Things I'm thinking about
ditching minimum viable products
Happy Sunday, friends. Keeping it short this week because I am hunkered down on this magic summer month focusing on shipping Sublime. I’m anxious/nervous/excited and can’t wait to share more – soon.
Things I’m thinking about:
The prevailing dogma over the past decade is that startups should ship fast and iterate as quickly as possible. Ten years ago, Reid Hoffman said: “if you’re not embarrassed by the first version of your product, you’ve launched too late.” That was great advice in 2013, but a lot has changed in the last ten years. I’ve read a lot of thoughts on this topic from people like Scott Belsky, Sam Altman, and others, and I’m going to try to list some reasons why this is not necessarily great advice:
To create a product that grows organically, you need to “surprise and delight” your customers. You can’t do that by simply meeting a user’s expectations, you have to surpass them. And doing that takes time. (h/t Scott Belsky)
Markets with lots of competition require finding new, better ways to solve existing problems. In the short term, investing in that sort of innovation often doesn’t look like progress. It’s the classic Abraham Lincoln “give me six hours to chop down a tree and I will spend the first four sharpening the axe.”
Everyone wants to have a unique idea but the alternative is to just to be ten times better. The market is full of crap. An underrated way to succeed at anything is simply to be better. Anyone can build a "quick and dirty" tool to chat with your PDF. Few people can build incredibly polished products. And those that do need time, effort, and intention.
Most companies are not OpenAI but it’s still worth watching this snippet by Sam Altman. OpenAI raised $1B before they had a product and it took them 4.5 years to release it. There are a lot of other examples of companies that were not the first to market but still won because they made an existing idea better. Figma famously took 5 years before making their first dollar. Zoom and Rippling built for 2-3 years before releasing. This breaks every “rule” in the startup world around MVPs and being capital light pre-PMF. This is Figma’s revenue chart 🤯
When you’re an early stage startup, your most important job is to get clarity on what you’re building. When your product is buggy and low quality and the first mile of the product is ignored (the welcome/tour, the onboarding, the explanatory copy, the empty states), you won’t know if the poor retention numbers are due to a bad product or a bad market. A buggy or incomplete product obscures quality. (h/t Dani Grant)
Launching a product has more gravity than we assume. Once you’ve launched, you’ve dropped a heavy anchor. The natural tendency at this point is to incrementally improve the product and it becomes exponentially harder to explore new terrain. (h/t Scott Belsky)
The point is not that you shouldn’t ship fast. Or force fast feedback loops. Real learning occurs only when real customers are using a real product. And it usually takes several iterations to get the implementation of an idea to the point where it delivers customer value.
The point is: 1) There are no hard rules. You have to define your own rules based on the specific context of your market, your product, and your values. 2) No customers wants to use MVPs that the creators are embarrassed by. Instead, aspire to build products that are Simple, Lovable, Complete.
Resonated a lot!
Thanks for sharing great insights all the time!
This could have been helpful to read for Jasper. It's an AI startup that hit PMF early by shipping features for AI copywriting fast. Last year, this was a blue ocean - there were hardly any major products for this, little competition. But in the red ocean (very competitive market) of 2023, Jasper struggled. It was building a large-scale product for millions of users on a shaky foundation of lots of technical debt and product problems. With so many competitors, these quality issues stood out, and it was far too hard to compete with lots of free AI tools. Their interface was not high-quality enough, they didn't have a technical edge, and they had too many path dependencies to rethink the product from the ground up. Now they're having to lay off 35%+ of employees and are losing revenue and investors. This story seems like a validating example for ditching MVPs - at least in red oceans!