Improbable Founder: Why Web3 Games Are Prone to Failure

Author: Herman Narula, Co-founder of Improbable

Compiled by: GaryMa, Wu Says Blockchain

Note: Improbable, a gaming technology company, was officially founded in 2012 and received its first funding of approximately $20 million from a16z in March 2015. In July 2015, led by a16z, Improbable completed a second funding round of $30 million, with Temasek also as an investor. In May 2017, it raised another $502 million, with SoftBank as the lead investor. In July 2018, it secured another $50 million from NetEase. In April 2022, Improbable's new project M² raised approximately $150 million, led by a16z and SoftBank.

Herman Narula, co-founder and CEO of the gaming technology company Improbable, recently tweeted a summary of his experiences over more than a decade in the gaming industry, highlighting the pitfalls and lessons learned in assessing the success potential of future MMO/Kickstarter/web3 games, such as the importance of experienced leadership in driving game success, technology application, optimal capital investment windows, team crisis management capabilities, and community communication. The compilation is as follows:

1. Due to factors like budget and technology, online games are very difficult, and the failure rate is very high, so the experience is crucial. If you commit to AAA quality, make sure experienced leadership is in place. Without leadership, there is no game.

2. Speaking of technology. If the technology has not been publicly proven to work for real users and deliver high fidelity experiences, don't even think about it. Ask questions like "operations per second," per-user bandwidth, server costs, etc. If they don't have good answers, run away.

3. If people claim to have "millions of players," "seamless integration," etc., assume their performance will remain at the level they announced the project, rather than magically improving hundreds of times later. Then it may all be downhill from there.

4. Interestingly, in our new architecture, we can handle far more than 15,000 concurrent players, but we won't publicly say this because it hasn't been fully tested. Even 15,000 is with very complex load testing bots. So we only have absolute confidence in fewer cases.

5. Generally speaking, technology is now a commodity; people buy it from other companies instead of developing it themselves. They should use off-the-shelf engines, Unity or Unreal, and services like PlayFab.

6. Never bet on something launching in more than 1-2 years. A 3-4 year development cycle for an online game has too many chances for things to go wrong. Buy in closer to launch, as more things will have been proven by then.

7. Unless the team regularly conducts public tests to show you progress, never accept a multi-year technology roadmap with absurd visions. Without public testing, do not participate in pre-sales. Demand independent reviews of the technology. (In fact, we sometimes offer this as a service).

8. Large online games are very expensive. You can't just put a demo version from the game engine asset store online; you need to build and run complex infrastructure. Even the largest studios outsource these, like our MPG team. Therefore, it is very questionable to release a quality online game with a budget below $30 million, especially when making AAA quality commitments.

9. Ensure that the founding team cannot access personal funds for use in any form before the game is released. If they get rich before you, that's not good.

10. Continuous communication is essential. At least quarterly updates are important, preferably quarterly, along with significant milestone updates. Leadership should be findable on social media.

11. Mistakes are inevitable. Every game I've seen or been involved with over the past decade has had issues. The standard for judging a team is how they communicate and resolve problems. If they deny/avoid, that's not good.

12. In the context of a crazily crowded market in 2022, creating a new audience for a high-budget game is almost impossible. Buying some land or assets does not create meaningful buzz for hundreds of thousands of users. So they need a marketing plan or a huge brand. Before spending money on marketing, check the community's engagement level and talk to community members. It's a small thing but always worth doing. In one case, we found that all the founders overlooked this!

13. Finally, I personally would not buy land in any virtual world unless it has some capability to support high CCU (concurrent users) or some known business model or some unique distribution capability. Only buy land when you believe in it 100%.