"Could work" vs. "Weird if it didn't work"

A subtle, but important, mindset shift

I make B2B startups take off (more info HERE), grab time to see if you’re a fit (HERE).

---

We talked about this on YouTube + Spotify:

Most startup advice can work… but frequently doesn’t work.

This is fine if you are diversified (e.g., a VC). This is not fine if you are an undiversified founder. In our shoes, we aren’t optimizing for the chance to succeed, but instead for minimizing the chance of not-succeeding.¹

More simply: We don’t want to do things that “could work”; we want to do things that would be “weird if they didn’t work.”

Here’s how to visualize the difference:

We understand this intuitively in most parts of our lives:

• If the pilot comes on the intercom and says, “we could survive this landing”, you’re going to start screaming.

• If your pregnant spouse says, “this could be your baby”, you’re going to have a few follow-up questions.²

• If your engineering intern says, “I just YOLO-deployed my vibe code to the demo branch, so it definitely could work for that mission-critical demo you have in five minutes”, how are you going to feel?

Yet because we don’t understand the physics of startups, or maybe because we’re foie grassed with so much bad advice that we’re all nihilists now, we wind up wasting our lives doing things that could work but don’t, and slowly go crazy.

We can escape this fate by approaching each part of startups in a way that is weird not to work:

In the idea stage

The typical idea-stage startup comes to me with a very impressive team that’s done tens or hundreds of customer discovery interviews. They have their list of pain points and problems that they’ve uncovered from their research; they’ve force-ranked these pain points and designed a solution to these pain points and have an impressive market thesis.

This could work; they will find out when they start trying to sell! (Spoiler: It very rarely works.)

Occasionally, I get a different story: “I went in-person with potential customers, really got a firsthand view of their world, and developed a hypothesis of what they need from reality, not a whiteboard. Then I tested that hypothesis by pre-selling the solution to five people who I got in touch with cold.”

This is the weird not to work approach to the idea stage. (See: Where to find ideas)

In the “first customers / early traction” stage

I get messages about companies like Loom, Slack, or (ffs) ChatGPT: “They just released a 10x product and it went viral, so doesn’t that mean I should just put my head down, build what I think the world needs, and launch on Hacker News?”

Sure, fine, this approach could work. It has worked. But most likely, you’ll wind up looking at your website traffic, wondering why nobody’s converting. In other words: It’s not weird if this approach doesn’t work.

Instead, the approach I advocate forces you to have 5-10 sales conversations per week with potential customers, and use that as an engine to get early traction and your first case studies. This is the weird not to work approach to early traction. (See: Sales sprints)

In the “figuring out repeatability” stage

The typical early-stage startup comes to me with a bunch of different potential customers who could buy - people who have pain points and problems, people for whom their product would be objectively valuable.

There are two elements here that could work:

1. You could figure out how to make multiple different kinds of potential customers work

2. And, people who have pain points / problems / would get value from our solution could buy

BUT! Here again, you usually wind up hitting your head against the wall when things that can work, don’t. You wind up having nice sales conversations that don’t turn into revenue, you have a wildly complex product roadmap based on all the feedback you’re getting, and can’t figure out why it’s not working.

Instead, designing around who would be weird not to buy our product brings you to the PULL framework and case study model:

• When startups use the PULL framework, they figure out their ideal customer profile based on who is in a situation where they can’t not act - where their existing options are seriously limited or unworkable. This is the best way I’ve found to define your ideal client profile. (See: The PULL framework; the ICP)

• When startups use the Case Study model, they focus on repeating ONE customer case study (based on who pulls hardest). If you spread your attention across multiple segments and value props, it would be weird if you figured any single one out; if you focus narrowly, and ran after that with all your might, it would be weird if you didn’t figure it out.

So what?

This is a lens you can apply to just about everything. When YC says “do things that don’t scale”, I interpret that advice as: “Tackle everything in such a way that it would be weird if it didn’t succeed.”

This lens will change your approach to hiring, fundraising, priorities/goals, product, sales calls, GTM (see: “the toll booth”), who to take advice from (would it be weird if they were wrong?).

A good prompt to figure out what is “weird not to work”: How would I approach X if my life depended on it?

This concept could expand into one of those insufferable business books - “the catchy idea that could have been a blog post.” Please… never let me write a book like that.

1. Minimizing the chance of not-succeeding =/= minimizing the chance of failing; the latter is obviously desirable but isn’t enough. We don’t just want to not die, we want to thrive.

Also, this concept is, I think, related to the “not-not” concept from the book the Heart of Innovation, but I am not sure because not-not is confusing to me. If anyone understands it, please advise.

2. My wife: “Please clarify that this did not happen to you.”