Finding PMF when you have nothing

Most advice for finding Product Market Fit starts with an idea (“what if…”) and outlines a process to test and iterate. But what if you don’t even have an idea?

Part of what makes totally-from-scratch PMF difficult is that we’re having to solve a multi-variable equation, with our variables being 1) Customer problem, 2) Our solution, 3) Who the customer is, 4) How we go to market and make money.

Like a crossword or sudoku puzzle, we’re having to pencil in what we think the answer might be for one or two of these variables, play out the tape for how that would lead us to answer the other variables, and see if it fits. If it doesn’t, we will need to revisit some or all of our work, again and again, until we get it right.

So, while what I lay out below looks like it’s a step-by-step instruction, it’s really not. We’re skipping around, holding some ideas constant while we pursue others, then rewinding when we hit a dead end.

1. Pick a problem space and start learning everything you can about it.¹  

2. Sketch out some ideas for how you could solve that problem that would transform those feelings into joy, confidence, safety, intimacy, relief. Boil it down to a headline, 3 supporting bullet points, and one picture.²

3. Do a quick business model sanity check - if we built this, how would we make money? Would it be enough? Does it seem realistic?

4. Now, mock it up as a landing page and include some kind of action button people can tap / click if they are interested (“Join the waitlist”, “Contact us”, etc).³  

5. Create some Google ads, buying keywords drawn from the words you heard people say during your user interviews, and your headline as the ad copy.

6. Now do this again 2-4 more times, starting back at step 1, 2, or 3.

7. Run your ads, and see which ones get the most traction. If you have a winner, start interviewing people to understand why it’s a winner. If you don’t rewind and try again.

What we’re doing here is using the ad keywords to test if we’ve found a meaningful problem people have. We’re using our landing page conversion to test if people will buy into the promise of our solution. We’re using our opportunity analysis model to confirm the TAM and GTM pass the sniff test. All without having actually built any software.

The Workshop

This is a newsletter-only section where I share a half-baked idea in hopes that y’all who are smarter than me can work it out with me.

A few quick hits:

1. Wired’s Steven Levy perfectly captures my feelings regarding the recent turn towards Trump of rich, entitled Silicon Valley types like Marc Andreessen and Ben Horowitz. Who needs democracy when you know you’ll be safe, thanks to your money, power, and influence?

2. Crowdstrike, wow. Anyone who’s been an engineer has been there - the crappy spaghetti code where one seemingly innocent change somehow brings down something supposedly unrelated. The other thing is how system failures can cascade, causing other systems, that are technically perfectly functional, to fail because they weren’t built to handle upstream failures. See: Delta struggling more than other airlines.

3. OpenAI launched GPT4o-mini, which is both 60% cheaper than GPT3.5 and double-digit % better (depending on the benchmark). This is a big deal, reminiscent of Moore’s Law 30 years ago when computers were getting faster and cheaper every year.

4. 2024 Mid-year consumer trends report from Dan Frommer. Topics: the economy, the beauty industry, GLP-1, Zyn and cannabis, YouTube is even bigger than you think.

1  Pick one of the problems that really stood out to you, that you heard from multiple people and, when they talked about it, you could really feel the depth of emotion tied to it. We’re not looking for minor annoyance, we’re looking for resentment, despair, frustration, fear, longing, anxiety.

2  What we’re looking for is the “Aha” moment that will make someone tell their friend / family / coworker, “You gotta check this out.”

3  A lot of tools can do this, but make sure you can track your conversion rate somehow. # of clicks on that button / pageviews.