We found PMF. Here were the signs.

People make product-market-fit seem like this mythical thing you can pinpoint to an exact moment, but it’s never that. Different folks have different definitions of it, so I won’t bother trying to come up with a unanimous one. Paul Graham, for instance, says you’ve achieved PMF when you no longer ask yourself whether you have. IMHO, that’s a rather simplistic way of looking at it (it’s not actionable and just relies on vibes). 

“Companies without product-market fit feel like they're pushing a boulder uphill. Companies with product-market fit feel like they're guiding a boulder downhill.”

Instead, I believe it comes in stages. It’s not a milestone, it’s a continuum. That’s what it was like for us, at Command AI. First Round Capital puts it very eloquently in their framework (we’ll talk about this later): 

“Our contention is that product market fit is not binary — and it's not something that you get to overnight, despite what all of those “we just went viral” founding stories may have you believe.”

Why even measure PMF?

PMF is your start-up’s reality check. It won’t tell you what to build, but it serves as a guiding compass for your company’s direction. Without it, you’re flying blind. Assessing PMF helps you understand if your product resonates and people want what you’re building or if you’re merely filling up space in the market. 

What framework to use?

Superhuman famously published its PMF framework a few years ago. Super straightforward. You ask your users how disappointed they would be if they could no longer use the product. If over 40% say they would be very disappointed, you’d considered that a strong indicator of PMF. It’s similar to an NPS survey (which you can easily build with Command AI, *wink wink*). It also asks: 

• “What type of people do you think would most benefit from Superhuman?”
• “What is the main benefit you receive from Superhuman?”
• “How can we improve Superhuman for you?”

My take on this framework is that it’s good, but mostly applicable in the very early   stages of the company. It gives you just enough data to either (1) continue with a vision or (2) change and adapt based on user feedback. It also carries the risk that you end up whatever people are requesting, which as counterintuitive as it may sound, is not always the way to go (“faster horses”, etc etc). 

The one framework we’ve used a lot at Command AI is this one from First Round  Capital, called The 4 Levels of PMF. Todd Jackson, Partner at First Round, also recently shared it on Lenny’s Podcast. 

In this article, I’m going to talk about how we used First Round’s framework for each of the levels and what we’ve learned from implementing it.

To understand what follows, you’ll need to understand what Command AI does. Suppose you’re not yet familiar with our product (we’ll forgive you this time - just this once!!). In that case, Command AI is a user assistance platform that allows product teams to embed an intelligent agent on top of their software to help perform actions, fetch data, and co-browse with them to show how the product works.

It does more than responsive support, too. You can configure Command AI to proactively nudge users when they seem confused or encourage them to do things—like a friendly human assistant standing by your side, not an annoying barrage of popups.

We didn’t start as this, though. We talk more about our repositioning here, but our initial product was what we now call Spotlight, a universal search bar for any app. 

Stage 1 PMF

First Round’s framework describes this as: You have a handful of somewhat engaged and happy initial customers but things still feel early and messy. Focus on increasing satisfaction.

What it felt like

This was definitely us after our seed raise through our launch. We had some really exciting logos, but each deal we were closing felt very different. Each company thought we were something different:

• Some thought of us as a power user shortcut tool
• Some thought we were a dev-tool-specific alternative to onboarding
• Some thought we were a way to reduce support tickets

Our buyers were also very disparate. We had leaders from Product, Engineering, Growth/Marketing, Customer Support/Success. 

Note: All these deals were inbound! We did very, very little marketing at that point (a couple of blog articles here and there, like this one). It would’ve been easy to say “Oh look, this company has a ton of inbound from great companies–PMF! The market wants this!” but in reality, we were sending up a bat signal and getting responses from very different people. 

Satisfaction:  7/10 (the people who were using it loved it!)

Demand:  4/10 (inbound, yay! but very dispersed)

Efficiency: 0/10 (didn’t have a care in the world for efficiency at this point)

Stage 2 PMF

First Round definition: “You have more engaged, paying customers and less churn — you need to work on driving demand.”

What it felt like

We first felt this leading up to our Series A after our public launch. We were being bought by core Product teams to solve in-product challenges more and more. Increasingly we were being used for eliminating user friction problems versus power user use cases. 

Our biggest problem was reliably and predictably driving demand. Some weeks inbound would be great, some weeks it would be shit. It felt like every week we were shaking a magic eight ball. We did some marketing, but couldn’t figure out what was driving inbound. No prospects seemed to remember how they found us.

Satisfaction: 7/10 (I don’t think this actually changed between these two stages)

Demand: 6/10 (unreliable, but much more consistent)

Efficiency: 5/10 (inbound was working, but completely word-of-mouth driven and unpredictable)

Stage 3 PMF

First Round definition: “Momentum is picking up and you're finally feeling the "pull" of demand. It's time to focus on increasing efficiency.”

What it felt like

Things really accelerated when we positioned ourselves as an alternative to digital adoption products. Here were the signs:

• We started being able to reliably drive demand. Blog posts drove leads. Conferences drove leads. Lenny Podcast ads drove leads. And all the leads looked similar like before.
• Our customers started to play back our messaging to us. “I found you on ___ and it seemed like you had a different solution to user onboarding that was more intent based.”People would use our own words, like “we want to take a user assistance approach.” What?? You know that word? 🥹
• The first 10 minutes of every Gong recording of a discovery call have a very similar feel. It doesn’t mean they’re all identical. We still sell to a variety of personas (Support, Product, Growth, Marketing) and cover a variety of use cases (ticket deflection, activation, engagement) but I can tell you in 2 minutes which path it’s going down. When it does, I can tell you what the key questions are going to be and what objections the prospect will raise.

On the last point  I think there’s a subtle point here around what ICP really means. We Before our positioning change, we were reaching product, growth, marketing, and CX people. After our positioning change, we were still reaching product, growth, marketing, and CX people. BUT, we were reaching people in those departments who (a) had immediate pain that we could solve and (b) knew something about what we were offering. This made a HUGE difference. The repeatable point here I think is that you shouldn’t stop at “we’re reaching folks with the right title”. You need to go deeper and say “are these the best folks with this title that we could meet”.

Lots of people talk about PMF as a driver of growth, but we’ve felt an equally bigger change in customer usage, particularly from Level 2 to Level 3. In Level 2, customer engagement spiked in the beginning when they were really excited to execute on a specific use case, and then would typically wane. It was hard for us to get it back when this happened. We’d throw feature releases at people, come up with new use cases, but it felt like we’d have to give a customer 100 reasons to re-engage for them to pick up on one. That changed when we started selling against existing categories. Suddenly, we were reaching the “right” buyers and usage of our product massively increased. Instead of pleading with our customers asking for monthly check-ins, our customers started asking us for weekly check-ins. Again, if you just looked at the titles of the people using our product before and after, you wouldn’t see a change. But the product usage data helped us feel great that we were finally, consistently reaching the right people.

Satisfaction: 9/10 (surprising movement here)

Demand: 8/10 (yay)

Efficiency: 6.5/10 (a lot of our marketing channels are still in the earlier stages; we’re not yet 100% confident which are going to be driving the most leads for us in a year)

What we’re doing to enter Stage 4 PMF

First Round definition: “You're repeatably and efficiently solving an urgent problem for a large number of customers who need your product.”

We feel we’re currently navigating from Stage 3 to Stage 4. I thought it’d be useful to highlight some of the things we’re doing to validate and move through this stage. 

1. We’re going deeper into one of our key ICPs: CS/CX leaders and entering the chatbot category. We’re at a point where we know we’ve made selling to our other two main personas (Product, Marketing/Growth) very efficient. Officially positioning ourselves into the chatbot category also makes a ton of sense and feels like a natural product expansion. We’ve already had customers use HelpHub/Copilot for those exact things. In fact, most of the customers who discover us through the other Command AI experiences (like nudges), end up quickly adopting our Assistance tools (Copilot). 
2. We’re focusing more on brand awareness and educating the market that the status quo of the digital adoption space doesn’t need to be as bleak as it is now. We still talk with loads of folks who say something like “oh man, I just signed with <loathsome competitor>, I wish I knew about y’all a month ago.” Nooooo. Nice problem to have, but we want to get ahead of as many of these conversations asp osible.
3. We’ve been lucky enough to have some of the best customers come TO us and not need to do a lot of marketing and outbound until Stage 3. Stage 4 will be taking what we know about our customers and their needs, and amplifying our efforts to get in front of the right people. 

Plus, we gotta keep up with this crazy hockey stick growth on our blog. Congrats, you are now part of the solution.

Conclusion

My one piece of advice is: don’t dwell on the mechanics of PMF too much. Instead, focus on doing as much as possible, learning as much as possible from what you’re doing, and talking to your customers. Too many times I see founders overoptimize for the idea of PMF and spend too much time wondering if “they got it.” 

My more tactical pieces of advice are:

• PMF is more about product love and usage than sales. Stick with PG’s “make 100 people love you not 1000 people like you” in the early days, instead of trying to grow revenue before you’re really loved.
• When going after your ICP, don’t stop at title. Push deeper into pain points, company structure, what other tools the company uses, etc.
• “We have lots of inbound” is not automatically great unless the inbound is clustered around your ICP.
• Choose a framework (First Round’s great), stick with it, and have regular check-ins to see where you’re at. 
• Lastly, have a founder friend audit your PMF. As helpful as this framework is, it’s easier to use in hindsight. If PMF is one of those IYKYK things, then find someone who knows to tell you when you have it.