Your ARR Looks Great. VCs Say Check These 7 Red Flags.

Here's a fun way to ruin your afternoon. Take the revenue number you tell people about and ask one question: would I bet my own money that this number is still here in twelve months?

Most founders flinch. And that flinch is exactly what a panel of investors was poking at during TechCrunch Disrupt last week. They weren't there to hype anyone. They listed the warning signs they see when a startup says it has product-market fit but doesn't quite have it yet.

The list is short, specific, and useful even if you've never pitched an investor in your life. Let's walk through it.

Three people shared the red flags: Anu Bharadwaj, a partner at Iconiq; Rajeev Dham, a partner at Sapphire Ventures; and Rahul Vohra, the founder of Superhuman Mail. TechCrunch covered the session in a piece published October 23, 2026.

Two investors and one operator is a nice mix. Vohra built a product people obsess over, and he's spent years measuring why.

Their advice agrees more than it disagrees. Fit isn't a trophy you win once. You keep earning it.

Bharadwaj pointed at a habit that's spreading, especially among AI startups. A founder says "we have $1.2 million in ARR," but when you dig in, the figure is a run rate. Take one busy month, multiply by twelve, and call it annual.

That's not the same thing. Annual recurring revenue means money you can reasonably expect to keep collecting. A run rate built on a hot month only tells you what happens if that month never ends.

The worst version involves pilots. If a customer pays for a two-month trial and you annualize it, you've counted ten months of revenue that doesn't exist. Investors catch this fast.

A fix that takes ten minutes: write two lines in your own notes. One says "contracted recurring revenue." The other says "everything else." Report them separately.

Customer concentration means a handful of buyers make up most of your revenue. If three customers pay 70% of your bills, you don't have a market yet. You have three relationships. Lose one and your month looks very different.

Swap-ability is the quieter one. Ask yourself what happens if a competitor offers your customer a lower price. If the answer is "they'd probably switch," your retention depends on nothing but cost. The panel flagged this as a real risk, particularly in categories where products look alike.

Here's a quick self-test. Pull your customer list and sort by revenue. Look at the top five. For each one, write down why they'd stay if a cheaper option showed up tomorrow. If you can't write a real answer, that's your homework.

Bharadwaj said churn and retention numbers depend heavily on revenue quality. Translation: a low churn rate means little when your revenue came from customers who were never going to stay.

Then there's the pilot gap. Getting pilots is easy these days. People like to try new things, and AI tools made trying cheap. Turning a pilot into a production contract is the hard part, because that's when the buyer asks real questions about security, support, and who is going to be on the hook when something breaks.

So track pilots and production deals as two separate counts. If you ran twelve pilots and signed two production contracts, your conversion rate is 17%. That's a number worth knowing, even if it stings.

Dham's warning was blunt. Don't mistake a compelling vision for a good business. A big idea gets you the first meeting. It doesn't pay anyone's salary.

The seventh flag is more of a trap than a flag. Dham also mentioned a company at roughly $1 billion in revenue that still had to re-earn its fit. Markets move, competitors copy, and customers change what they need. Even giants don't get to coast.

If that's true at a billion dollars, it's true at your size. Fit is a condition you check on regularly, a bit like checking the oil in a car you drive every day.

Vohra shared the survey method behind Superhuman's rise, which he first laid out in a First Round Review piece. The core question is simple: how would you feel if you could no longer use this product? The choices are very disappointed, somewhat disappointed, or not disappointed.

Sean Ellis, who first popularized the survey, found that startups with about 40% of users answering "very disappointed" tended to grow. Superhuman started at 22%. By splitting responses by customer type and focusing on the people who loved the product, it moved to 33%, and then to 58% after roughly three quarters of focused work.

At Disrupt, Vohra added a harsher number. If only around 15% or fewer of your users say "very disappointed," he suggested thinking seriously about a pivot, reusing the code and the team to build something else. At 20%, he said you can optimize your way toward 40%.

You don't need a big sample. The First Round piece notes that roughly 40 respondents can be directionally right. Survey people who used the product at least twice in the last two weeks, ask the four questions, and you'll have a real answer by the weekend.

This part gets skipped in most summaries, and it's the useful part. Superhuman ignored the "not disappointed" crowd entirely. It studied the "very disappointed" group to learn who its ideal customer was, then built for them.

For the "somewhat disappointed" group, it only listened to people who cared about the product's main benefit, speed. Those folks were close, and something small was holding them back. Fixing that something turned them into fans.

Roughly half the roadmap went toward doubling down on what fans loved. The other half removed blockers for the almost-fans. Not a bad split to steal.

Vohra also described an internal metric called "delighted ARR," which counts only revenue from highly engaged customers. It's a nice trick. A single number that says how much of your revenue comes from people who'd miss you.

Good news. Bharadwaj said that at the seed stage she isn't looking for product-market fit yet. She wants a plausible thesis and a founder who adapts as conditions change. Dham added that he likes founders who've pivoted several times, since each pivot means a hard decision was made.

So early on, your job isn't to look perfect. It's to learn quickly and show your reasoning. Paul Graham's classic essay on doing things that don't scale makes the same point from another angle: recruit your first users by hand, talk to them directly, and treat that contact as the best product research you'll ever get.

If you're organizing those early learnings, you can use a spreadsheet, Notion, or a planning tool like Foundra to keep your customer assumptions and evidence side by side. The tool matters less than writing it down.

A simple weekly habit: keep a page titled "What I believed, what I learned." Two columns, dated rows.

Try this plan before your next investor chat or board update.

Monday: separate recurring revenue from pilots, one-time work, and run-rate guesses. Put each in its own column.

Tuesday: list your top five customers by revenue and write why each would stay at a lower competitor price.

Wednesday: send the disappointment survey to 40 or more active users. Keep it to the four questions.

Thursday: sort answers by who is "very disappointed" and look for patterns in role, company size, and use case.

Friday: pick one blocker to remove for the almost-fans and one strength to double down on. Write the numbers down and schedule the re-survey for next quarter.

None of that needs a data team.

Report recurring revenue separately from run rate and pilots. Watch concentration, swap-ability, churn, and the pilot-to-production gap. Don't confuse a great vision with a working business.

Measure fit with the "how would you feel" question, aim for the 40% benchmark, and read answers by customer type. At seed stage, show your learning and your ability to change course, because that's what early investors say they want to see.

And remember, fit is something you keep earning.

What are the biggest product-market fit red flags investors mention? At Disrupt, the panel named unclear revenue metrics, inflated ARR from short pilots, customer concentration, easily replaceable customers, weak churn and retention, a gap between pilots and production, and confusing a vision with a business.

What is the 40% product-market fit rule? It comes from Sean Ellis's survey of nearly 100 startups. Companies where about 40% of users said they'd be very disappointed to lose the product tended to see strong growth.

How many people do I need to survey? The First Round Review piece says roughly 40 respondents can give directionally correct results. Survey people who used your product at least twice in the past two weeks.

What did Rahul Vohra say about pivoting? He said that if 15% or fewer of users would be very disappointed to lose the product, you should consider a full pivot. At 20%, he said you can optimize toward 40%.

Do seed-stage startups need product-market fit? Anu Bharadwaj said she doesn't look for it at seed. She looks for a plausible thesis and a founder who adapts. Rajeev Dham likes founders who've pivoted several times.

What is the difference between ARR and run rate? ARR is recurring revenue you can expect to keep collecting. Run rate multiplies a short period, such as one strong month, by twelve, which can overstate what's actually recurring.

1. VCs, Superhuman Mail founder share red flags about product-market fit, TechCrunch, October 23, 2026
2. How Superhuman Built an Engine to Find Product/Market Fit, First Round Review
3. Investors break down how founders can raise a Series A today, TechCrunch, October 22, 2026
4. Do Things That Don't Scale, Paul Graham