Very important talk. Not sure when the phrase was popularized, but my introduction to it was in Marc Andreessen's blog post [1]. This entire series of posts in his blog are worth reading, by the way. Marc calls this "The only thing that matters", and it is definitely true.<p>I like this video because of the emphasis on the role our psychology plays in the search for product-market fit. Understanding people/users well - what they <i>mean</i> when they say "this is interesting" is really important. Systematic de-biasing of our own mind is an important idea in this talk.<p>Richard Feynman's famous conclusion to his report on the shuttle Challenger accident said "Nature cannot be fooled"-similarly, markets/users cannot be fooled. Either a product or service solves their problem or it doesn't. Anything in between needs to questioned carefully and observed (in an unbiased way) closely. Feynman also says "The first principle is that you must not fool yourself – and you are the easiest person to fool". Same idea as Peter's -- and a very important one!<p>I wish more founders talked about "what product-market fit feels like" (once they found it) - because only with enough samples, new startup founders can better understand whether they have achieved pm-fit. However, I wish video shouldn't be mistaken for thinking: "my Show HN post tanked, so I don't have pm-fit". I think it is worthwhile trying to understand what he means. It is quite clear when you <i>really</i> solve a problem that someone else has. It is also a great feeling.<p>One of the reasons why finding pm-fit is hard is because there are too many variables. Peter simplifies it to some extent when he says that while searching, one must be clear if you are changing the market or the product. This is correct, but in reality, these adjustments are quite subtle. So paying close attention to these constant/rapid adjustments (stepping back and thinking about them) is quite important.<p>I like the conclusion of his talk: you will know when you find it.<p>[1] <a href="http://pmarchive.com/guide_to_startups_part4.html" rel="nofollow">http://pmarchive.com/guide_to_startups_part4.html</a>
I found segment.io in the early days of a new job and quickly set it up. I started playing with it and siphoning off some live data using a pretty rough shell script calling curl. Peter and another colleague saw what I was doing and came to my company to meet in person. I recall booking the largest, most impressive conference room we had and a 10AM meeting.<p>Peter and his colleague showed up and walked me through using Segment. The told me that I was the only one using their API via curl and that they were impressed with what I had been able to cobble together so quickly.<p>Peter took out a tiny Moleskine-type notebook and drew a diagram of what was next on the roadmap for Segment. I recall how neat and detailed the diagram was and how thoughtful Peter was about what he was doing next.<p>Peter and his colleague also revealed to me a sort of sister company they were familiar with that was creating an API around professional identity as a service. This sounded a lot like <a href="https://www.fullcontact.com/" rel="nofollow">https://www.fullcontact.com/</a> but was something else.<p>I've since moved on from focusing on that type of data. HN seems to feature a fair amount of content about Segment. I enjoy keeping up with their progress and wish them all well.
<a href="https://news.ycombinator.com/item?id=4912076" rel="nofollow">https://news.ycombinator.com/item?id=4912076</a> is the show hn for anyone interested. It's funny reading that post I would not have realized that product market fit has been found.<p>In fact, there's been so many show hn's that have been way way way way more successful and I have seen nothing from them.
Besides it being a great and no-BS talk which I very much recommend, I think the title is a bit misleading.<p>The talk does a good job explaining how to <i>identify</i> PMF (i.e. how to tell fake PMF apart from real PMF), but unfortunately doesn't do much explaining on how to <i>find</i> it.
He keeps warning not to write thousands of lines before talking to customers (as you often hear from biased survivors). But it sounds like they didn't find PM fit by talking to customers. Rather they identified -- perhaps out of desperation and exhaustion -- 0.1% of the code they had written that they thought might be useful and threw it out there. The takeaway really seems to be write as much code as you can, if one-tenth of one percent is going to be the golden nugget. Also, I suspect that they were able to use that other code and the processes around it to capitalize on their good fortune.
This is a great and really honest talk. I really enjoyed it. In particular the description of how one risks spending too much time building something that no one really wants but everyone is kind enough to tell you is a good idea.
This is one of the best talks of the startup school for early stage companies, Peter could not be more direct and clear to define what exactly is a product market fit. I hope this is something which is consistent with all successful companies.
Most important takeaway for me was "fail faster". He talks about the first two attempts going on for a year and a half and burning through 5/6th of their money before they hit product market fit.