A bad idea is harmless. You see it’s bad and move on.

The dangerous idea is different. It looks brilliant, obvious, and urgently needed. But it has a fatal structural flaw hidden just beneath the surface.

These are tarpit ideas.

Like the La Brea Tar Pits, they look like a pristine watering hole from a distance. Up close, you find the remains of ten thousand founders who came before you.

They’re seductive for one reason: they come from genuine frustrations. The founder spots a daily friction point and thinks, why has no one fixed this yet?

Every semester, a group of engineering students decides to solve the “where’s the crowd tonight?” problem.

The pitch is always the same. Going to an empty bar sucks. What if a real-time heatmap showed you where everyone is going?

The problem is real. The tech is simple. The mortality rate is near 100%.

Here’s why.

• The cold start problem is fatal. The app needs users to be useful. But it needs to be useful to get users. A blank map on a Friday night and the app gets deleted forever.

• Venues are adversaries. When crowdsourcing fails, founders pivot to bars sharing door count data. No venue will do this. A nightclub sells status and the illusion of popularity. Broadcasting “we’re empty” is financial suicide. Never build a product that requires a partner to act against their own interests.

• It misreads human nature. People don’t want the coordinate with the most bodies. They want to go where their friends are. A heatmap is an engineering solution to a social problem.

Three patterns show up across almost every tarpit idea.

• High frequency, low pain. The problem annoys everyone, but adopting your solution takes more effort than living with the problem. Loyalty card consolidator apps are a good example. People hate carrying cards more than they want another app.

• Behavioral dictatorship. The product requires groups to abandon how they naturally interact. Dinner scheduling apps fail here. People prefer the chaos of a group chat over structured software.

• The two-sided mirage. Two distinct groups must adopt the platform simultaneously. Local skill-swapping networks never work. Matching exact needs, at the right time, in the same zip code, is nearly impossible to bootstrap.

The counterintuitive lesson: if an idea seems obvious, targets a massive market, and no successful company exists yet, that absence is almost never because you’re the first person smart enough to think of it.

It’s the market signaling the problem is structurally unsolvable.

The best startup ideas often look like bad ideas that happen to hide a brilliant secret.