December 12, 2022

Crypto Products, But Increasingly Useful

Written by:
zee prime blog marginal utility.png

“We’re in this love together,

And like berries on the vine,

It gets sweeter all the time”

Al Jarreau, We’re In This Love Together

Zee Prime harvesting that sweet sweet utility

This is how the blog article from Bill Gurley started in June 2003, just around the time when web1 was giving way to web2. It describes business positioning which leads to increased customer engagement thanks to marginal product improvements every time it is used. If we believe we are on a junction of web2 and web3, we should be asking how useful these web3 products actually are.

The idea behind Increasing Marginal Utility (let’s call it IMU in this article) is to build something that delivers a better service or product for consumers every time they interact with it. Instead of locking in your customers through “switching cost” (the cost/effort required to switch), you should lock them in through “switching loss” (the benefit lost when leaving the platform). This induces a healthy form of FOMO, driving user retention and engagement higher. Without these two attributes, no application, protocol or company makes economic sense.

Network effects and IMU

As with any domain, there is rarely a pure one-sided interaction with the product. Buyers need sellers, service providers need someone to provide services to, and social products need more than one participant to make it worthwhile. Everything on the internet is in some shape or form a network.

During the wildest decades of internet capitalism, web2 has been able to completely reshape distribution of most goods and services. Instead of watching TV all day long like our parents did, we now interact with the unbundled online equivalent. Spotify, Youtube, Netflix, Linkedin, Tinder, Twitter, Google–these are all byproducts of The Great Unbundling. First they unbundled the once-powerful platform, and then they bundled the relevant content into unique points of integration, targeting specific consumer interests.

In theory, we think of network effects and IMU as two separate concepts. Take a marketplace as an example for network effects. If it doesn’t serve product suggestions to buyers based on their purchase history, there is no IMU to buyers. But the network effect is still there, because there is a group of transacting buyers and sellers. Paradoxically, a new user can passively benefit from the activity of others, while the positive experience of network effects need not follow the user's engagement with the product linearly. An example for this is Bitcoin–a first time user has the full benefit of the existing network security, and the transaction fees paid in comparison to block rewards for miners are so negligible (low single digit %) that they don’t provide IMU in the form of increased security.

That being said, it’s almost impossible to ignore the positive feedback loop between IMU and network effects. If a platform provides IMU to network participants, it strengthens the network, and the network strengthens the IMU. So in practice, network effects and IMU often go hand in hand because, as we mentioned above, everything online is a form of network. IMU is created through competition for counterparties, their attention and/or their purchasing power–and competition does not exist in isolation.

In web2, the IMU can be stronger on one side of the relationship–the example of Airbnb, where hosts live and die for positive reviews, whereas guests don’t have nearly as much of a switching loss. And yet guests are the indirect beneficiaries of IMU too, because they benefit from competition among hosts. IMU is the tide that lifts all boats. There are also examples where IMU is direct on both sides–content creators on many platforms have data analytics tools that allow them to manage their messaging to reach broader audiences and monetize better, whereas content consumers are served with suggestions that match (more or less) their taste.

The open network nature of blockchains and products built on top of them exemplify the idea of “come for the tool, stay for the network”. We hope the idea of web3 can take the relationship between networks and IMU even further. Absence of gatekeepers can lead to greater user experiences, where unique products are not marginalized as features. There should be increasingly more examples of IMU benefiting all product users directly, as opposed to indirectly, as we’ve shown above. We discuss these ideas in this article.

IMU in our digital la-la land

At this point in time there are very few crypto applications to which we could point and argue they solved a real world problem. Self-custody, sovereignty and permissionless access are obviously important, but that’s not the point of our argument. These are more like features of the industry, and not of a specific product.

The idea of IMU simply does not apply to DeFi core primitives. Every DEX swap, money market loan and leveraged trade are all zero-sum interactions by definition. If a user would get a discount on every incremental AMM trade, a liquidity provider (and/or application’s treasury) would be accumulating an incremental loss; also, this doesn’t accrue advantage to a single product, because it can be replicated by any competitor.

DeFi summer of 2020 and liquidity mining was not a good example of IMU. If a user becomes a liquidity provider (LP), he is passive by definition. Yes, LPs earn more every time there is an interaction with the product, but the earnings are not increasing marginally (the fee is fixed, and earnings are volume-dependent). And, most importantly, the interaction is required on behalf of the active trader, not the passive LP.

The LP has no incentive to continuously use the product, and the active trader doesn’t experience IMU. The concept of impermanent loss kills IMU, because users are constantly at risk. Even if there was a marginal improvement, it could get degraded by marginal loss.

That being said, we could see IMU at DeFi infrastructure/tooling level. This would mean ML-enabled risk management and prediction products, and algorithmic decision making which improves upon more data being fed. This can also benefit users who contribute (data) to these open products.

At its core, we think IMU is best enabled via content and monetization–it could be social, media or commerce. This is mostly what we discuss in the following sections. We like these categories because they have high user retention which makes them perfect for designing products with consistent and improving engagement loops.

Marketplaces (e.g. Opensea) in web3 are not as prone towards IMU as their web2 counterparts. Thanks to smart contracts, sellers no longer need to rely on buyers’ reviews to establish transactional trust. The financial aspects of a transaction are enforced in a credibly neutral way, whereas the nature of digital assets removes buyer’s concerns regarding the product quality. It turns out the social aspect–which is lacking in web3 marketplaces–is a feature, not a bug. Creators need to be able to communicate to collectors, either directly or via curation medium.

In the following section we show some project examples that we think have a clear direction toward creating IMU for their users.

Lens is a protocol representing an on-chain social graph, on top of which various social applications can be built. Most notable so far is Lenster, best understood as Twitter of web3.

The social graph can be a vague term, but you can think of it as the entirety of your interactions across all web2 social platforms. We’ve written about it before. Lens’ social graph is constructed through NFT ownership. There are different NFTs users own that represent their profiles, publications/content, follows, etc. On aggregate, they represent the user's social graph, because all on-chain social interactions can be traced to a specific NFT.

Let’s unpack a few ideas how Lens users can use this graph to experience IMU.

The problems of content creators in web2 are well known–risk of deplatforming and loss of followers, hefty take rate charged by the underlying platform, limitations around monetisation models and distribution of content. Web3 social and Lens thus bring the most immediate benefits to Content Creators, and the ideological objective is to tackle all of those issues.

The history of social products tells us how the nth user is more valuable and adds more value to other users than the 2nd or 3rd user did thanks to compounding network effects. Every interaction with the web3 social graph by the creator can expand their follower list in a valuable way; every interaction can make the existing followers more engaged. If web3 social graphs are immutable and truly owned, while the monetisation model is more lucrative compared to web2 alternatives (more earnings kept by the creator), that alone becomes an interesting proposition to keep engaging with this graph.

However, we think there is one underappreciated aspect in web3 creator monetisation. This new system is competing against the web2 that unbundled the traditional media channels, as we discussed in earlier sections. The web3 creators now have to unbundle the already unbundled products even further, and make the content of an individual creator more appealing than the aggregate of alternatives.

Example would be an individual video producer on Lens competing for users’ attention against the entirety of content available on Youtube. This paradigm will favor the few who stay relevant on their own; the others will have to rely on curators (i.e. bundlers–the new aggregators), whose role in web3 will be more important than people anticipate. Web3 social will again prove the great truth of business–there are “only two ways to make money in business: One is to bundle; the other is unbundle.

The bottom line is that a higher share of earnings to the creator alone does not guarantee lucracy–distribution and visibility does. The ability to get discovered early and easily is what propelled some of the most viral web2 products (TikTok), but the web3 curator incentives and user’s ownership of social graph will prevent it from plateauing and undoing itself.

IMU for Fans on Lens will result from their interactions with their favorite creators and brands, because their loyalty, continuous engagement and super-fan status can be verified on-chain. Creators will select, engage and reward specific fans for their actions, which creates a better incentive model for both parties in this interaction to continue their relationship on-chain, rather than off-chain.

Regular Users will be able to look into their social graphs and understand their content preferences, professional and personal networks better. Lens allows more choice for the average user across the entire social bundle stack. This stack includes a platform where content is displayed, moderation of it, algo / discovery preferences, etc.

Of course, users already “have” web2 social graphs–but they cannot separate them from a platform like Twitter, and they don’t own them in a true sense, the openness is limited, and the graphs are fragmented per each platform. It’s not possible to use any single web2 social graph in a multi-dimensional or unified sense, where it’d be used to access products and services of different providers. If a web3 social graph can be used like this, then every interaction brings IMU–the graph becomes richer.

We think this is where content indexing, labeling and preferences become important. To help with this growing need, Curators will work with creators to deliver their content in a way that optimizes for both the need for discoverability (aggregate reach) and preferences expressed through users’ social graphs (relevant reach). The majority of web2 and web3 creators cannot stay relevant entirely on their own–without curators, most of them are simply screaming into the void. The importance of curators grows non linearly as the amount of content increases, and the underlying platform matures.

The successful curators will be those that can bring the most relevant content to the most relevant group of people, and do it dynamically. The incentive for them will be the curation earnings (can be monetized on both ends of the spectrum–creator and consumer) and their own social graph (followers). We think web3 curation barriers will be lower. No longer you need to work at Spotify to curate music–instead, anyone can help distribute content and be rewarded for it through Lens’ mirroring functionality. This allows far more authentic social discovery which we already started trending towards over the last few years thanks to Substack and the likes.

We think there will be a big loop of increasing marginal utility on Lens, and multiple platform participants will be a part of it. People joke about “Twitter on blockchain”, and who’s gonna use it. We agree–posting, tweeting, commenting–these are all commoditized functions. However, our belief is that real adoption will start with creators and their content being brought over. They will be the real pull for regular users and content consumers towards web3 social we presently fantasize about on web2 mediums.

Finally, social graphs could be used to establish credibility, and develop into a scoring system that allows people to engage in DeFi with some trust assumptions. For example, having a reputable social metric on Lens might allow people to take out under collateralized loans from Aave. The more someone cultivates their Lens social graph, the higher chances of getting more favorable lending terms. We also see Sismo being part of this picture, which allows verifiable credentials through zero-knowledge attestations. Verifiable credentials will be important to protect parts of the social graph which are deemed sensitive by the user.

Our bottom line is we think web3 social is the ultimate medium to create IMU on both sides of the fan-creator interaction. Web2 platforms are strong at creating IMU for creators, but they’re not perfect in respect to fan engagement. That’s where web3 can do more and better.

Tribz is building a reputation system for web3 devs that can attest to their open-source contributions through soulbound NFTs. They reflect different skills a developer has, and indicate the depth of contribution.

Once there is a market / platform to showcase these contributions at scale and for them to be recognized, there is a clear reason for the developer to continue engaging with Tribz. This reputation system could be used for hiring, accessing gated content and building social credibility. Verifiable demonstration of knowledge and skills–or IMU–accumulates at no cost. is a platform for displaying and shilling NFTs with social features layered on top. What’s the value of a collectible, if it cannot be shown? also allows curation of NFTs to be done on behalf of others. If a curation proposal is accepted, the curator gets paid by the NFT collector.

If curators do well, they increase the visibility and outreach of their work–especially if it includes highly valuable NFTs. This can help them get noticed more frequently and establish credibility, and credibility should lead to more curator earnings. This is similar to the curator role we discussed in the Lens section, but not identical–the NFT collector is unlikely to monetize, so the distribution and reach is not as relevant. Instead, the curator has to optimize for the aesthetics.

From the perspective of an NFT collector, it would be no different from paying up for prestigious and/or luxurious services in pursuit of excellence in real life. That being said, we can see where IMU accrues to NFT collectors too. If you have a highly valuable collection that can help curators get noticed, then such a collection should attract many of them, and carry strong incentives to provide the best curation to the NFT collector.

Sepana allows applications to build dedicated web3 search engines for their end-users. This infrastructure should enhance content discoverability, and improve the search quality through customized algorithms. This idea goes beyond the search for financialized data–Sepana has already provided custom search engines for social platforms like Lens and Mirror.

An example of the custom Sepana search engine for

Users use web2 search daily, but they face ads or the generic search results that are forced onto particular users based on some attributes (e.g. geographic location). Users are also served content based on criteria into which they have no visibility into. The end result is either poor user experience or having to learn advanced search commands to filter out the irrelevant results.

The reason why we think Sepana plays into the IMU thesis is because it opens up customizable search algorithms based on the user’s social graph. If you can tailor your search experience based on your preferences–using the search labels you find useful, ignoring the labels which are distracting–you’re able to build a personalized vertical search experience.

This process is distinctly different from both web2 vertical search (Google, Meta and others who determine your information discovery process) and horizontal search (context-independent information like financial transaction history). Every interaction with web3 vertical search brings incremental data points which people can use to make the next search more accurate, more relevant, and perhaps even less so if that’s the objective.

To help people start searching and provide a signal in the noise– functions like filtering and ranking–search algorithms can be contributed by outside users (third parties). This is where search can become explosive. Not everyone will tailor their own search algo–a broad selection of public algos will prove to be useful. The content owners will incentivize algo builders to ease the discovery process. Unlike in web2 search, the end-user will not be stuck with a single, black-box-like search algo. Web3 search engines have the potential to disassemble TikTok the same way TikTok is disassembling Google; algo builders have the potential to eclipse the creators in importance and demand.

Web3 Dating x Fashion x Curation

While observing two acute problems of our industry–single developers and embarrassing outfits–we came up with a product idea that truly showcases our IMU ideas on multiple dimensions. This is the web3 intersection of fashion, curation and dating!

Developers want to meet women, but first they need to level up their fashion game. Brands will provide their inventory for selection, but a newbie dev will likely need advice from experienced clothing curators first. After all, you can dress expensively, but no one else besides your mother will admire you! Once there is a girl who likes dev’s looks and wants to date, devs can proceed to the next step. A purchase of the outfit from brands is then required, following which brands get paid and compensate the curators, and the dev goes on a date. Through these steps, dev is building his social and commerce graphs.

Who is experiencing IMU here?

  • Developers. They get instantaneous feedback from the market (ladies) on their outfit. The more they interact with this product, the better they become at picking slick clothes and (hopefully) dating girls. The social graphs of their dating history elevate their status to super devs, and appeal to young aspiring devs.
  • Brands. They access the open commerce graph of devs, and can see what pieces of clothing sell best. Brands can target the super devs who prove themselves by dating successfully, or they can offer different inventory to those devs who are down on their luck. Point being, the more you sell, the more info you get which can improve product distribution strategy.
  • Curators. If you prove that you turned a shy dev into a dating star, imagine how many more of them will want to hear your fashion advice. Being a successful web3 curator means building a name for yourself which everyone can easily verify.

IMU meets tokens

The token models for IMU-creating products is still a niche and underexplored topic. There are no perfect answers, but we think we can point to some principles around which tokens should be used to reinforce IMU loops. After all, a token is not the business model–just a pull for certain participants in the system and a direction towards specific action. Not every web3 application can create IMU, and our following thoughts on tokens apply to those that do.

To start with, the burden to actively engage with the token should not be actively placed on the user who is the largest beneficiary of IMU. Take the example of Spotify. Every week you are presented with the fresh Discover Weekly, which is based on your musical taste. You get to pick new songs to your playlist, and some people absolutely love it. Now, would these same people still love it and pay for the service, if at the end of each week they were required to write a review on Trustpilot?

To illustrate this point in web3, we wanted to find an example where the token function actually disincentivizes IMU. However, after giving the most fee-generating applications a quick glance, it becomes clear that the vast majority of these projects either a) don’t have a token, or b) don’t actively use a token in a relevant manner for the product. It is common practice to design tokens in a way for them to be functional as a byproduct of the application being used. Yet these designs are not optimal for building strong IMU loops–they optimize for the preservation of token price, and not for the utility to the end-user or enhancement of the product.

That being said, we want to spend some time discussing dYdX which we think did a good job with bootstrapping its perpetual futures exchange by allocating some tokens to both sides of the market–makers and takers. If you’re a taker, the incremental fees you paid on dYdX increase your allocation of dYdX rewards relative to other takers. If you’re a maker, the incremental volumes you provided within allowed market depth and spread parameters increase your allocation of dYdX rewards relative to other makers. Unlike with AMMs, where provision of liquidity is passive, the order book model can be better used to reward specific (active) participation.

Based on the above, every taker carries an incremental incentive for makers to provide better quotes. The incentive is not only the typical maker spread earned, but also dYdX token rewards. The more makers compete for this, the better pricing is offered to takers–they bring even more volumes, and the IMU cycle starts again. This design improves the market functioning and liquidity on dYdX.

Even in today’s bear market dYdX is doing 500M-1B in daily volumes, which is several orders of magnitude larger compared to the next biggest on-chain competitor (who’s doing 100M-300M in daily volumes).

dYdX daily and cumulative volumes

One of the problems dYdX was unable to solve in the past–admittedly it’s hard–is combining the IMU with token utility. As a result, there was no good reason to hold dYdX in the past.

In this regard, dYdX's decision to build their own Cosmos chain becomes very interesting. The selling pressure from today’s token sellers (IMU beneficiaries, i.e. makers and takers) will be at least to me extent balanced by demand from dYdX chain validators and stakers, whereas token itself could have value accrual in the form of fees and/or MEV capture. Meanwhile, the IMU cycle for makers and takers will be preserved.

Another token design principle is using it as an incentive for builders of IMU-enabling product functionality. Founders should identify the primary IMU beneficiaries of their application, and what features are needed to get it running. For example, web3 social needs content curation to help creators with distribution; web3 search needs custom search algorithms for users to plug into and discover the content.

This is not to say that these parties should all be incentivized in perpetuity–after all, IMU should be the tide that lifts all boats and eliminates the never-ending token incentives. What we mean is that a token could be the answer to the cold start problem IMU products have in the traditional web2 product landscape. Surely this “solution” has been applied in web3 far and wide, but we haven’t yet seen examples of organic IMU that would far outweigh the importance of a token.

Next, we think if the product can prove PMF without a token, then it should also be able to sustain itself without one, even if / when a token is introduced at a later point. This does not apply to the governance and decentralization aspect–only to the core functionality and how the user flow is impacted by the token.

As we’ve already shown, IMU in web3 can be bidirectional. This implies that the application's native token should not be the main component of incentive design over a long period of time. If applications treat tokens as a marketing budget (which is what they so far proven themselves to be), then they better think what to do when this budget is depleted.

Lastly, for tokens to be recognized as valuable incentives, they should be desired, and the surest way of getting there is attaching tangible value. We’ve all learned that governance power alone does not warrant economic premium for a token. TradFi has a perfect illustration of how much governance power is worth in one of the most valuable companies in the world, where voting power is concentrated in the hands of co-founders. This is not much different from many crypto projects, where founders either own the governance outright or can exercise power over other actors to effectively seize the unchecked power.

Alphabet’s (Google parent co) two stock classes–one with voting rights, the other without

Tokens need not require a direct monetary accrual. Instead, a user who builds IMU-enabling features can be incentivized with an application's token that unlocks extra product features for whoever is using the product.

The end goal, in our view, is for the IMU apps to become appchains and/or superapps, which would go a long way in addressing the “valueless governance token” meme and the Tragedy of Commons. We will address token topics in more detail in part 2 of this article.

Our marginal takeaways

We hope this article and concept of IMU can inspire better crypto product and engagement ideas. After all, the sentiment is at the bottom. People are looking for clues in CPI prints and FTX bankruptcy documents. We urge you to cease these activities immediately.

Instead, ask yourself how your product can improve for your users over time. To us, the learning process we went through here has reaffirmed that adoption hope is not lost in vain.

Disclaimer: Zee Prime has invested in Sismo,, Tribz and Sepana