The Next Usability Frontier
How web3 changes the Friction Layers model, increasing interface variance, and why this matters
Hey V&V fam ✌️
Got a good one for you today, lots to say on this topic. This piece was originally published on April 4, 2022 - bit of a delay in sending the newsletter. As always, thank you for reading. If you’ve found your way here but haven’t subscribed yet, join us.
A little over a year ago, I wrote about a concept I called Friction Layers. The gist: Most products miss opportunities to reduce friction at various “layers” of the user experience. Each layer represents a set of potential interaction points or modes as a product mediates the exchange of data, dollars, and ideas with and among the audiences it serves.
At the time of writing, web3 wasn’t really a thing yet, at least not as an organizing idea commonly known as “web3”. But now…it’s a thing. I think web3 maps to the Friction Layer model but warps it in a way that opens up some big implications:
TL;DR—Friction Layers in a web3 world
With a paradigm shift to “ownership” and private key responsibility, web3 inverts the friction layer model, reintroducing major friction to the gen pop user experience
Why it matters: Usability becomes a differentiator again
Web3 unbundles interfaces from databases (composability), freeing developers and designers to focus on a particular friction layer, like usability
Why it matters: web3 necessitates—and uniquely enables—design innovation that will break through decades-old tropes
Web3 will increase interface variance, augmenting user choice in how to experience the web
Why it matters: The relationship between interface, brand, and audience will tighten—increasingly, we’ll triangulate a product’s position in the culture based on its interface(s)
I. Web3 is bringing friction back
The average person, a gen pop internet user, let’s say, may be aware of crypto but is not conscious of “web3.” And even for most of the early adopters currently participating in web3, the learning curve to basic participation is steep. Precariously so, at times. Am I doing what I think I’m doing with this wallet or address that I control? Is it properly backed up? How do I confirm the transaction went through? Am I making a security misstep here—how do I know this is safe? Perhaps, like me, you’ve experienced lots of web3 click hesitation along these lines.
Why may seem obvious, but let’s break it down.
As @cdixon and others have noted, web3 marks a step change to the ownership era of the web. Many of us believe there is unbounded good in this, with incredible benefits conferred to users—to the PEOPLE—through ownership. But the jump to ownership hinges on a user’s ability to adapt to a paradigm shift in usability: it demands the user be responsible for their private key. Herein lies something beautiful and treacherous.
NOTE: For simplicity, I’m equating private keys with seed phrases throughout this piece.I’m also making an assumption that the reader understands why encryption and private keys are so important. To sum it up, I like this post from @nneuman, in which he writes: “Private keys are a critical technology for empowering individuals…Control is power, thus encryption endows you with digital power. Private keys unlock a future where you have true ownership over your money, data, identity, and more.”
“Ownership” on the web does not map to non-internet analogs like “reading” and “writing” on the web do. We all grow up reading and writing, then apply those ingrained skills to web-based behaviors (with lots of help from interfaces to abstract away the complexities, of course). But managing and wielding private keys, and understanding the risks and implications of doing so, presents an unprecedented usability challenge for the mainstream.Yes, there are custodial services that allow users to offload this burden and acclimate to web3 via a fiat ramp and familiar web2 UX, but let’s focus on the true vision of web3, the non-custodial experience.
I agree that Read > Write > Own is the best pithy mental model we have for web3. But “own” doesn’t adequately capture the mindset and behavior required of the user; it suggests passivity. If I were to rephrase it in terms of what the user actually needs to do, I might offer something much more clunky, like:
Web3 invites us to participate in a trustless system. To participate, we need to accept a minimum viable friction bar:being a custodian of our own keys and authenticating and transacting without a safety net if the UX gets too complex or if the things we own are lost or sent in error. Owning in this context demands a new type of active orientation toward the web, and that’s not implicit in the word “own”.
As much as web3 unlocks and liberates, in doing so it puts fresh onus on the user. This is the usability paradigm shift. It’s not that web3 doesn’t deliver wonderful, fluid user experiences that are improvements over web2 (it does); the point is, there’s a cognitive cost just to walk in the door of web3, which the average consumer is not prepared to pay right now. Historically, we’ve relied on institutions as the interfaces through which we interact with, and participate in, complex, technical systems most of us don’t understand. But web3 won’t afford us the same type of hand holding.
This is different, but related to, an important point that @moxie makes in My first impressions of web3: “People don’t want to run their own servers, and never will…I don’t think this can be emphasized enough—that is not what people want.” Many people will not want to switch over to “ownership mode” on the web (not in the current state of web3 UX, at least) once they find out what it requires of them.
Considering crypto at large, surveys suggest we’ve reached mainstream awareness—but we’re at a much earlier point on the web3 adoption curve. If we sketch a funnel, with awareness at the top, moving down to → ownership and then to → usage, at each level down we must discount the population size significantly, taking us further away from the mainstream.
86% of American adults say they have heard at least a little about cryptocurrencies
16% of U.S. adults (~33 million) report having invested in or traded cryptocurrency. This number jumps to 43% among men 18-29 (Nov. 2021)
Globally, estimates put total crypto ownership at ~295 million users, ~23 million of which are ETH users
Usage (regular activity)
Metamask, the leading non-custodial wallet, surpassed 10 million global MAU in 2021
Instagram, which we could use a web2 exemplar, has well over 1 billion MAU (100x)
Even with rapid growth in key web3 apps like Metamask, over the last year daily active ETH addresses (uniques) have hovered around 600,000 and have not surpassed the 1 million mark
The web3 user base is relatively small and frequency of use among early adopters is limited
Usability - examples of major issues
(These examples serve to qualitatively illustrate the seriousness of irreversible usability gaffes, focusing on user error—these examples do not speak to hacking or fraud, other vulnerabilities.)
Billions worth of Bitcoin have been lost by people who’ve mismanaged their private key(s)
Over 12,000 ETH lost forever due to typos (sending to a non-existent address)
$300k Bored Ape NFT accidentally sold by owner for $3k in “fat finger error”
D.C. family can’t access $5.8M in ETH; the Ethereum Foundation's legal team responds, “We continue to believe that Ethereum has no liability for lost wallets, passwords, and Private Key”
Overall ownership and usage are growing fast, the momentum is there. But the population currently ready and willing to participate in web3—active users of non-custodial apps (i.e., people who control their own private keys)—shrinks considerably as we move down the funnel, because web3 introduces major usability friction ubiquitously.
Consider adoption of two-factor authentication, which requires users go a step beyond the basics of username + password. Referencing this report from Duo, we find:
79% of people (US & UK adults) have used two-factor authentication—we can interpret this as high adoption with significant YoY growth (up from 28% in 2017)
However, “only a minority of respondents (32%) report using 2FA on all applications that offer it…For example, some respondents may enable 2FA on their bank account but not their email.” So, most people are selective with 2FA, not ready to apply it as part of their default web login experience.
SMS and Email are far and away the lead options that people use as a second factor, but these factors are conveniently sent by another party or institution, and they’re easily regenerated—lose it or forget it and another one can be sent.
Meanwhile, adoption of other factors like security keys and hardware tokens lags far behind. I think there’s important signal here: These factors, very different from SMS & Email, are more akin to a web3 token (private key or wallet seed phrase) in that they’re things you need to have rather than know (one can’t be expected to memorize a 12-word seed phrase), and they can’t be sent to you or recovered by another party.
We might interpret these 2FA data as proxy evidence that the main mechanism of web3 (trustless authentication) will present a usability challenge. To “log in” to web3 and safely participate, one must assume responsibility for a key they likely don’t know how to handle.
I love this visual from @jackbutcher in which he shows the flip side of this, the promise of improved UX in web3. We’ll seamlessly connect to apps and transact using a much more elegant form of identity verification, a new universal login.
This is right on, but the chasm we need to close to make the trustless system feel this usable by the gen pop is completely non-trivial. I’m betting on it happening and trying to contribute to the effort, though IMO it’s not an inevitability. Mass adaptation to a paradigm shift never really is. How might we design for a web in which user fallibility at scale AND cryptographic immutability are part of the laws of physics of that web?
Usability will once again become a differentiator. And that brings us back to the Friction Layers model.
A quick recap (web2 version). If we assign a quality to each friction layer, against which a product or brand can be assessed, it might look like this:
Web3 almost wholesale inverts the model, dissolving friction at the more “advanced” layers— Authenticity, Lucidity, and Liquidity—while reintroducing major friction to the baseline UX for gen. pop. users. Web3 usability is not table stakes, mainstream participation is not a given.
Reimagining Friction Layers for web3, it’s striking how much of the original model remains intact and is simply inverted:
Authenticity, which becomes the baseline friction layer in web3, is the one layer that requires a new definition when updating the model. In web2 terms, Authenticity is about engendering trust through design: UI/UX that is non-derivative, showing the brand’s/creator’s ability to provision interfaces that are both efficient (enabling fluid value exchange) and evocative (communicating values, meaning). In the language of a recent meme, web2 Authenticity is about authentically rotating shapes, not just wordceling. It’s about ingenuity that proves genuineness.
In web3, the Authenticity layer is hard coded, part of the protocol, and enables the system to be a trustless one. An inversion of web2, the design job-to-be-done now rests at the Usability layer instead.
Thinking of a web2 world, I wrote in the original Friction Layers piece, “better usability is not defensible in a world where fluidity can be engineered and distributed with zero capex required.” This did not account for the step change to the ownership era of the web and the usability paradigm shift it will entail. In web3, usability will once again become a differentiator.
II. Composability helps solve the friction problem
We’re early on the web3 adoption curve because the jump from Read/Write to Own introduces a fundamental usability challenge. What’s TBD is how early we are as measured in time (years). Can the velocity of web3 innovation bridge this design gap relatively quickly?
Making a trustless system sufficiently usable for the gen pop is not an inevitability. User fallibility at scale and cryptographic immutability need to somehow coexist in mainstream web3, and resolving that tension is a gnarly problem, kind of like trying to rework the physics of the web. So what’s encouraging people to move toward this new design frontier?
Composability. Heard of it? If you read and listen to the web3 cognoscenti, you probably can’t get away from it. So I’ll spare you my derivative take. What I’m interested in is what composability has to do with Friction Layers.
Some quick level setting. Composability is about leveraging the shared infrastructure of a blockchain, allowing developers to access and use compatible, permissionless building blocks from one program to another. For more, read Data composability: what it is + why it matters by @dazuck. I’ll use a few ideas from that post here, verbatim, to concisely make the point:
“Applications are basically interfaces + some business logic + databases…Today, databases are basically siloed…every application must have its own database to feed its logic and its interface…If an entrepreneur has a vision for an improved service or interface, she can't just build that. She has to build the whole stack - from scratch - and compete on all of it.”
“When database functionality is not siloed but open…any app can build on the same data. No app is a gatekeeper to it…This enables ‘permissionless innovation’ - anyone can build any new service (logic) or any new interface (app) on the same data layer.
“Why should every potential business need to build all 3 parts of this stack when their core innovation or value add comes primarily from one or two?”
Now, applied to Friction Layers: being able to unbundle the interface-logic-database stack means that teams have a chance to focus primarily on one friction competency instead of trying to boil the ocean across four. And because the trustless, composable system by nature dissolves much of the friction at the more “advanced” layers—Authenticity, Lucidity, and Liquidity—much energy can now be unleashed at the Usability layer.
In a reversal from web2, usability becomes a higher-order need in the hierarchy—the actualization of web3 depends on it. A lot of the design action toward this aim is currently happening with wallets, because that’s where one’s “key” to web3 lives. At root, if wallets aren’t usable (in the most complimentary sense of the term), the mainstream never gets onboarded to web3. So there’s been plenty of buzz, both sarcastic and earnest, about what “first principles wallets” look like:
(Is it just a coincidence that both of these tweets featured the typo “principals”? I dunno)
Many teams concurrently rethinking a design problem from first principles is a sign that ample resources have been freed to attack that particular problem. In this case, the nature of web3 shakes up the Friction Layers model, promoting usability to a higher-order, high differentiation position; and the composability of web3 allows teams to be hyper-focused at the interface layer, prioritizing the usability challenge without having to divert energy and resources to the rest of the stack (logic, data).
I think even looking at the names of emerging, fan-favorite wallets gives some signal about this: We’re starting to move from a “MetaMask” world to a more friendly environment with a “Rainbow”, a “Family”, and a warm “Glow”.
Web3 is bringing back harsh friction, unpalatable for the mainstream; this inspires design innovation (fresh UI patterns, embedded education, etc.) to make something that’s wholly unintuitive for most people start to feel accessible, safe, user-friendly; and composability allows builders in the space to double-down on this investment in interface design. Web3 necessitates—and uniquely enables—design that will break through decades-old tropes.
III. Interface variance and the web3 kaleidoscope
The previous sections posit that web3 brings about paradigm shifts in usability and composability. Here we explore what this may give rise to: a cambrian explosion in interfaces, anchored in authenticity by the protocols of a trustless system.
I’ve been thinking about this as a sort of kaleidoscopic internet experience. On one end of a kaleidoscope is the ‘object chamber’, where the objects to be reflected are stored. This is the data layer, the blockchain—what’s inside may tumble around and expand, but the contents are consistent for any viewer, openly knowable, verifiable. As a viewer peers into the kaleidoscope, rotating the device, she may alter the image and affect her interaction with the underlying objects by rotating the scope, changing the interface. And so the same data produces an almost endless number of novel patterns and experiences.
As @balajis says, the internet increases variance. Web3 will increase interface variance. “May the best frontend win.” This is about more than marking an exciting new era in design land—it holds deep implications for how products will form relationships with their audiences.
Currently, “databases are basically siloed — every application has its own”and this means we, as users, have little choice as to how we interact with the underlying data powering an application; we need just accept the corresponding interface, whatever’s draped over the database.
Interface variance at scale can create an abundance of choice in how we view and interact with the web. The magic of web3 is that regardless of our choice of modality, preserved is an underlying consensus reality with which we’re interfacing. Now, just because two people want to access and leverage the same set of information doesn’t mean they must conform to the same ways of seeing the world (the web) or moving through it. People may gravitate to applications based on whether the interface matches their particular usability needs, tastes, etc.
Following this logic, interfaces will become the primary selection criterion for where we spend our time and focus our attention, where on the web we let our loyalties develop. And therefore interface design will play a more important role in how products cultivate a brand and “communicate” values in the culture. Your UI/UX becomes your positioning (messaging still matters, but I think claims and persuasion around a value proposition don’t hit as hard when you don’t own the database from which much of that value is derived).
Q: If the underlying data or building blocks are open and composable, what will make people rotate over to your “channel” on the kaleidoscope?
An ability to dissolve the type of friction that’s most relevant to a particular set of users
An ability to unlock novel modes of experience: delivering a UX that satisfies the preferred phenomenology of a particular cohort, or of a mainstream moment-in time (interface zeitgeist)
In a 2020 piece called The New Mainstream, I wrote, “While we wait for the next big unifying idea or piece of content to come along, our attention runs through sub-communities: some relatively tiny and esoteric and exclusive, others millions of people strong but still not quite representative of the masses. We might think of this as provincial popularity—except the boundaries of a province in The New Mainstream are drawn by shared affinities and content habits and UX preferences, not geography and demographics.”
This was pre-web3 thinking but directionally points to where we might be headed in the next phase of the web. Going forward, audiences will be segmented, and will self-segment, differently (phenomenologically). However, we have not yet crossed the threshold into consumer web3. We’re at the first stage of audience segmentation by interface. Early adopters of crypto/web3 represent a certain level of sophistication: the kind of person who, with at least some know-how and confidence in their ability to possess a private key, is willing to participate in a “trade” without an institution mediating the exchange.
The root usability issue, the key friction layer at play, looms like a final boss before the mainstream can come online to web3 and a long tail of interface choice makes economic sense. To move from Read/Write to Own, we must onboard the mainstream to private key proficiency and help the average user ASSUME RESPONSIBILITY.
Thanks for reading this week’s V&V. Shout me on Twitter @marc_it. Oh, and help yourself to some footnotes below, if you’re into that sorta thing.
The non-technical user onboarding to web3 likely does not have a good grasp on what a private key is and may misconceive of keys/seed phrases, often confusing these concepts with wallets and addresses. Some may find it helpful to review these quick definitions before reading further: Private key: A secret number that allows users to prove ownership of an account or contract, by producing a digital signature; Wallet: Software that holds private keys, used to access and control Ethereum accounts; Address: A source (sender) or destination (receiver) involved in transactions on the blockchain. For full definitions, check out the ethereum.org glossary.
The original thinking on Friction Layers, and the focus in this piece, is mostly about the dissolution of friction. But the value-creative side of friction, key to crypto and web3, must be acknowledged. The simplistic logic: Ownership depends on asset security, security requires friction, and therefore friction plays an important role in enabling ownership. I don’t think we’re talking enough about how these principles of crypto and web3 align with the broader tech trend toward security, privacy, user sovereignty, and “ownership” of data across all types of applications that don’t involve cryptocurrency (see, for example: Apple, Messenger, Roam research). The common denominator is encryption and we will only see more traces of it in all forms of the web.
The main “unit” of web3 friction discussed in this piece is private key custody (one’s responsibility for their own seed phrase). There are multiple other friction points we could include, e.g. the best practice of using cold and hot wallets, understanding gas fees, etc. All contribute to the steep learning curve for people onboarding to web3. But the root of the usability paradigm shift we’re discussing here is the fact that non-custodial web3 puts the onus on the user to store and manage their private key(s).
Why say “ubiquitously” here? Because the paradigm shift-friction we’re discussing—user accountability for private keys, enabling trustless transactions—is a standardized part of the web3 user experience and will be everywhere, a fabric of the new web. Meaning, it’s not the typical “local” friction that arises from a particular UI pattern, or from some unique wicked design problem that afflicts a given app or industry.
It’s certainly possible for a committed user to memorize a mnemonic seed phrase, but the average person will fail or not even try in the first place. I’ve seen data on this in my research for an end-to-end encrypted product; in simulation studies in which gen pop users are asked to save and retrieve a private key passphrase (a mnemonic seed phrase or similar), success rates are pretty abysmal. Think about it intuitively: how many forgotten passwords have you needed to reset over the course of your web-browsing life?
Back to My first impressions of web3. @Moxie raises some caveats that should be noted here as I’m kind of casually waxing poetic about this ‘kaleidoscope effect.’ The viewer (user) is not peering directly and independently into the database, the client is not necessarily interacting directly with the blockchain - there are centralized platforms involved, directing traffic via APIs. E.g., “Metamask doesn’t actually do much, it’s just a view onto data provided by these centralized APIs” and “I don’t think it should be a surprise that we’re already at a place where your crypto wallet’s view of your NFTs is Opensea’s view of your NFTs. I don’t think we should be surprised that OpenSea isn’t a pure ‘view’ that can be replaced…”
Of course there’s plenty of choice on the web: I may choose from hundreds of social media apps, and maybe UI/UX considerations drive my time spent on Twitter vs. Instagram vs. TikTok, but it’s much more likely that I’m choosing one over another for its unique social graph or content. I may choose Spotify or Apple music based on UI/UX preferences (e.g., which offers the discovery algorithm or playlist features I like), but these services still differ at the data layer—the music catalogs are not identical and the podcast offerings vary considerably.