Writing

Marc Andreessen famously said:

While this does provide an intuition as to whether you have product/market fit (PMF), it implies that PMF is binary. In my experience, PMF is not binary, it is a gradient. You can have very intense PMF, a lot of PMF, some PMF or you can have zero PMF.

These thoughts are centered around getting new products from 0 to 1. In my opinion, they apply to almost every kind of business, except for consumer social.

The closest literature I have found that matches my thinking on this is from First Round Capital, which suggest there are levels, almost suggesting there are stages, or even a progression, to PMF. This still feels off to me; PMF feels like a heart rate - something with continuous measurement.

I'm going to go over how I think about PMF, using tactics and measures I've used to build multiple products that do more than $1m ARR over the course of my career. 

>> 🧮 If you just want to calculate your own PMF Score. Use this calculator.

Let me explain more.

You can have intense PMF (PMF Score: 0.99) one day, and over time, let it slide to zero PMF (PMF Score: 0.0). This is to say, you're not "good" once you achieve a sufficient level of PMF - it very easy to regress. 

Getting your PMF Score is not quite a science for me yet, but there are clear factors that you're moving up or down on the PMF Score scale. I do think that there are some products where there is an upper limit on your PMF score, either because of the product, or the market - so you may not ever be able to get to hyper-scale.

With that in mind, let's talk through my PMF hit list.

I've compiled these examples first hand over the years and can count on them to provide real signal when evaluating a product or investment. These factors apply to almost all types of businesses except for consumer social. That means it should not matter if you're building a hard-tech product (like Sonic Fire), or a B2B SaaS company (like PassNinja) - they will apply.

ICP willingness to talk is nice - Getting your ideal customer profile to talk to you is a very helpful signal, the further they are from you socially, the better. The higher the likelihood the better. Why? Your mom loves you. Your moms co-worker cares about her and by proxy cares about you. Some dude you met in a telegram group? Probably doesn't care about you at all, but does care about his own problems.

If I can convince at least 1 in every 10 strangers I reach out to directly to respond to me, I will assign this factor a value of 0.8. I will add 2 basis points for each additional person, per 10 strangers I reach out to - stopping at 0.99. Anything below 1 in 10, is 0.5.

I deliberately use the word stranger, because you should not know these people!

This is an easy way to start to understand if you will have PMF before having a product, but it can easily be a false positive if you're not discerning about your ICP. For instance, if you reach out to 10 BD or sales people, you will likely get 8+ of them responding to you emphatically - it's their job!

Pre-sales are top tier signal - Getting strangers you do not know personally, to pay you for your product before it even exists is typically the strongest indicator that you are onto something. It implies the problem you're aiming to solve is so pressing, that people will use their imagination and faith in the hope of solving their problem. The further you are from product, the better. So just an email describing the solution with no pictures is best, conversation second best, mockups third, etc etc

If I can pre-sell some product once, I'll assign this factor a value of 0.80. Each additional pre-sale adds 1 basis point to the factor up to 0.99. So 20 pre-sales would give us 0.99.

You can't use this factor in perpetuity, as once you have a product, it's moot. If this is the case, leave the factor out of the equation.

Support tickets are premium - Having people pay for pre-sales is a very strong signal, but sometimes there are externalities that are not accounted for in pre-sales. Things like recommendations from bosses to solve a problem (even if the subordinate doesn't think it's a problem) can lead to false positives on the pre-sale factor. As an additional factor, I like to include the number of support tickets that come in once the product is live. Having people open support tickets usually indicates that at a bare minimum, people are trying to use your product, couldn't and care enough about their problem to make it yours too. In the ideal case, they accomplished their goal, but have some feedback about how they wish it worked instead.

Now you don't need some fancy support ticket software like Zendesk for this, just having open lines of communication with users is good enough. SMS, WhatsApp, iMessage, Slack, etc are all valid for getting your support tickets from users.

For the first 25 product iterations or 12 weeks (whichever happens last): If I can get at least 1 support ticket for every 20 customers, that I'll assign this factor a value of 0.75. For each additional unique customer writing in per 20 customers, I'll add 5 basis points up to 0.99.

After the first 25 product iterations or 12 weeks (whichever happens last): The factor inverts. Meaning, ideally you get less than 1 support ticket for every 50 customers. If you're at 1 support ticket per 50 or under, you set the factor to 0.95. For each additional unique customer writing in per 50 customers, deduct 3 basis points - all the way down to 0.5.

This factor should change over time, to go down.

Growth rate speaks volumes - Paul Graham's essay on Growth should be a canonical text for all startup founders. It is an amazing composite of all of the above factors that you can measure weekly. The takeaway for us here is that you should be growing your core metric (which should be either revenue, or active users, in my opinion) at a minimum of 5% week over week, once you have launched your product. The cool thing here is that this is very easy to do at the outset, and becomes much harder as time goes on, so you don't get to play on easy mode once you get decent at this.

Looking at this as the average weekly growth rate for the last 4 weeks, I assign this factor a value of 0.4 for anything under 5%, 0.7 for 5-6% and 10 basis points for each break above 6% until 0.99. So an average weekly growth rate of 8.1% for the last month is 0.9 and a growth rate of 12% for the last month, 0.99.

Retention is the last factor - There are a hand full of ways to measure retention, but my favorite for evaluating PMF is to use Net Dollar Retention. There are some good resources for understanding the underlying methodology. Here are my two favorite resources: stripe + saas cfo. This retention number basically tells you if your customers find what you're doing for them valuable enough to spend more money with you over time. Hugely valuable for understanding PMF.

This factor takes time to crystallize, so in the early months (up to about month 4 of the product being live), I just leave it off the equation. Once I have some data, I score anything under 100% NDR as 0.5, 100-110% as 0.8 and then add 1 basis point for each 5% increase over 110% NDR up to 0.99. So 200% NDR would be 0.98 factor.

Depending on how far along your startup journey you are, you may not have sufficient data. When this is the case, you leave off what you don't have, but it's very important to be honest with yourself about why you're leaving it off and if you should even leave it off in the first place. Let's talk through a few examples.

If you are developing a treatment for cancer, and you're very early in the process - you should definitely be able to talk to customers, and pre-sales should certainly be on the table, depending on the progress you've made. If you have actually developed the treatment for cancer and are starting to commercialize it, growth rate would be a factor, and retention would be a factor. For obvious reasons, support tickets should be 0.

On the flip side, if you've developed the best video editing solution for marketing departments then you should be including all factors except for pre-sales when evaluating your PMF Score.

Now let's take a real example, I'm sharing metrics from a company I invested in with permission from the founder:

ICP Willingness 0.5 - based on his latest outreach to about 60 potential customers where only 2 of them responded.

Pre-sales 0.0 - based on the fact that the product is already live, we'll leave this out of the equation.

Support tickets 0.5 - They have been live for more than 12 weeks, customers are writing in about technical and billing support issues continuously.

Growth rate 0.4 - They grew an average of 3% week over week last month.

Retention 0.8 - They have been expanding users spend over time in the 100-110% range.

Note: I want to bring back that the idea that some companies may have an upper limit to their PMF. It could be a natural upper limit, due to the market, or a superficial upper limit based on product constraints - it's important that you figure out which upper limit you're hitting, if you do indeed think you're tapping on a ceiling.

Now, getting our PMF score is pretty easy, leave off pre-sales. The PMF score is calculated to be 0.08. Not great. This feels about right, the growth rate and retention are OK, but not great. The support tickets indicate there is some issue that is still unresolved for a lot of customers. ICP willingness tells me the message is not resonating. In it's current state, this company seems like it won't be a big business. 

I discussed this with the founder, and he told me he thinks that's about right. However, he thinks the constraint is actually because a lot of customers come to them for a product that doesn't quite exist yet, but is possible with some work. He says that roughly 90% of all signups are under the impression they can use them for this purpose, but it's not easy to do out of the box. So their PMF Score is actually being limited by the product, not the market, which is a good thing. This company's fate is not yet written, they can change their trajectory yet.

I do want to note that even though the company has a very low PMF Score, it is running with net profit each month. This is a mostly-separate thing from PMF, and very important to consider on its own.

What I'm getting at is that PMF can exacerbate poor unit economics. In a simple example, if it costs you $50 to get a customer that will only make you $4 over the entirety of their time with you as a customer, and you have strong PMF - you will go out of business.

This post is only on PMF, not on profitability, unit economics, etc - that is it's own set of very long blog posts. 

PMF is not binary, it's a gradient and your company exists somewhere on this gradient. You move on this PMF continuum with time, in either direction, closer or farther from true product/market fit.

>> 🧮 Remember there is a PMF calculator if you don't want to do a bunch of math yourself.

I'm curious to hear your thoughts on where I'm off in my thinking and hope someone finds this helpful!