The five most common mistakes founders make in trying to find Product Market Fit (PMF)
I spend time with the twenty or so seed founders in my portfolio every week searching for product market fit. My guiding document is this one. I believe that there is a lot of art in finding PMF, but also a lot of math. Almost all founders have a background that causes them to lean heavily on one of the metrics in my report card. This strength often becomes a weakness because their internal narrative is that if they just keep leaning into their strength, everything else will take care of itself. While a strength is super helpful, PMF doesn’t really care. You have to hit these metrics or you end up with a mediocre company. So if your strength is acquisition but your retention sucks, you are out of luck. As a teacher, I often leaned into student strengths as a way of giving them confidence in areas for development. The same is true with founders, but you have to work on the areas for development, because they are usually your blind spots.
Here are the five problems I see most often:
1 — Not being clear on the metrics for PMF.
Hopefully, my founders are pretty clear on the metrics they need to establish PMF. We focus all the time on the report card in the document above — magic, d7 and d30 retention, engagement, channels and organic growth. But I rarely find a later stage seed founder who knows the metrics for PMF.
2 — Not picking one metric at a time.
Founders typically react to every piece of new information they get. That thrashes them from worrying about the magic, to worrying about d7 retention to worrying about channels often in the same week. The thing about experimentation, which is the only way to find PMF other than luck, is that it benefits from focus. When you are trying a bunch of things every week to move a metric, you are totally immersed, you have better insights, and multiple failed experiments can point you in the right direction. When you thrash around, your brain isn’t focused enough on the user behavior and you don’t have enough data points to gain insights. You should focus on the most important metric until you move it. For example, let’s say you are at 5% d7 retention instead of 80%. You can’t build an epic company until you at least 10x that number. Almost certainly that means some experiments to reframe your magic. If you focus on that for two weeks, you are very unlikely to do that. Persistence is almost as important as focus here.
3- Not running five experiments per week.
I have to admit that not all of my founders achieve the cadence of five experiments per week. Since I stress experiments around discovering user desires rather than fulfilling them, this is almost never about the ability to write code. Rather it is an intentional or unintentional laziness around experimentation, often caused by the confidence that their gut instinct will show them the way without so much work. The scientific method is disconcerting, because we want to believe we ‘just know’. Sadly, that does not work because user behavior is far too complex to develop a good gut instinct even if you have been working with those users in other contexts for years.
4- Not putting a strong analytics and attribution system in place.
Now we are getting to the good stuff. I have teams that understand what PMF takes, have a focus metric, and are running five experiments per week, but they are failing because they don’t have the accounting systems which allow them to understand their users properly. If you can’t answer the question of where a user comes from (attribution) precisely, you are just guessing about what acquisition strategy is working. For example, if you have a bunch of tik tok videos but users watch them and then type your site name into google, you can’t differentiate between tik tok traffic and other organic traffic.
Likewise, if you aren’t using at least a basic analytics system like mixpanel, you often can’t answer questions like ‘what is your d7 retention?’ in less than a minute. When you can’t do that, and it takes work to find the answer, you are naturally going to use that data less, and use your gut instinct more. Go with the data and make it easy to do that well. This is the best investment an early stage startup who has traction can make for discovering PMF.
5 — Using user interviews as a way of deciding what to experiment with.
If you are detecting a theme here, good for you. Many of the mistakes here are laziness rather than lack of information. There are very good reasons to talk to users. They are exceptionally good when you are trying to understand ‘why’ users are doing something. A question like ‘Why did you abandon sign up on the last page?’ is totally gold. What users are not good for is telling you what experiments to run or what you should build. Users are stimulus-response machines. Based on the stimulus, their actions will be totally different. When you talk to them, they will come up with a story around why they felt some way, but most of the time, the issue is the exact situation you put them in. So mostly depend on the quantitative aggregate behavior numbers you collect, and spend time with users to get color on those. When you start with user interviews, you get biased by very small sample sizes of very articulate users justifying their behavior rationally.
I hope that spelling out these gotchas may help you avoid some of the pitfalls. I hope for a world where 50% of founders can find PMF instead of the 5% or so that I see in the market today.