Stelo Labs: Three Reasons for Our Failure That May Provide Some Insights for Other Startups

Original Title: Sunsetting Stelo

Original Author: Stelo

Compiled by: Qianwen, ChainCatcher

The Web3 security company Stelo, which was previously backed by a16z, terminated all products on October 31. According to the RootData page, Stelo completed a $6 million seed round financing on February 16 this year.

The company announced on its official website in October, explaining three reasons for the product failure. Below is the full announcement released by the official team, compiled by ChainCatcher:

We made a difficult decision to sunset all Stelo products on October 31, 2023. This includes the Stelo extension, stelo.com, approvals.xyz, and the Stelo API. This decision was very hard, but we want to ensure the safety of all users and protect them from future phishing attacks. To that end, we recommend that all users immediately download one of the following products: Pocket Universe, Fire, or Wallet Guard.

When we started developing Stelo, our vision was to tackle the technical and design challenges of security and usability. We believed the most pressing issue was preventing users from being scammed or falling into phishing traps. Therefore, we decided to focus on ensuring transaction security and making it easy for users to understand.

By becoming an integral part of the transaction flow, we believed we could evolve into a product ecosystem that includes MPC wallets, fiat on/off ramps, API products leveraging the scam/phishing data we collected, DAO voting tools, and more.

Numerically, we achieved the following results:

• The Stelo extension has over 10,000 downloads.
• Daily active users reached 1,500, with monthly active users at 7,000 (despite not developing or marketing products in the past 6 months).
• The assets we protected totaled over $200 million.
• Our platform blocked hundreds of scams, saving users from potential disasters.

In addition to our extension, we also developed a developer API for wallets, integrating our transaction risk engine, token approval site, and a universal interface for interacting with any smart contract.

However, after a year of building the Stelo product line, we clearly recognized that some of our assumptions were not valid or did not apply to the early stage of a startup.

Assumption 1: Transaction Security Has Data Network Effects

Our viewpoint was based on the "data network effect," which means that the more users we have, the more attacks we see, and the stronger our system becomes at detecting and blocking attacks, leading to more users. We particularly hoped for off-chain network effects, where data about websites and URLs is crucial. Only products that are part of the transaction flow can access this data and see where malicious transactions originate. It turns out that simply looking at the transaction payload and applying simple heuristic algorithms can detect most malicious transactions.

We recognized that when there is evidence showing otherwise, we should not just trust our own narrative . We believe there are network effects, but we should not confuse potential long-term effects with short-term realities. This network effect may still play out to some extent (almost all fraud networks have this effect), but we have not reached that scale yet.

We also recognized that competition is prepared for losers. In a market with no entry barriers, no strong network effects, and filled with competitors, no one is a winner. We need to devise effective strategies to avoid competition and identify the market.

Assumption 2: Every Cryptocurrency User Uses Independent Wallets

We adopted cryptocurrency adoption metrics because we believed all transactions needed to be secured. This assumption implies that all transactions would go through dedicated wallets. However, many of the most promising cryptocurrency use cases ------ gaming, decentralized social, and stablecoins ------ are likely to have embedded wallets that abstract the complexity and risks of using dedicated wallets away from users. If users do not directly use wallets, they will not face phishing risks, and our product suite will not add any value.

We learned to clarify our assumptions and question them when new evidence arises. We believe our approach has "lowered" the risk of the maximum use case form for cryptocurrency. While we know we are betting on the competition between independent wallets and embedded wallets, we did not truly focus on this issue to find its underlying implications. If we had really thought about which solution developers would choose for each product, we might have taken a different approach or chosen different problems to solve.

Assumption 3: Consumer Adoption of Cryptocurrency is Imminent

Most importantly, we believed that mass consumer usage of cryptocurrency was just around the corner. We are still excited about many cool projects, but cryptocurrency is not yet mainstream. AAA-grade crypto games or decentralized social platforms may soon lead to explosive growth, but the Stelo product suite does not offer much benefit to these users nor can it profit from this growth. Importantly, we do not believe our current users can represent the new wave of users we expect to come with broader adoption. Without feedback from the "right users," we feel like we are flying in the dark, without direction.

We recognized that we need to focus on significant end-to-end issues. On the surface, we are concerned with the problem------we protect people from scams. However, empirically, our user base primarily engages in gambling and speculative activities, which means that if we truly think about the end-to-end user issue, it is about helping people gamble and speculate, which is not the problem we want to solve. We do not need to solve the entire problem from the start, but we should understand the entire user journey and ensure we help users address important and ongoing issues.

Moving Forward

We will restart building and engaging with users. As we develop the next generation of Stelo Labs, we will keep these lessons in mind. We will be bold in questioning our own narrative , focusing on finding the right market, listing various assumptions, and grounding ourselves in end-to-end user issues.