Cluely's Fake Revenue Was the Feature, Not the Bug

On March 5, 2026, Cluely's CEO admitted on X that a $7 million revenue figure shared with TechCrunch was fabricated. The retraction was explicit: the number wasn't real.

This gets covered as a scandal. It's more useful as a diagnostic.

What the incentive structure produces

Cluely isn't an isolated incident. It's the end-state of a specific incentive structure: build narrative, generate press coverage, attract capital, then fill in the revenue. The metrics serve the story, not the other way around.

In a narrative-first funding environment, the metric that matters is the one that supports the pitch. Revenue figures, user counts, engagement rates — when these are reported by founders who know they're in a signaling game, the incentive to optimize the report rather than the underlying reality is structural. Cluely's CEO didn't fabricate revenue because he's uniquely dishonest. He did it because the environment rewards the story that $7M revenue tells.

The gap between internal and external models

I've mispriced freelance projects before. The failure pattern was specific: I believed my internal model of the situation more than was warranted. I estimated execution difficulty based on how it looked from outside, not from the actual complexity I discovered building it.

Fabricating revenue is the same failure pattern, deliberately applied. The founder knows the gap between the internal model (actual revenue) and the external model (reported revenue) and chooses to close it by updating the external model rather than the internal one.

This is more rational in an environment where the external model is what gets funded. The internal reality is unfunded. The external narrative is funded. Updating the external model while the internal reality catches up is just strategy. The environment rewards it.

What this means for any metric

The broader problem isn't that founders lie. It's that the metrics that matter in the VC funding cycle are the ones that can be narrativized, not the ones that describe underlying reality. Revenue can be claimed. User counts can be inflated. Engagement can be gamed.

The metrics that describe whether the product actually solves a real problem at a real margin are harder to narrativize and slower to move. They don't fit pitch decks. So the incentive is to focus on the metrics that do.

I build Ordia outside the VC system deliberately. Not because VC money is bad — it's useful for specific trajectories — but because the metrics I care about don't fit the VC timeline. Does the system work? Does it keep working? Does it solve the coordination problem it was designed to solve?

These are slow metrics. A fund with a 7-year return horizon needs faster ones. The result of building for the fund's timeline is building for the metrics the fund tracks, which are not always the same as the metrics that describe whether the product is real.

The AI-specific version

The AI funding environment amplified this. The premium on being "AI-native" compressed the normal evaluation cycle. Companies got funded on category membership before demonstrating that the category applied to them in a meaningful way.

Cluely's product — AI-assisted real-time coaching — is a real category. Whether Cluely's specific implementation generated $7M in revenue is a separate question. The funding environment said "AI-native company in real category" was sufficient to fund. The revenue figure was needed to close the gap between "interesting category" and "investable company" — and when the actual revenue didn't close that gap, the reported revenue did.

The deflation of the AI hype cycle runs this mechanism in reverse. As the novelty premium compresses, investors ask what the actual revenue is. Companies that inflated external models while waiting for internal ones to catch up are now in the gap.

What honest metrics look like

The useful question is not "what metrics should you report" but "what metrics would change your behavior."

Metrics that change behavior: does the product retain users month over month? Does solving the problem continue to matter after the initial sign-up? What's the churn and why?

Metrics that narrativize: total sign-ups, revenue during a promotional period, user counts that include trial accounts.

The difference isn't always visible from outside. It's visible from inside, to the founder, at the moment they're deciding which number to put in the pitch deck.

The structural conclusion

Cluely's CEO knew which number was real. The environment made the choice for him.

Building outside that environment is one way to stay accountable to the metrics that matter. Not because bootstrapped founders are more honest — but because the incentive structure to optimize the report over the reality doesn't exist in the same form. The only person who cares about the metrics is the founder. That's a useful constraint.

The scandal framing misses this. Cluely's problem isn't that its CEO made a bad decision. It's that the system he was operating in made the decision predictable.