Let me tell you about a founder I backed last year. Smart guy. MIT grad. Came to our angel group with a pitch deck that’d make a VC weep with joy. MRR growing 40% month-over-month. User sign-ups through the roof. Their new AI feature? 73% adoption rate.

Six months later, they’re dead.

Here’s what those beautiful metrics didn’t tell us: their customers were churning faster than they could replace them, nobody was actually using the product beyond day three, and that AI feature? People clicked it once, got confused, and never came back.

I’ve been on both sides of this table—as a founder, as an investor, as the poor bastard who had to tell the team we’re shutting down. And I’m sick of watching brilliant founders drive off a cliff while staring at metrics that lie straight to their faces.

Lie #1: MRR Growth Rate (Without the Full Story)

Every pitch deck I see these days has that hockey stick MRR chart. “We’re growing 35% month-over-month!” Sure, mate. But you started at $200 ARR, so congratulations—you made an extra seventy bucks.

What founders don’t show you is the denominator. When you’re doing $2K MRR, a couple of enterprise pilots can spike your growth rate to absurd levels. Lose one of those pilots next month? Your growth rate doesn’t just slow—it goes negative, fast.

I saw this firsthand with one of my own exits. We were crushing it, 50% monthly growth, until we hit about $40K MRR. Then reality set in. The easy customers were gone. Sales cycles got longer. Growth dropped to 10%, then 5%. We had to completely rebuild our go-to-market strategy because we’d been hypnotized by a metric that only works when you’re tiny.

Here’s what actually matters: cohort retention and revenue concentration. If your top three customers represent 60% of your revenue, you don’t have MRR growth—you’ve got a consulting business with extra steps. And if your three-month retention is below 70%, you’re filling a leaky bucket while doing a victory lap.

Lie #2: User Sign-Ups Without Activation

This one kills more startups than bad code ever will.

A Melbourne fintech I advise had 12,000 user sign-ups in their first six months. They were celebrating. Writing press releases. Posting on LinkedIn about their “explosive growth.” According to SmartCompany’s latest research, Australian fintech startups are particularly vulnerable to this exact vanity metric trap.

Know how many of those users completed their first transaction? 340. That’s a 2.8% activation rate. They were essentially running an expensive email collection service.

The problem is that sign-ups are easy. Any founder with a decent landing page and $5K in Google Ads can generate sign-ups. But activation—getting someone to experience your core value proposition—that’s where real product-market fit shows up.

I don’t even look at sign-up numbers anymore. Show me D1, D7, and D30 activation rates. Show me how many users complete the key action that makes your product valuable. If you’ve got 10,000 sign-ups but only 200 activated users, you don’t have a growth problem—you’ve got a product problem.

Lie #3: AI Feature Usage Without Retention

Oh boy, this is my favourite one right now. Every SaaS product on earth added an AI feature in the past 18 months, and they’re all patting themselves on the back for “high adoption.”

A portfolio company added an AI-powered insights feature last quarter. 81% of users tried it in their first week. The founder was ecstatic. “See? People love AI! We’re going to raise a Series A on this!”

Except nobody came back. Week two retention on that feature? 12%. By week four? 3%.

Users clicked it because it was new and shiny and had “AI” in the name. But it didn’t actually solve a problem they had. It was a feature looking for a use case, and TechCrunch has documented dozens of startups making this exact mistake throughout 2025.

This is the most dangerous lie because AI features are expensive to build and maintain. You’re burning engineering time and infrastructure costs on something that gooses your usage stats for exactly one week before becoming a ghost town.

Real AI product-market fit looks different. It’s not 80% trying it once—it’s 30% using it every single day. It’s users complaining when it goes down. It’s organic word-of-mouth because it genuinely saves people time or makes them money.

What to Track Instead

Stop lying to yourself. Start tracking metrics that actually predict whether you’ll be here in two years:

• Net revenue retention: Are your existing customers spending more over time, or are you constantly replacing churned revenue?
• Activation rate: What percentage of sign-ups experience your core value within seven days?
• Feature retention: For any new feature, what’s the week-four retention rate among people who tried it?
• Payback period: How many months until you recover your customer acquisition cost?

These metrics are brutal. They don’t make pretty pitch decks. They won’t get you retweets. But they tell you the truth.

And in startups, the truth is the only thing that keeps you alive.

Look, I get it. Vanity metrics feel good. They’re easy to explain to your board. They make you feel like you’re winning. But feelings don’t pay the bills, and pretty charts don’t stop your burn rate.

Next time you’re about to celebrate a metric, ask yourself: if this number doubled tomorrow, would my business actually be healthier? If the answer isn’t an immediate yes, you’re tracking the wrong thing.

Stop measuring what makes you feel good. Start measuring what keeps you alive.