLachlan, at 21:05

Two columns. Left, dashed: fear marketing, how security is often sold. People should fear data leaks. And what leaving employees take with them. The most costly ad Dashlane ever ran was a Super Bowl ad, by Dominik's account. Right, filled: productivity, what heylogin pitches instead. Show how heylogin gives employees more productivity. Jon has never seen fear pay off in long-term success. Dominik's line at 20:02: So IT security products are often sold with what I call fear marketing.
Two ways to sell security: fear of what goes wrong, or time people get back.

The worst day

Jon asked for the worst day. Dominik went back to the start of heylogin.

“So, you know, when we started developing heylogin in March 2020, we were able to start with a grant.”

Dominik Schürmann, at 21:17

“We got a big grant from the German government that was pretty great to do the development.”

Dominik Schürmann, at 21:25

As the grant ran down, they went looking for a first venture capital round.

“And we were pretty confident that it shouldn’t be a problem.”

Dominik Schürmann, at 21:40

“We had a pitch at a very, very local VC.”

Dominik Schürmann, at 21:48

Listen · Dominik Schürmann on episode 30, from 21:48
About 30 seconds: Dominik on a failed first pitch, and a week of being down. The clip stops on its own. Read the full transcript.

It was the fund for their own state, Lower Saxony, and partly governmental, he said, and the money didn’t come. Looking back, he called the confidence naive.

“So that was a really bad experience, and it kind of grounded us.”

Dominik Schürmann, at 22:10

“So, yeah, I was down for a week at least.”

Dominik Schürmann, at 22:14

Lower Saxony is a coastal state in northwestern Germany (Wikipedia), and Braunschweig, where heylogin has its headquarters, is one of its cities (Wikipedia, LinkedIn). Jon thanked him for telling it. Those are vulnerable moments, he said,

“not just for companies, but for the leaders in those companies too.”

Jon McLachlan, at 22:38

A crash course

Taking in the feedback and starting a new round took a long time, Dominik said.

“There was a lot of stuff that we needed to learn before.”

Dominik Schürmann, at 22:53

“So we took an external consultant for this specific round, and then it took some months until we started talking to VCs again, and then we were better prepared.”

Dominik Schürmann, at 22:57

Jon called it a crash course in an MBA.

“we had like three meetings per week with this consultant.”

Dominik Schürmann, at 23:20

“So, yeah, it was kind of a crash course.”

Dominik Schürmann, at 23:23

A timeline. March 2020, a grant: development of heylogin starts on a big grant from the German government. As it ran down, dashed, the first pitch: a very local fund for Lower Saxony, partly governmental, says no. He's down for a week. After, a consultant: three meetings a week for one round, then back to investors, better prepared. Fall 2022, filled, pre-seed: local business angels and Mozilla Ventures. heylogin's press page puts the round at 900,000 euros. November 2, 2022, filled, Mozilla Ventures: announced with 35 million dollars, naming HeyLogin among its first three investments. On the tape, the plan: very conservative, cash flow positive the next year, the same development team, money into sales and marketing. Dominik's lines: at 21:48, We had a pitch at a very, very local VC. At 26:06: I think we have a pretty interesting combination of business angels and one bigger brand.
From a government grant to Mozilla: heylogin's funding story as Dominik told it, with the dates from heylogin and Wikipedia.

A very conservative finance plan

Jon asked what success looks like. Dominik answered with a plan.

“Yeah, so currently we have a very conservative finance plan.”

Dominik Schürmann, at 23:47

The plan was to be cash flow positive the following year without a lot of growth, and he was hesitant to raise another round in the venture market of the time. Jon pointed to the responsibility that comes with it.

“I mean, being a leader of a company, there’s people that depend on you for their livelihood”

Jon McLachlan, at 24:46

Nine pregnant women don’t make a baby in a month

The plan held the engineering team steady.

“The development team stays the same after a current round.”

Dominik Schürmann, at 25:05

One more developer, he said, doesn’t make product development faster.

“So what we do is put our money into sales and marketing and keep the current headcount in development.”

Dominik Schürmann, at 25:21

Listen · Dominik Schürmann on episode 30, from 25:05
About 25 seconds: Dominik on keeping the development team the same size and spending on sales and marketing. The clip stops on its own. Read the full transcript.

Jon reached for the old line about software schedules.

“that nine pregnant women don’t make a baby in a month.”

Jon McLachlan, at 25:36

The phrase comes from Brooks’s law, Fred Brooks’s observation in his 1975 book The Mythical Man-Month that “adding manpower to a late software project makes it later” (Wikipedia). Jon added that he wished there were a gender-neutral version.

It’s not like Sequoia

Jon asked whether the investors pushed for more hiring and spending. Dominik described a mix.

“I think we have a pretty interesting combination of business angels and one bigger brand.”

Dominik Schürmann, at 26:06

The angels are very local, he said, and the brand is Mozilla. The angels have a German perspective and are happy with cash flow positive.

“They don’t necessarily plan for growth.”

Dominik Schürmann, at 26:38

Listen · Dominik Schürmann on episode 30, from 26:28
About 35 seconds: Dominik on local business angels, Mozilla, and why there wasn't much pressure. The clip stops on its own. Read the full transcript.

Mozilla, he said, isn’t like Sequoia or another venture firm, and there wasn’t a lot of pressure. Mozilla announced Mozilla Ventures on November 2, 2022, at Web Summit in Lisbon, with a starting fund of $35 million, and named HeyLogin among its first three investments (Wikipedia). heylogin dates the investment to fall 2022, in a pre-seed round alongside German business angels (heylogin), and its press page puts the round at €900,000 (heylogin). Sequoia Capital is a venture firm in Menlo Park, California, that invests from seed to growth stage (Wikipedia).

“So this is a good thing, and I’m happy for you”

Jon McLachlan, at 27:24

Fail fast is only partly right

Jon’s last question was the advice Dominik wishes he had heard younger. He took on a Silicon Valley favorite.

“there is this idea of fail fast and try a lot of things and go to market early.”

Dominik Schürmann, at 28:02

“I think it’s only partly right.”

Dominik Schürmann, at 28:11

His example was Figma, which took years of trying different things before it settled on a product, and then more time to build a basic version.

“So after trying a lot of stuff, you should at one point stick to something.”

Dominik Schürmann, at 28:28

“And then it takes some time until you get the first user base.”

Dominik Schürmann, at 28:33

Listen · Dominik Schürmann on episode 30, from 28:02
About 35 seconds: Dominik on why fail fast is only partly right, and what Figma shows. The clip stops on its own. Read the full transcript.

Fail fast is the management idea that people should use trial and error to judge a product or strategy quickly and move on, and it’s part of the culture of Silicon Valley (Wikipedia). Figma, Inc. was founded in 2012 by Dylan Field and Evan Wallace, and its first public release came on September 27, 2016 (Wikipedia). Jon added the obvious footnote, maybe a pivot or two, and Dominik agreed. If you look closely, he said, heylogin has had more than one, and the big one was in March 2020.

Four boxes joined by arrows: try a lot, fail fast, early on. Stick to one, at some point. A basic version, takes time. First users, take more time, filled. A timeline: 2012, Figma, Inc.: founded by Dylan Field and Evan Wallace. September 27, 2016, filled, first public release: about four years later. March 2020, filled, heylogin's big pivot: one of more than one, Dominik said. Dominik's line at 28:11: I think it's only partly right.
Try many things, then stick to one: Dominik's advice, with Figma's dates.

In the market for a new password manager

Jon closed with his own situation. He was shopping for a password manager, because his 1Password had dropped the peer-to-peer sync between his own devices in its latest version.

“And I just don’t want my stuff in their cloud with only a password protecting it.”

Jon McLachlan, at 29:22

“So, long story short, I’m in the market, and I’m definitely going to give heylogin a try.”

Jon McLachlan, at 29:27

Listen · Jon McLachlan on episode 30, from 29:00
About 30 seconds: Jon on leaving 1Password and giving heylogin a try. The clip stops on its own. Read the full transcript.

Wikipedia records the change: 1Password 8 removed the option to store vaults locally, which drew criticism, and syncing over the local Wi-Fi was only in earlier versions (Wikipedia). Jon invited listeners to try it too.

“To be a small part of that journey, to be a small part of history.”

Jon McLachlan, at 29:48

After the tape

heylogin’s LinkedIn page lists it as founded in 2020, with headquarters in Braunschweig and 11 to 50 employees (LinkedIn). Its website now says logins can be confirmed with the phone app’s swipe to login, a FIDO2 security key, Touch ID, Windows Hello or a smartwatch, lists passkeys and two-factor codes among its features, and says heylogin GmbH is ISO 27001 certified (heylogin). A passkey is a key pair created by the user’s own authenticator, built on standards from the FIDO Alliance and the W3C’s WebAuthn (Wikipedia), and the FIDO Alliance launched in February 2013 to reduce the world’s reliance on passwords (Wikipedia).

The press noticed before the episode aired. heise online tested heylogin in January 2023 (heise online), and Golem.de wrote about the password manager that needs no password itself in June 2023 (Golem.de).

Dominik still makes the argument from the show. In September 2026 he wrote that a password manager isn’t automatically more secure because it can fend off more attacks, and that it’s only secure once as many people as possible actually use it.

Ein Security-Feature, das niemand nutzt, ist nur eine weitere Zeile im Whitepaper.

Dominik Schürmann · September 4, 2026
Dominik on usable security, in German. The line translates as: a security feature nobody uses is just another line in the whitepaper. View on LinkedIn.

It’s the OpenKeychain lesson, three years after the episode: security that people don’t use doesn’t help.

A timeline. January 2023, heise online: tests the password manager without a master password. June 2023, Golem.de: writes about the password manager that needs no password itself. July 3, 2023, filled, episode 30 airs: on Podbean, Apple Podcasts and Spotify. September 4, 2026, filled, Dominik's post: a security feature nobody uses is just another line in the whitepaper, in translation. heylogin today: logins confirmed by a swipe on the phone, a FIDO2 security key, Touch ID, Windows Hello or a smartwatch. Passkeys and two-factor codes among its features. ISO 27001 certified, founded in 2020, headquartered in Braunschweig, 11 to 50 employees.
After the tape: heylogin today, the press before the episode, and Dominik's post on usable security.

Did you catch it? A seven-question quiz

Seven questions from the episode. Each answer links the moment in the transcript.

Quiz · seven questions

Did you catch it?

Pick an answer for each. The explanation shows once you choose.

  1. What did heylogin start as in 2018?

    1. COTECH
    2. OpenKeychain
    3. Stratum0

    a. "So actually, we started as COTECH in 2018." Transcript, 1:13.

  2. How many people took part in Dominik's key fingerprint studies, as he put it on the tape?

    1. Over 100
    2. Over 1,000
    3. Over 10,000

    b. "We did, for example, studies with over 1,000 people to find which kind of key fingerprints compare best." Transcript, 3:35.

  3. What do some people call security protocols designed by a committee?

    1. Security theater
    2. Defense in depth
    3. Death by committee

    c. "That's why some people call it death by committee." Transcript, 5:59.

  4. What takes the place of the master password in heylogin?

    1. A longer passphrase
    2. The phone's security chip, plus a PIN, face or fingerprint
    3. A code sent by text message

    b. "So it's two-factor secure by default, and even easier, because you have just the factor of possession, the security chip in your smartphone, and the second factor is the PIN or face unlock or fingerprint." Transcript, 12:19.

  5. How often do you swipe to log a browser in?

    1. Once per day
    2. For every login
    3. Once per week

    a. "And when you're doing this, you just need to do it once per day." Transcript, 14:28.

  6. What did the review on heylogin's wall say at the end?

    1. That setup took too long
    2. That it was too expensive for a small team
    3. That it was so easy to use it didn't feel secure

    c. "it's so easy to use that it doesn't feel so secure." Transcript, 17:53.

  7. Which bigger brand invested alongside heylogin's local business angels?

    1. Sequoia
    2. Mozilla
    3. Proton

    b. "I think we have a pretty interesting combination of business angels and one bigger brand." The brand is Mozilla. Transcript, 26:06.

Does your security get adopted?

Dominik’s test for a security measure is whether people use it. Eight questions on whether yours pass.

Scorecard · eight questions

Does your security get adopted?

Answer for how your company works today. If two answers fit, pick the lower one.

  1. Do the security tools you pick work for people outside IT, not just for the IT team?

    • Only IT can use them
    • Some of them
    • Yes, we pick tools everyone can use

    "the tools are still targeted to an IT audience." See 110 software as a service platforms.

  2. If you've rolled out a password manager, do people use it without writing the master password down?

    • We don't have one, or we don't know
    • We think so, but we don't check
    • Yes, and we check

    "their employees which are forced to use it just write down the master password on a piece of paper and put it on their desk." See The master password on a piece of paper.

  3. Is two-factor enforced, not just offered, on the accounts that matter?

    • It's optional
    • Enforced for admins
    • Enforced for everyone

    "but most people don't enable it." See Most people don't enable it.

  4. Do your logins rely on something people have, like a phone's security chip or a security key, and not only on something they know?

    • Passwords only
    • For some apps
    • Yes, by default

    "the PIN is bound to your security chip in your phone, and there is a limited number of tries." See The PIN is bound to your security chip.

  5. Does access on a laptop expire on its own, and can people revoke it from their phone?

    • Sessions last until someone notices
    • Some expire
    • Yes, short-lived and revocable

    "You can always also open the app on your phone and revoke at any time you like, if you don't like to follow the 24-hour protocol." See Once per day.

  6. Before a security control ships, do you try it with the people who will have to use it?

    • No
    • After launch, if people complain
    • Yes, before it ships

    "My research was always like finding trade-offs between usability and security." See Trade-offs between usability and security.

  7. Do you track how many people use each security tool, and not just whether it's deployed?

    • No
    • For a few tools
    • Yes, every tool has an adoption number

    "You have the most secure piece of software or hardware in the world, and if nobody uses it, are we really making a difference in the world?" See If nobody uses it.

  8. When you ask people to adopt a security tool, do you lead with the time it saves them rather than with fear?

    • Mostly fear
    • A mix
    • Mostly the time it saves

    "So IT security products are often sold with what I call fear marketing." See Fear marketing.

People work around it

Your controls look good on paper, and people are probably working around them. This month, list the apps outside single sign-on, enforce two-factor on admin accounts, and ask ten people how they really log in.

Adopted in places

The basics are in place for some people. Next, enforce two-factor for everyone, move the important logins to something people have, like a phone's chip or a security key, and give laptop sessions an expiry.

Security people use

Your security is built to be used. Keep measuring adoption for every tool, test new controls with the people who use them before they ship, and keep pitching security as time saved.

How many logins live outside your single sign-on?

Dominik’s statistic was 110 SaaS platforms at the average B2B company. This calculator starts there and counts the logins that single sign-on doesn’t cover. Every default besides the 110 is our example, not a benchmark.

Calculator

How many logins live outside your single sign-on?

Four numbers: the SaaS apps your company uses, the share behind single sign-on, the share a typical person uses, and how many people you have.

–apps outside single sign-on
–of your apps
–separate passwords per person
–passwords across the company

With 110 apps and 40% behind single sign-on, 66 apps still need their own logins. If a typical person uses 30% of them, that's about 20 passwords each, or 990 across 50 people.

The 110 is the statistic Dominik cited on the tape for B2B companies on average. The 40%, the 30% and the 50 people are our examples, not benchmarks. Put in your own.

From psychological acceptability to swipe to login

Everything on this page sits on one timeline, from a 1975 design principle to Dominik’s post in 2026. Filter it by thread.

Timeline · 1975 to 2026

From psychological acceptability to swipe to login

Filter by thread. Dates are from the public record except where marked as Dominik's account or part of this show.

  1. Psychological acceptability

    Jerry Saltzer and Michael Schroeder make usability a security design principle (Wikipedia, on this show).

  2. The Mythical Man-Month

    Fred Brooks coins the law about adding people to a late software project (Wikipedia, on this show).

  3. PGP

    Phil Zimmermann develops Pretty Good Privacy (Wikipedia, on this show).

  4. The BSI

    Germany's Federal Office for Information Security is founded in Bonn (Wikipedia, on this show).

  5. LastPass

    The password manager is founded (Wikipedia, on this show).

  6. Tesla Roadster

    Production of the electric sports car begins and runs to 2012 (Wikipedia, on this show).

  7. Dashlane

    The password manager is founded in Paris (Wikipedia, on this show).

  8. The electronic ID card

    Germany's ID card gets a chip and a six-digit PIN for logging in online (Wikipedia, on this show).

  9. OpenKeychain

    The Android app starts as a fork of Android Privacy Guard (Wikipedia, on this show).

  10. Aalto University

    Dominik's first stay as a guest researcher, on delay-tolerant networks (LinkedIn, on this show).

  11. Model S

    Series manufacture of Tesla's second vehicle begins (Wikipedia, on this show).

  12. Figma, Inc.

    Dylan Field and Evan Wallace found the company (Wikipedia, on this show).

  13. OpenKeychain 2.0

    The app releases version 2.0 (Wikipedia, on this show).

  14. The FIDO Alliance

    Launched to reduce the world's reliance on passwords (Wikipedia, on this show).

  15. TU Braunschweig

    Dominik becomes a research associate at the Institute for Operating Systems and Computer Networks (LinkedIn, on this show).

  16. LastPass sold

    LogMeIn, now GoTo, buys it for $110 million (Wikipedia, on this show).

  17. Freelance cryptography

    Dominik starts analyzing and developing cryptographic protocols for the BSI and ENISA, until 2018 (LinkedIn, on this show).

  18. Key fingerprints at USENIX Security

    The study of 1,047 participants is presented in Austin (USENIX, on this show).

  19. Figma's first public release

    Four years after the company was founded (Wikipedia, on this show).

  20. Aalto again

    A DAAD scholarship, for research on gait-based pairing of mobile devices (LinkedIn, on this show).

  21. COTECH

    Dominik co-founds the company that becomes heylogin (LinkedIn, on this show).

  22. The big pivot

    Development of heylogin starts on a government grant, by Dominik's account (LinkedIn, on this show).

  23. Maintenance mode

    OpenKeychain goes into maintenance mode (Wikipedia, on this show).

  24. Mozilla invests

    Mozilla Ventures joins German business angels in heylogin's pre-seed round (heylogin, on this show).

  25. Mozilla Ventures

    Announced at Web Summit in Lisbon with $35 million, naming HeyLogin among its first three investments (Wikipedia, on this show).

  26. The LastPass breach

    Attackers reach LastPass user data and vaults (Wikipedia, on this show).

  27. heise online's test

    A review of the password manager without a master password (heise online, on this show).

  28. Golem.de

    A story on the password manager that needs no password itself (Golem.de, on this show).

  29. Episode 30 airs

    On Podbean, Apple Podcasts and Spotify (Podbean, on this show).

  30. A security feature nobody uses

    Dominik's post on usable security (LinkedIn, on this show).

What to do this week

Five things, in the order the episode suggests.

  1. Find out whether your security gets used. Take the scorecard with whoever owns IT and whoever owns security, and compare answers.
  2. Count the logins outside single sign-on. Run the calculator with your real numbers, then connect what you can and give everyone a password manager for the rest.
  3. Enforce two-factor, don’t just offer it. An option most people never turn on doesn’t protect them. Our post Enabled Isn’t Enforced has the checks.
  4. Move the important logins into hardware. Passkeys and keys bound to a device’s security chip can’t be written on a sticky note. Our post on device-bound credentials has a rollout plan.
  5. Get help with the parts you can’t staff. Our Product Security team helps build security your users will adopt, our Penetration Testing team tests what an attacker would try, a vCISO can own the program, our Sales Support team answers the security questionnaires, and you can book 15 minutes with us.

Key takeaways

  1. Adoption is the test

    If security measures prevent adoption, they don't help. Dominik learned it maintaining OpenKeychain, where covering every attack scenario led to no adoption.

  2. Attacker thinking has a cost

    Security people are trained to think like attackers and close every corner case, which leads to extreme choices in protocols, especially when a committee designs them.

  3. Death by committee

    German ID cards have interesting cryptography but are so hard to implement that, by Dominik's account, most Android phones and card readers don't work with them.

  4. Fingerprints are a usability problem

    In a USENIX Security study of 1,047 people, a sentence-based fingerprint did best, and plain numbers beat hexadecimal. Phrases only work in English, which is why Signal went with numbers.

  5. Some standards can't be fixed

    After years of trying to innovate on PGP, Dominik and his co-founder decided it wasn't fixable and started over on a different problem, from the ground up.

  6. Every app outside single sign-on needs a password

    Dominik cited 110 SaaS platforms at the average B2B company. Any app that can't be connected to single sign-on needs a password manager, and most of those tools are built for IT.

  7. Forced tools get worked around

    In consulting work in 2018 and 2019, employees forced to use a password manager wrote the master password on paper and put it on the desk, which Jon called a single point of failure.

  8. Optional two-factor isn't two-factor

    Password managers offer two-factor as an option, and most people don't enable it, so for most users the vault still rests on one factor, something they know.

  9. Two factors without a password

    heylogin has no master password. The factors are possession, the security chip in the phone, and a PIN, face or fingerprint, which are easier to remember and use.

  10. A PIN can beat a long password

    The PIN is bound to the phone's security chip and allows a limited number of tries, which makes it more secure than a master password, in Dominik's view.

  11. Short-lived access by default

    One swipe a day gives a browser a temporary key that the phone can revoke at any time. Jon's read: sensitive data on the laptop gets a time to live.

  12. Free for people, paid for companies

    heylogin's private plan is free and the B2B plan is the business model, a product-led growth strategy built on people switching from paid tools.

  13. Sell productivity, not fear

    Security is often sold with fear marketing. heylogin shows the productivity it gives employees instead, and Jon has never seen fear pay off in long-term success.

  14. A failed pitch is a lesson

    heylogin's first pitch to a local, partly governmental fund failed. An external consultant, three meetings a week and months of preparation came before the next round.

  15. Growth doesn't come from headcount

    The plan kept the development team the same size and put the money into sales and marketing, because one more developer doesn't make product development faster.

  16. Fail fast is only partly right

    Try things early, but at some point stick to one. Figma took years before its product and more time for a basic version, and a first user base takes longer still.

Chapters

Tap a chapter to play from there.

Pull quotes

“So Dominik is the founder and CEO of heylogin.”

Jon McLachlan ·

“So actually, we started as COTECH in 2018.”

Dominik Schürmann ·

“And I always say 2018 and 2019 were kind of the years of finding an actual problem that we should solve.”

Dominik Schürmann ·

“So it's such a good pivot that they look like two distinct companies.”

Jon McLachlan ·

“I started with Visual Basic for Kids. It was a great book.”

Dominik Schürmann ·

“But when I really got started with cryptography, it was during my university time.”

Dominik Schürmann ·

“And political involvement sometimes leads to cryptography, I would say.”

Dominik Schürmann ·

“My research was always like finding trade-offs between usability and security.”

Dominik Schürmann ·

“We did, for example, studies with over 1,000 people to find which kind of key fingerprints compare best.”

Dominik Schürmann ·

“This was one of the USENIX Security papers I did with a colleague of mine.”

Dominik Schürmann ·

“So a lot of different ways, but they all have some upsides and downsides.”

Dominik Schürmann ·

“Security people are trained to think like an attacker, and that means they are always trying to find corner cases.”

Dominik Schürmann ·

“So all these, like, these sorts of processes lead to extreme choices in security protocols, especially if you design security protocols by a committee.”

Dominik Schürmann ·

“Oh no, by a committee, oh no.”

Jon McLachlan ·

“That's why some people call it death by committee.”

Dominik Schürmann ·

“It's some very interesting cryptography involved, but it's so difficult to implement that these ID cards don't work with most of Android phones, and they don't work with most card readers.”

Dominik Schürmann ·

“It's called OpenKeychain, which still has 300,000 users.”

Dominik Schürmann ·

“You have the most secure piece of software or hardware in the world, and if nobody uses it, are we really making a difference in the world?”

Jon McLachlan ·

“PGP is one of the worst examples, and it's not getting better.”

Dominik Schürmann ·

“I know it's very easy to get sucked into rabbit holes.”

Jon McLachlan ·

“Just my point of view now is they're prioritizing the wrong stuff.”

Dominik Schürmann ·

“And at one point, we just thought, okay, it's not fixable.”

Dominik Schürmann ·

“Let's do something else where we can design everything from the ground up.”

Dominik Schürmann ·

“Like, 1Password is 16 years old now.”

Dominik Schürmann ·

“the tools are still targeted to an IT audience.”

Dominik Schürmann ·

“So heylogin is a password manager that is obsessed with the usability of the experience and trying to aim for security through usability.”

Jon McLachlan ·

“So one of our experience during our time as consultants in the early years, like 2018, 2019, was that even if companies deploy a password manager, their employees which are forced to use it just write down the master password on a piece of paper and put it on their desk.”

Dominik Schürmann ·

“and now in security terms, you have a single point of failure for all your passwords.”

Jon McLachlan ·

“And it's still a one-factor solution.”

Dominik Schürmann ·

“but most people don't enable it.”

Dominik Schürmann ·

“In heylogin, there is no master password, no password at all.”

Dominik Schürmann ·

“So it's two-factor secure by default, and even easier, because you have just the factor of possession, the security chip in your smartphone, and the second factor is the PIN or face unlock or fingerprint.”

Dominik Schürmann ·

“the PIN is bound to your security chip in your phone, and there is a limited number of tries.”

Dominik Schürmann ·

“There are several cryptographic keys involved here, but let me just say the most important key is the one that is generated at first-time use inside the security chip of your phone, and this key cannot be extracted from your phone.”

Dominik Schürmann ·

“On your phone, you can just swipe to login.”

Dominik Schürmann ·

“And when you're doing this, you just need to do it once per day.”

Dominik Schürmann ·

“I do like that, because then now my computer is much more secure, in the sense that any sensitive data is going to have a TTL, essentially.”

Jon McLachlan ·

“You can always also open the app on your phone and revoke at any time you like, if you don't like to follow the 24-hour protocol.”

Dominik Schürmann ·

“So it's kind of like the typical product-led growth strategy, where you have a freemium plan and then do an upsell.”

Dominik Schürmann ·

“He said it's like the best user experience ever.”

Dominik Schürmann ·

“it's so easy to use that it doesn't feel so secure.”

Dominik Schürmann ·

“So it's quite funny, because it's a compliment at the same time.”

Dominik Schürmann ·

“You know, I imagine this is why Tesla, the very first car that they built, was basically a race car.”

Jon McLachlan ·

“So IT security products are often sold with what I call fear marketing.”

Dominik Schürmann ·

“So that's why we try to go the way of showing how heylogin provides more productivity for your employees.”

Dominik Schürmann ·

“I've never really seen it pay off in any long-term success.”

Jon McLachlan ·

“We had a pitch at a very, very local VC.”

Dominik Schürmann ·

“So, yeah, I was down for a week at least.”

Dominik Schürmann ·

“So we took an external consultant for this specific round, and then it took some months until we started talking to VCs again, and then we were better prepared.”

Dominik Schürmann ·

“Yeah, so currently we have a very conservative finance plan.”

Dominik Schürmann ·

“So what we do is put our money into sales and marketing and keep the current headcount in development.”

Dominik Schürmann ·

“that nine pregnant women don't make a baby in a month.”

Jon McLachlan ·

“I think we have a pretty interesting combination of business angels and one bigger brand.”

Dominik Schürmann ·

“I think it's only partly right.”

Dominik Schürmann ·

“So after trying a lot of stuff, you should at one point stick to something.”

Dominik Schürmann ·

“And I just don't want my stuff in their cloud with only a password protecting it.”

Jon McLachlan ·

“So, long story short, I'm in the market, and I'm definitely going to give heylogin a try.”

Jon McLachlan ·

About the guest

Dominik Schürmann, Co-Founder and CEO, heylogin

Dominik Schürmann is co-founder and CEO of heylogin, a password manager with no master password, based in Braunschweig, Germany. He holds a PhD in IT security, was a research associate at TU Braunschweig's Institute for Operating Systems and Computer Networks from 2014 to 2018, analyzed and developed cryptographic protocols as a freelancer for the BSI and ENISA, and was a guest researcher at Aalto University. He maintained the Android encryption app OpenKeychain, co-wrote a 2016 USENIX Security study of key fingerprint formats, and co-founded COTECH in 2018, the company that became heylogin.

About the host

Jon McLachlan is co-founder of YSecurity, the on-demand cybersecurity team for startups, and of Cyberbase.ai. The Security Podcast of Silicon Valley is hosted by Jon McLachlan and Sasha Sinkevich. Jon McLachlan on LinkedIn ↗

Context and sources

  • Dominik Schürmann's LinkedIn: CEO and co-founder of heylogin since March 2020, CEO and co-founder of COTECH from May 2018 to November 2021 with a Hardware Security SDK for security keys and smartcards on Android, research associate at TU Braunschweig's Institute for Operating Systems and Computer Networks from January 2014 to March 2018, freelance cryptography engineer for the BSI and ENISA from 2015 to 2018, and guest researcher at Aalto University in 2012 and, on a DAAD scholarship, in 2017. Source ↗
  • Episode 30 on Podbean, July 3, 2023, with show notes on Dominik's journey from research toward more usable security, how usable security differs from formal or provable security, and how to build more secure software that works with human nature. Source ↗
  • heylogin's about page: founders Dr. Dominik Schürmann, CEO, and Vincent Breitmoser, CTO, a company that grew out of their encryption research and security consulting, and Mozilla Ventures' investment in the pre-seed round in fall 2022 alongside German business angels. Source ↗
  • heylogin's security architecture: vaults encrypted with the security chips of users' devices, a seed passed end-to-end encrypted to the browser extension, cloud storage that can't decrypt the vaults, XSalsa20, Poly1305 and Curve25519, recovery through administrators' chips, and an ISO 27001:2022 information security management system. Source ↗
  • heylogin's website: logins confirmed with a swipe on the phone, a FIDO2 security key, Touch ID, Windows Hello or a smartwatch, passkeys and TOTP codes among its features, and ISO 27001 certification. Source ↗
  • heylogin's press page: a €900,000 pre-seed round with Mozilla Ventures and business angels, and coverage including heise online in January 2023 and Golem.de in June 2023. Source ↗
  • heylogin on LinkedIn: founded in 2020, headquartered in Braunschweig, 11 to 50 employees. Source ↗
  • An Empirical Study of Textual Key-Fingerprint Representations, USENIX Security 2016, by Sergej Dechand, Dominik Schürmann, Karoline Busse, Yasemin Acar, Sascha Fahl and Matthew Smith: 1,047 participants and six textual formats, with hexadecimal more vulnerable to partial preimage attacks and a sentence-based format best for attack detection and perceived usability. Source ↗
  • Dominik Schürmann on LinkedIn, September 4, 2026, in German, on usable security: a password manager is only secure once as many people as possible use it. Source ↗
  • heise online, January 2023, testing heylogin, the password manager without a master password (in German). Source ↗
  • Golem.de, June 2023, on heylogin, the password manager that needs no password itself (in German). Source ↗
  • Wikipedia on usable security: the part of computer science, human-computer interaction and cybersecurity concerned with the user interface of security systems, going back to Saltzer and Schroeder's psychological acceptability in 1975. Source ↗
  • Wikipedia on public key fingerprints: short sequences of bytes, made by hashing a public key, used to identify a longer key. Source ↗
  • Wikipedia on Signal: users compare fingerprints out of band, and one study found 21 of 28 participants failed to compare them correctly. Source ↗
  • Wikipedia on USENIX: a nonprofit membership organization founded in 1975 in Berkeley, California, that organizes conferences. Source ↗
  • Wikipedia on design by committee: a pejorative term for a project with many designers and no unifying plan or vision. Source ↗
  • Wikipedia on the German identity card: an electronic ID card since November 2010, with a six-digit PIN for online authentication and a PUK after three wrong tries. Source ↗
  • Wikipedia on OpenKeychain: a free and open-source Android app for OpenPGP encryption, forked from Android Privacy Guard in March 2012, version 2.0 in January 2013, and in maintenance mode since August 17, 2021. Source ↗
  • Wikipedia on Pretty Good Privacy: developed by Phil Zimmermann in 1991, following the OpenPGP standard in RFC 4880, which Wikipedia calls historical, without support for the updated formats in RFC 9580. Source ↗
  • Wikipedia on Mozilla Thunderbird: a free and open-source email client made by MZLA Technologies, a subsidiary of the Mozilla Foundation. Source ↗
  • Wikipedia on Proton Mail: an end-to-end encrypted email service launched in 2014 in Geneva. Source ↗
  • Wikipedia on K-9 Mail: a free and open-source email client for Android. Source ↗
  • Wikipedia on password managers: software that stores usernames and passwords in a vault whose key is commonly derived from a master password. Source ↗
  • Wikipedia on 1Password: from the Canadian company AgileBits, with a vault locked by a master password and hosted by default on 1Password's servers for a monthly fee. Local vaults were removed in 1Password 8, which drew criticism. Source ↗
  • Wikipedia on single sign-on: logging in once, with one ID, to several related but independent systems. Source ↗
  • Wikipedia on software as a service: begun around 2000, and by 2023 the main way software was deployed. Source ↗
  • Wikipedia on multi-factor authentication: access only after two or more distinct types of evidence, which protects against someone who has discovered a single password. Source ↗
  • Wikipedia on secure cryptoprocessors: dedicated chips for cryptographic operations with physical security measures that make them tamper resistant. Source ↗
  • Wikipedia on time to live: a mechanism that limits the lifespan of data in a computer or network. Source ↗
  • Wikipedia on freemium: a pricing strategy with a free basic product and paid features, used in software since the 1980s. Source ↗
  • Wikipedia on LastPass: founded in 2008, bought by LogMeIn, now GoTo, for $110 million in 2015, and breached in late 2022. Source ↗
  • Wikipedia on Dashlane: a subscription password manager founded in Paris on July 6, 2009. Source ↗
  • Wikipedia on the first-generation Tesla Roadster: a battery electric sports car on the Lotus Elise chassis, produced from 2008 to 2012, 0 to 60 mph in 3.7 or 3.9 seconds. Source ↗
  • Wikipedia on the Tesla Model S: the company's second vehicle, a four-door car whose series manufacture began in June 2012. Source ↗
  • Wikipedia on Brooks's law: Fred Brooks's 1975 observation that adding manpower to a late software project makes it later. Source ↗
  • Wikipedia on Mozilla: Mozilla Ventures, announced on November 2, 2022, at Web Summit in Lisbon with a $35 million starting fund, named HeyLogin among its first three investments. Source ↗
  • Wikipedia on Sequoia Capital: a venture capital firm in Menlo Park, California, investing from seed to growth stage. Source ↗
  • Wikipedia on fail fast: the business idea of using trial and error to judge a product or strategy quickly and move on, part of Silicon Valley's culture. Source ↗
  • Wikipedia on Figma: founded in 2012 by Dylan Field and Evan Wallace, with its first public release on September 27, 2016. Source ↗
  • Wikipedia on passkeys: key pairs created by the user's authenticator, built on FIDO Alliance and W3C WebAuthn standards. Source ↗
  • Wikipedia on the FIDO Alliance: launched in February 2013 to reduce the world's reliance on passwords. Source ↗
  • Wikipedia on TU Braunschweig: founded in 1745 as the Collegium Carolinum, and a member of TU9. Source ↗
  • Wikipedia on Aalto University: a public research university in Espoo, Finland, formed in 2010 from a merger of three universities. Source ↗
  • Wikipedia on the Federal Office for Information Security (BSI): the German federal agency founded in 1991 in Bonn. Source ↗
  • Wikipedia on the German Academic Exchange Service (DAAD): the joint organization of German universities, founded in 1925. Source ↗
  • Wikipedia on Lower Saxony: a coastal state in northwestern Germany, where Braunschweig is. Source ↗
  • Wikipedia on the Pirate Party Germany: a party founded in September 2006, part of the international movement of pirate parties. Source ↗
  • Wikipedia on Visual Basic: Microsoft's name for a family of programming languages, with the original version supported from 1991 to 2008. Source ↗

Transcript

Automatic transcription. Jon McLachlan hosts this episode alone with Dominik Schürmann, co-founder and CEO of heylogin. There is no diarization on the tape, so speaker labels were assigned from the content and checked against reference recordings of each voice from this episode, passage by passage and on the short turns. The two voices are close, so a few short lines may sit with the wrong speaker. Names and product names were normalized where the recognizer misheard them (Jon McLachlan, Dominik Schürmann, heylogin, COTECH, Aalto University, TU Braunschweig, OpenKeychain, USENIX Security, 1Password, LastPass, Dashlane, Mozilla). Words in [brackets] are best guesses where the audio is unclear. Answers are transcribed as spoken, with repeated words and false starts trimmed. Short backchannels such as "yeah" and "right" are left out where they overlap the other speaker.

Read the full transcript 4,136 words

Jon McLachlan

Hello and welcome, everyone, to another episode of the Security Podcast in Silicon Valley. I'm here today with a very special guest, Dominik Schürmann. Welcome to the show, Dominik.

Dominik Schürmann

Thanks for being here, Jon.

Jon McLachlan

No, thank you. So Dominik is the founder and CEO of heylogin. Before that, going all the way back, he was actually a visiting researcher at Aalto University, a research assistant at TU Braunschweig. You're also a visiting researcher at Aalto University. You did some freelance cryptography engineering work for a couple of years there, it looks like. That's awesome. You came back to TU Braunschweig as a scientific staff, where you focused in on difficult and open security problems, it sounds like things like end-to-end encrypted communications. And you ended up starting a company, COTECH. You are the CEO and co-founder of COTECH. And now you are heads down, focused on heylogin as co-founder and CEO of heylogin.

Dominik Schürmann

Right, that's correct. So actually, we started as COTECH in 2018. And I always say 2018 and 2019 were kind of the years of finding an actual problem that we should solve. So at one point, we found the problem of password management, and then we focused 100% of heylogin, and then we renamed the company. So actually, it's the same entity, but it's now called heylogin.

Jon McLachlan

Awesome. So it's such a good pivot that they look like two distinct companies. Good job. I think usually sometimes when people do a pivot, it's much softer around the edges. But it looks like, you know, from your experience in laser focus and security, maybe you wouldn't mind sharing with our audience just maybe a pinch about your childhood, or maybe what parts of that shape who you are today.

Dominik Schürmann

Yeah, yeah. So, I mean, I wasn't a popular kid, right? But I started, actually, I started with Visual Basic for Kids. It was a great book. I think I started at 14 years doing some programming using Visual Basic. And then I was deep down into modding of a computer game and stuff like that. But when I really got started with cryptography, it was during my university time. So I was quite interested in political movements in Germany, like the Pirate Party, and got some involvement there. And political involvement sometimes leads to cryptography, I would say.

Jon McLachlan

Yeah, yeah, [sometimes it] does, doesn't it? That's funny how that works.

Dominik Schürmann

So that got me involved into cryptography. I did a lot of studies in cryptography during my university and bachelor and master, and then I focused even more doing my PhD and work as the scientific staff. My research was always like finding trade-offs between usability and security. So it's not that easy, but it's an interesting way to think about. We did, for example, studies with over 1,000 people to find which kind of key fingerprints compare best. So stuff like that.

Jon McLachlan

Very nice. I imagine that came up with some interesting results there. So, different representations of key fingerprints.

Dominik Schürmann

Yeah, it's really interesting. It was one of the best papers I did during my time. This was one of the USENIX Security papers I did with a colleague of mine. And the way you can think about key fingerprints is you can just do it's hexadecimal, like in PGP, you can do it numeric, like in Signal, or you can even, like, have phrases, like in these backup phrases for the backup [seeds] for Bitcoin. So a lot of different ways, but they all have some upsides and downsides. Like, for example, the phrases, they are only in English, so they are not universally appliable. That's why [Signal switched] to numeric. And yeah. And we also did some studies on how long people need to compare these fingerprints and stuff like that. So that was pretty amazing.

Jon McLachlan

So as you were in your studies and you were focused on all of these difficult and open security problems, and really it sounds like the intersection between security and usability, was there something that started to emerge that was a truth that you really saw in the world that wasn't noticed by most people?

Dominik Schürmann

Yeah, I thought about that question. So, especially in the IT security space, there are a lot of people, especially at an older age, who think like an attacker, right? Security people are trained to think like an attacker, and that means they are always trying to find corner cases. Like, this is a very specific attack, and this means the system is not secure. So all these, like, these sorts of processes lead to extreme choices in security protocols, especially if you design security protocols by a committee.

Jon McLachlan

Oh no, by a committee, oh no.

Dominik Schürmann

That's why some people call it death by committee. Yeah. In Germany, there's a lot of examples, like the ID cards in Germany. It's some very interesting cryptography involved, but it's so difficult to implement that these ID cards don't work with most of Android phones, and they don't work with most card readers. So I think this is a bad way to go at security protocols. And I had some very, very interesting experiences over several years where I maintained the most widely used PGP implementation for Android. It's called OpenKeychain, which still has 300,000 users. And it's, I came from the same point of view. You need to take care of all attack scenarios. But that just leads to no adoption. So I think the big insight is here to find a trade-off. And if your security measures prevent adoption, then they don't help, right?

Jon McLachlan

Right. You have the most secure piece of software or hardware in the world, and if nobody uses it, are we really making a difference in the world?

Dominik Schürmann

Exactly. That's like, PGP is one of the worst examples, and it's not getting better. So I'm a little bit following the current standardization process of the next PGP version, and it's a disaster. From my point of view, now, looking at it from a distance.

Jon McLachlan

I'm sure there's a strong sense of appreciation for some of the dedication that's behind it. It's just maybe if there was more of a balance between the usability and the grappling with those difficult security problems. I know it's very easy to get sucked into rabbit holes.

Dominik Schürmann

Right. I mean, I know some of these, most of these people personally. We had a lot of meetings of all the PGP implementation people, like the Thunderbird guys, then the Proton Mail people. But it's, yeah, it's a difficult subject. Just my point of view now is they're prioritizing the wrong stuff.

Jon McLachlan

And was this some of the creative juices that nudged you towards building heylogin?

Dominik Schürmann

I mean, we, during my time as a PGP tool maintainer, we had a lot of insights. My co-founder then also got involved in this project. That's why I know him. And we always try to innovate on this very broken standard. And at one point, we just thought, okay, it's not fixable. Let's do something else where we can design everything from the ground up. And I think password managers are a very similar concept that is pretty, let's say, it hasn't changed a lot since its inception. Like, 1Password is 16 years old now. So in the IT world, that's a long time. So it's a similar space where there needs to be new [thoughts] and new ways to fix the problem, because if you look from the business point of view, more and more passwords are being used every day. I just saw a statistic that many B2B companies, on average, B2B companies have 110 software as a service platforms in use. So if you cannot connect them to a single sign-on, you need a password manager. So obviously, the market is growing, but the tools are still targeted to an IT audience. And that's what we are trying to change. heylogin focuses on usability, and that's why we think differently in a lot of aspects of password management.

Jon McLachlan

So heylogin is a password manager that is obsessed with the usability of the experience and trying to aim for security through usability. Because once you get the usability, you get the adoption, you get the ease of use, you are actually taking the burden off the shoulder of the user.

Dominik Schürmann

Right, exactly. So one of our experience during our time as consultants in the early years, like 2018, 2019, was that even if companies deploy a password manager, their employees which are forced to use it just write down the master password on a piece of paper and put it on their desk. It's like a very similar problem as without a password manager. It's just moved to a different tool.

Jon McLachlan

It's the same problem. It's just transitively [shifted] to 1Password, and now in security terms, you have a single point of failure for all your passwords.

Dominik Schürmann

And it's still a one-factor solution. If you just use the master password, it's just one factor, something you know, so knowledge. 1Password and co, they have an option for two-factor security, but most people don't enable it. So it's not like there is a huge adoption of these options. So again, for, I don't know, 90% of our users, it's just the master password. And that's where we are trying to think differently. In heylogin, there is no master password, no password at all. Everything is end-to-end encrypted using the smartphone of the user. So it's two-factor secure by default, and even easier, because you have just the factor of possession, the security chip in your smartphone, and the second factor is the PIN or face unlock or fingerprint. And these two factors are easier to remember and use in comparison to a master password.

Jon McLachlan

Yeah, yeah, they are. I got the whiff that there's nothing really that I have to, well, maybe a PIN, but my Face ID or a biometric is quite convenient. [A PIN] is easy to remember, though, I think so.

Dominik Schürmann

Yeah. And the PIN is more secure than a master password, because the PIN is bound to your security chip in your phone, and there is a limited number of tries. So it's more secure.

Jon McLachlan

So help me with the user experience. If I'm on heylogin and I'm on my desktop and I'm in, let's say, just Safari, and that password that I need to enter that's 100 characters long needs to go, does it go from my phone into my local machine somehow?

Dominik Schürmann

Yeah, yeah, that would already be a deep dive, but I can try to explain [in short terms].

Jon McLachlan

If you just tease us a little bit.

Dominik Schürmann

There are several cryptographic keys involved here, but let me just say the most important key is the one that is generated at first-time use inside the security chip of your phone, and this key cannot be extracted from your phone. And from this key, we derive several other cryptographic keys for different purposes. And so if you go to a website in Safari on your laptop and you need to log in, and you click a button, and then you get a notification on your phone. On your phone, you can just swipe to login. That's how we call it, swipe to login. And then you're logged in on the desktop. So this is the user experience. And when you're doing this, you just need to do it once per day. Then a temporary cryptography key is generated for this browser instance for this day, and then the browser has access to the password [vault] for the specific time until this cryptographic key is revoked by your phone. So we don't need to transmit all passwords from the phone to the browser. It's a little bit more complicated, but that allows us better responsiveness.

Jon McLachlan

Very nice. Okay, I like it. So it sounds like we've got a very high security bar here, and things are expiring on a 24-hour basis, which is amazing. But still obsessed about the user experience, and just log in on your phone. I like that. I do like that, because then now my computer is much more secure, in the sense that any sensitive data is going to have a TTL, essentially. So I don't have to worry about future exposures to risk.

Dominik Schürmann

You can always also open the app on your phone and revoke at any time you like, if you don't like to follow the 24-hour protocol.

Jon McLachlan

What's your go-to-market strategy with heylogin? Are you a B2C play? Are you B2B? Are you top-down or bottom-up or...?

Dominik Schürmann

Bottom-up. So, I mean, 1Password, LastPass, Dashlane, they all have a private user plan. We currently don't have a private user plan, because our private user plan is free. You can just use it, it's for free. I don't think there is a lot of market left for B2C in this space.

Jon McLachlan

It's pretty saturated, huh?

Dominik Schürmann

Yeah, and people are, I mean, the IT people who happily pay, they all already have a password manager. They don't easily switch to another paid password manager, but they may switch to a free plan and a better password manager. So it's kind of like the typical product-led growth strategy, where you have a freemium plan and then do an upsell. So that's what we do. Our B2B plan is the business model.

Jon McLachlan

Yeah, exactly. It sounds like you've got that, and you got your go-to-market strategy locked in and everything. And so, hey, maybe a [softball] question. What's been the best day that you've had so far at heylogin, your journey?

Dominik Schürmann

Yeah, so when we got the first reviews of our B2B features, it was quite a great week. We made some smaller marketing campaign to get the word out, and we get some pretty good deals for early adopters, and they wrote really, really good reviews. So that kind of proved our strategy and our usability. There were some funny moments. Like, I remember one review, we printed it out and hang it on the wall. He was quite happy with heylogin. He said it's like the best user experience ever. And but then at the end, he wrote, it's so easy to use that it doesn't feel so secure. So it's quite funny, because it's a compliment at the same time.

Jon McLachlan

That it's...

Dominik Schürmann

Yeah.

Jon McLachlan

He's been so trained, or maybe he or she has been so trained to expect that security means it's going to be difficult in some way. It's going to be an impediment to the flow of your day, a roadblock, an obstacle, or something.

Dominik Schürmann

Yeah, that was pretty funny.

Jon McLachlan

This is always what gives me hope, that not just that we have smart people working on these difficult problems and we have a good handle in the direction that we want to take them, take these problems, but that we're focused on the human side of things, not just the technical side of things. And when you get a review like that, I'm sure you replied, it's something that maybe in 20 years we could look back on and be like, oh, do you remember when password management was so bad, and you had reviews like this, to try to break the mold and the stereotype and negative image around them. You know, I imagine this is why Tesla, the very first car that they built, was basically a race car. You know, they're trying to break the stereotype: electric cars are slow, less of a car, whatever. So break that stereotype with a very fast race car, then go out and build the sedan. It's almost kind of like what you're doing with security in the password management space. Breaking the stereotype [that password] management has to be an interruption to the flow that they, you know, [had in mind].

Dominik Schürmann

Regarding that topic, there is, so, in my opinion, there is even a second side to that. Not just the product side, product development side, where we try to be more usable and different than the existing solutions. There is even the side of marketing. So IT security products are often sold with what I call fear marketing. People should fear data leaks. They should fear what if people leave the company and take passwords with them. Like, there is all these kinds of scenarios. Like, the most costly ad Dashlane ever ran was a Super Bowl ad. So there is a guy in this ad, he dies, and then he gets to the heaven, I think, or something, and then he is asked what is the password. It's, again, it's kind of fear marketing. So I don't think that's the right way to find traction in this market. I think you need a different angle on that. So that's why we try to go the way of showing how heylogin provides more productivity for your employees.

Jon McLachlan

I'm with you there. Fear-mongering or even ambulance chasing never has, I've never really seen it pay off in any long-term success. Okay. The worst day that you've ever had in your journey with heylogin?

Dominik Schürmann

So, you know, when we started developing heylogin in March 2020, we were able to start with a grant. We got a big grant from the German government that was pretty great to do the development. And nearing the end, we thought, okay, now it's time for chasing the first venture capital round. And we were pretty confident that it shouldn't be a problem. That was quite naive in retrospective. We had a pitch at a very, very local VC. It's like the VC for this specific German state we are in, in Lower Saxony. It's partly governmental. So we were confident, just do it, and we get the money, and it didn't really work out. So that was a really bad experience, and it kind of grounded us. So, yeah, I was down for a week at least.

Jon McLachlan

Yeah. Expecting a round to be able to pull through and then having to fall apart can be detrimental to the business, but also to those that are shepherding the company through the process and everything. So I understand that completely. No, and thank you for sharing. So those are vulnerable moments, not just for companies, but for the leaders in those companies too.

Dominik Schürmann

It also took a long time to, let's say, to take all the feedback and start a new [round]. There was a lot of stuff that we needed to learn before. So we took an external consultant for this specific round, and then it took some months until we started talking to VCs again, and then we were better prepared. But yeah, it still took more time, let's say.

Jon McLachlan

Yeah, sounds like you had a crash course in an MBA.

Dominik Schürmann

Yeah, some kind of, yes. Yeah, we had like three meetings per week with this consultant. So, yeah, it was kind of a crash course.

Jon McLachlan

But good, really. The growth mindset at work and at play, it's been able to pull you forward.

Dominik Schürmann

Yeah, that's true.

Jon McLachlan

So if you fast forward into the future, and I'll let you decide how far into the future we want to look, but what does success look like in your mind for heylogin?

Dominik Schürmann

Yeah, so currently we have a very conservative finance plan. We plan for cash flow positive next year, without a lot of growth. I mean, the VC space currently, let's say I'm a little bit hesitant to raise another round right now.

Jon McLachlan

So, I mean, given the current economic state of the world, of Germany in particular, of the region, energy prices...

Dominik Schürmann

Ukraine, [Belarus]. Yeah, it's a difficult situation right now. I mean, we started during COVID in March 2020. We're kind of used to it, but yeah, it got even worse, let's say, the state of the world.

Jon McLachlan

You have to balance all of those responsibilities. I mean, being a leader of a company, there's people that depend on you for their livelihood, and you yourself perhaps have people that depend on you.

Dominik Schürmann

Yeah, definitely. Yeah, that's why we very currently just plan for a very, we plan for cash flow positive. The development team stays the same after a current round. We don't hire more developers, because, as you may know, just having one more developer doesn't mean you're faster in product development. So what we do is put our money into sales and marketing and keep the current headcount in development.

Jon McLachlan

I understand that entirely. It's almost where that phrase came from, that nine pregnant women don't make a baby in a month. I wish there was a phrase like that that was more applicable in a gender-neutral sense, but it's kind of a similar thing of what goes on in software development. Do you have pressure underneath you from the VCs to hire more and to spend more and to not be cash flow positive? Or are they pretty supportive? Do you find that your VCs and the folks that are invested... pushy?

Dominik Schürmann

So, or, yeah. I think we have a pretty interesting combination of business angels and one bigger brand. So it's very, very local business angels which live in this area here, plus Mozilla, who invested in our company. So the local business angels, let's say they have a pretty German perspective, so they are happy about being cash flow positive. They don't necessarily plan for growth. So on the other hand, Mozilla is not like the, it's not like Sequoia or some other VC. There is not a lot of pressure right now, so that's good, let's say, for the current situation.

Jon McLachlan

No, that's very good. I think that might actually be more of the exception that I hear, but that's great that you've got true partners with your investors, not just folks trying to [flip it] on a very strict timeline, and are working with you and are flexible. So this is a good thing, and I'm happy for you, and for heylogin, too.

Dominik Schürmann

Yeah, I hope [these things look good].

Jon McLachlan

I'm sure it will, especially with strong leadership at the helm. Dominik, thank you so much for your time today. Thank you for sharing a few vulnerable moments and the inside look at heylogin and what drives it and distinguishes you from the other players in the space. I'm curious, would you like to leave any of our listeners with words of wisdom or maybe advice that you wish that you received at a younger age, but never heard?

Dominik Schürmann

You know, there is this idea of fail fast and try a lot of things and go to market early. I think it's only partly right. I mean, for example, if you look at Figma right now, they also took a lot of years in the beginning to try different stuff before developing Figma, and then they took some time to get a basic version. So after trying a lot of stuff, you should at one point stick to something. And then it takes some time until you get the first user base. So, yeah, it's a [marathon].

Jon McLachlan

And maybe a pivot or two.

Dominik Schürmann

Yeah, maybe a pivot or two. I mean, if you look closely, you would probably find more than one pivot in our company, but it's like the big pivot in March 2020, so that's...

Jon McLachlan

Nice. Well, personally, I'm in the market for a new password manager, because 1Password is so old, and they recently, in their latest version, stripped their peer-to-peer service, where you can push passwords from your local network, from your local device onto your other local devices. And I just don't want my stuff in their cloud with only a password protecting it. So, long story short, I'm in the market, and I'm definitely going to give heylogin a try. And if anyone out there is listening and is curious to see what all of this is about, and curious to see what a good user experience might actually feel like, I would highly encourage them to give it a spin, too.

Dominik Schürmann

Thanks, Jon.

Jon McLachlan

To be a small part of that journey, to be a small part of history.

Dominik Schürmann

To be very special, you know. It's kind of like, thanks a lot.

Jon McLachlan

Why a lot of us go into security, you know. We notice something that's wrong with the world, and we want to be part of [it].

Dominik Schürmann

Definitely. Yep, definitely. Awesome.

Jon McLachlan

Okay, Dominik. Well, thank you again. Thank you so much. And thank you to all of our listeners for tuning in to another episode of the Security Podcast in Silicon Valley.

FAQ

Questions this episode answers

Who is Dominik Schürmann?
Dominik Schürmann is co-founder and CEO of heylogin, a password manager with no master password, based in Braunschweig, Germany. He holds a PhD in IT security, was a research associate at TU Braunschweig from 2014 to 2018, worked as a freelance cryptography engineer for the BSI and ENISA, was a guest researcher at Aalto University, maintained the Android encryption app OpenKeychain, and co-founded COTECH in 2018, the company that became heylogin.
What is heylogin?
A password manager with no master password, from heylogin GmbH in Braunschweig. Logins are confirmed with a swipe on the phone, whose security chip holds the key, or with a FIDO2 security key, Touch ID, Windows Hello or a smartwatch. It has a free private plan and a paid business plan.
What was COTECH?
The company Dominik co-founded in 2018. Its Hardware Security SDK let Android developers use security keys and smartcards. After 2018 and 2019 spent looking for the right problem, it focused on password management and was renamed heylogin.
What did Dominik's key fingerprint study find?
The 2016 USENIX Security paper tested six textual formats with 1,047 participants. Hexadecimal was more vulnerable to partial preimage attacks, a sentence-based format had the highest attack detection rate and the best usability perception, and plain numbers still beat hexadecimal.
Why does thinking like an attacker cause problems?
Security people are trained to find corner cases, and every corner case means the system is not secure. That leads to extreme choices in protocols, especially by committee, and to protocols so hard to implement that people don't adopt them.
What is OpenKeychain?
A free and open-source Android app for OpenPGP encryption that Dominik maintained for several years. It started in 2012 as a fork of Android Privacy Guard and has been in maintenance mode since August 17, 2021. Dominik said on the tape that it still had 300,000 users.
What does Dominik think of the next PGP version?
He was following its standardization from a distance and called it a disaster. In his view, the people behind it, many of whom he knows, are prioritizing the wrong things.
Why do companies need a password manager if they have single sign-on?
Because not every app connects to single sign-on. Dominik cited a statistic of 110 SaaS platforms at the average B2B company, and any app outside single sign-on needs its own password.
What goes wrong when companies force a password manager on people?
In Dominik's consulting work in 2018 and 2019, employees wrote the master password on paper and put it on their desks. Jon called it a single point of failure for every password.
How does heylogin work without a master password?
A key is generated inside the phone's security chip at first use and can't be extracted. To log in on a laptop, you click a button in the browser and swipe on the phone. heylogin's security page says the chip is unlocked locally with a fingerprint, face or PIN.
Is a PIN more secure than a master password?
Dominik says yes, because the PIN is bound to the phone's security chip and allows only a limited number of tries.
How long does a heylogin browser session last?
You swipe once per day. A temporary key is generated for that browser for that day, and you can revoke it from the phone at any time.
How does heylogin make money?
The private plan is free, and the B2B plan is the business model. Dominik calls it a typical product-led growth strategy: a freemium plan, then an upsell.
What did the review on heylogin's wall say?
The reviewer called it the best user experience ever and then wrote that it was so easy to use that it didn't feel secure. Dominik took it as a compliment.
What is fear marketing?
Dominik's name for selling security through fear of data leaks or of employees leaving with passwords. He pointed to a Dashlane Super Bowl ad, by his account the most costly ad it ever ran. heylogin pitches productivity instead.
Who invested in heylogin?
Local business angels and Mozilla. Mozilla Ventures, announced in November 2022 with $35 million, named HeyLogin among its first three investments, and heylogin dates the investment to its pre-seed round in fall 2022.
What advice does Dominik give founders?
That fail fast is only partly right. Try a lot of things, but at some point stick to one, and expect a basic version and a first user base to take time, as they did at Figma.
Keep reading on the YSecurity blog
New episodes every other Tuesday

Got a story like this one?

We're always looking for practitioners who've been in the room when it mattered.