Product Market Purgatory

A number of years ago, I joined an early stage startup. We had some funding, but not too much. We had some customers, but not too many. We went through a different variant on the product every few months to try to increase our traction. As we watched our runway slowly dwindle, we’d stress at late hours, fidget with numbers on the spreadsheet, and rearrange the office furniture in hopes of prompting inspiration.

We were in Product Market Purgatory. Product Market Purgatory is that long, awkward period of indeterminate length in which a company is searching for Product Market Fit, but doesn’t know exactly what it looks like yet.

When you’re in Product Market Purgatory, you feel like you’re productively working while at the same time find yourself making progress about 6 times slower than your brain thinks you should be. Both your thoughts and your words burst with your vision, but the details of your everyday actions look more like meticulous customer service and numbers tediously not matching up.

Product Market Purgatory, in short, is a state of moderate productivity that feels like it could go on forever, if only you had the runway. (In a larger company trying to make a new product, it often actually does go on forever, because you do have the runway.)

The Core Myth of Finding Product Market Fit

Product Market Purgatory happens because we subscribe to one critical myth about Product Market Fit. The myth is that the journey to Product Market Fit is supposed to happen like this:

1. Identify a market opportunity in the form of a problem a lot of people have
2. Come up with an idea for a solution to it
3. Ask some people if they think it’s a good idea, and adapt accordingly
4. Build a functioning Minimum Viable Product and test whether people buy into it, making tweaks as needed.
5. Now you either have Product Market Fit, or you don’t. If you don’t, pivot until you do.
6. Win. Or, run out of money and lose.

(In fact, a lot of startup guides tell you to do exactly this.)

But if we look closer, the problem with this model is that at almost every point you are operating based solely on guesses.

When you come up with a solution idea, that’s a guess. An educated one perhaps, but still a guess. When people tell you whether they think they’ll use it, those are guesses (and self-prediction is usually horrendously inaccurate). When your MVP starts returning data and you pivot, A/B test, or try other things, those new options are guesses.

In fact, the only point at which you don’t act off of a guess in this model is if the product successfully takes off, either because the initial MVP worked or a later iteration did. At that point, you can finally start optimizing based on successful data.

Which gets us to the corollary myth to the Core Myth: Finding Product Market Fit is a matter of trial and error.

But that’s not how it’s actually supposed to work. Floating a lot of guesses at a time just increases our uncertainty. Instead, we can strategize to decrease our uncertainty by grounding actions on incremental experimental research. Here’s how:

How Finding Product Market Fit Actually Works

Imagine you’re searching for something. Let’s say it’s not a needle in a haystack, because that’s too depressing, but maybe it’s a clearing in a wooded grove. This is a metaphor.

You could search for it by blindly walking up to different spots in the grove and hoping there’s a clearing there. If you get lucky, maybe you’ll find it on the first try, or the fourth. If you get there before you run out of daylight you’ll be great.

Or, you could make a systematic search of the grove. You could quarter it and walk back and forth over every inch. If you could keep it up forever you’d definitely find the clearing, but you can’t actually do that because it’s not literally practically possible to try every single thing. Before you have time, night will fall and you’ll get eaten by wolves, or you’ll die of old age, or at least you’ll have to go back to your day job.

Lastly, you can make an educated search. You can climb a tree to increase your visibility. Even though up is clearly not the direction you ultimately want to be going, it gives you valuable perspective. You can make a good guess about where to go, and then examine the space when you arrive and move next in the direction of sparser trees. You can look for patterns in mini-clearings to learn more about trees and clearings in general.

The Product Market landscape is a landscape, not a lottery. If something ends up gaining traction with customers, or fails to, there is always a why. The data behind that why is embedded inside the thoughts, feelings, and behaviors of the customer.

Specifically, we are looking for information like:

• How does the user currently experience the problem that this product is supposed to solve?
• Which components of that experience are most significant for them?
• How do they currently understand what’s possible, not possible, necessary, etc when operating in this domain?
• How does the user currently relate to the actions they are expected to take when using the product? What do those actions mean to them?
• When a user has a particular reaction, what’s going on underneath it?
• Where are the opportunities to create a “magic moment” where the user’s eyes light up from how valuable the product is?
• What narratives are necessary for both the product and the necessary user actions to make intuitive sense to the user?
• What needs to happen for the product to fit into their life in a streamlined way?

The more of that why, how, or what’s going on underneath data we get, and the faster we get it, the more we can navigate inexorably toward Product Market Fit.

Education is the only reliable route to Product Market Fit in a feasibly efficient amount of time.

High Bandwidth Learning

Because volume of learning is our key strategy, it is vital that we use techniques that give us as high a bandwidth of knowledge back from our users as possible.

Common early-stage techniques such as landing page testing, “would you use this” interviews, and overall MVP marketing are highly limiting because they return data in the form of booleans: did this thing work, or didn’t it?

Other techniques, such as more in-depth surveying, marketing analytics, heat mapping, and A/B testing, return data in the form of numbers and statistical distributions. This is still only about an order or two more complex. On the scale of mapping a landscape, the bandwidth is still much too low.

Analytical data is optimized for accuracy, not for richness.

We need a way to not just answer known unknowns, but to identify unknown unknowns.

In order to get the bandwidth necessary in order to navigate the customer landscape reliably and efficiently, we need to get vast quantities of qualitative data.

In other words, first we find the clearing through educated searching. Then we can go back and look for shorter, more efficient, and more reliable routes afterwards.

Qualitative Navigation

So how can we get qualitative user data? Here are some techniques:

Live Observation. Literally watch them use your product, live. In fact, if you haven’t already, watch them go through life doing what they would do without your product as well as what they would do with it. Pay very close attention to facial expressions and mood cues that imply something significant may be going on beneath the surface.

Simulate the Context. People behave really differently in different environments. Your user will approach your product differently at work vs at home. Identify the context they would realistically use it in, and test your product in that context or a simulation thereof.

Get Train of Thought. Ask the user for their stream of consciousness reaction as they go. Don’t interrupt their reactions to talk to them, ask questions, prompt actions, or take feedback. Save it for the end.

Ask Open-Ended Questions. The more specific the question, the narrower the bandwidth of response. The more open-ended and non-presumptive, the more the user is able to provide the critical data of how they internally organize their reactions to your product. Ask, “Tell me more about your reaction.” Get them to tell you a story — the highest-bandwidth form of human communication is narrative.

Actively Encourage Non-Courtesy. People really want to be nice to the person they’re talking to and pretend to care about that person’s life’s work. Emphasize to them that genuine reactions are the most valuable, even if they’re negative or indifferent.

Build a Minimum Unviable Product. Your product doesn’t have to be hooked up and working for someone to understand how it’s supposed to function in a live test. Build a simulation of it using written mockups, example text, voice acting, and whatever else you need, just so you can run testers through it one at a time.

Don’t Pivot, Stroll. The best way to learn more is to iterate more. Be flexible and responsive to data as it comes in. Since you can iterate 10x — 100x faster using Minimum Unviable Products (above), you’re no longer limited to a few major pivots, but can instead constantly adjust to each new learning in a meandering stroll.

Experiment with Users Constantly. No data of any bandwidth will come in without users. Experiment live with real users constantly. The two big experimental topics of Product Market Fit are “Is my product useful?” and “How will people discover and start using my product?” At any given time, there should be plenty to test toward both of these.

Investing in Bandwidth

That company that I used to work for eventually figured out their Product Market Fit by visiting users in their home and watching them work. Today, it is a very successful, fast-growing, series A business with no end in sight. Before they were able to do so, though, they had some very lean years, lost a lot of talent, and nearly went bankrupt twice.

These days, I work with hundreds of early stage startups and use qualitative testing to get them to Product Market Fit in a vastly shorter time. It’s a much more relaxing process because everyone understands the high level-strategy, and most companies will hit their fit with plenty of runway left.

In the process, I’ve learned that the single most important investment a company can make early on is in bandwidth of learning from users.

Concretely, that investment looks like doing the following:

• Taking the time to practice and learn good qualitative user testing skills
• Making sure the company tests with users regularly, about 3-5 people a week.
• Spending relationship, brand, and social capital on asking users the questions the company really needs to know early, even if it’s awkward, rather than maintaining a reserved and optimistic front for the sake of appearances, courtesy, or future sales chances.
• Building the skills and tools necessary to try a high volume of Minimum Unviable Product experiments, such as paper prototyping or a copy-and-paste design template set to make hundreds of mockups.

In my own spare time, I personally enjoy video games. In game logic, investing in bandwidth is like spending your first few skill points on the learning-related skills that will increase your longterm skill gain. It feels like a pain to start more slowly, but is undefeatable in the long run.

In pedagogical terms, it’s learning to learn.

Escaping Product Market Purgatory

Product Market Purgatory happens when, trapped by uncertainty, we wander semi-aimlessly in the forest and distract ourselves with the trees. Sometimes, it happens because the forest is vast and we have too may options. Other times, we’re afraid of admitting that we’re a little bit lost.

Building a company necessitates living in uncertainty. We all signed up to do that.

But, uncertainty doesn’t have to mean Purgatory.

We can take control of the uncertainty of creating a new product by navigating through qualitative learning. We can control the time it takes by accelerating our tests and increasing our feedback bandwidth.

This enables us systematically get closer and closer to the match we are looking for in a time- and resource-effective manner .In other words, it’s uncertainty on the ground, but certainty on the strategic- or meta-level.

The truth is that, in the program I teach, we encourage companies to not even touch code until they’ve narrowed down their Product Market Fit through a hundred or so Minimum Unviable Product prototypes. All this is doable, and it truly can be solved in a matter of weeks or months.

Ultimately, we’re all trying to solve a problem and make things better for someone. For me, I would be happy if founders never had to go through Product Market Purgatory again.