Nikita Bier's 3 analytics traps. I had all 3.

I spent Sunday learning growth from someone who has launched 15 apps. My own analytics couldn't count to 2,085.

Hey Builder 👋

This week I did something I should do more often: I read one expert properly instead of skimming twenty threads, then ran his playbook against my own product. This one is what he teaches, what it found in Careerkit, and how I'll use it.

This week:

Nikita Bier has a blunt answer for founders building social apps for adults: "adults have no friends." Careerkit isn't even social. Nobody shares their resume builder with friends. So I wanted to know what he teaches that doesn't depend on invites.

Some context on why he's worth listening to:

Ayush Garg collected his interviews and posts into an open playbook, every claim cited. On Sunday I read it and turned it into a Claude Code skill that audits Careerkit. Its first run, on Monday, found my analytics had missed more than half of my signups.

Below: five of his rules that carry over to a small product, what each one found in Careerkit, and what I'm doing about it.

He says founders who ask him about growth almost always fall into one of three traps:

Careerkit had all three:

All three fixes went live this week. The honest numbers land next Tuesday.

Check yours tonight: count signups in your analytics and in your database for the same 30 days. Then sign up once with every method you offer and watch which event fires.

He has a name for the one-off viral moment: "Spotify Wrapped syndrome. You might top the charts for a day but that is only going to create phantom validation." Real product-led growth, in his words, is "where the mechanics of the product are actually driving the acquisition."

For you: list everything your product produces in one session. Pick the one someone would screenshot, and make it one tap to share with a tagged link.

"If you can't demonstrate value in the first three seconds, it's over."

This is where I'm actually losing people:

He also says "there's crazy places that users drop off that you wouldn't ever anticipate," and his fix is to walk every step of the funnel yourself. I've queued 12 recordings of people who left. My plan is to watch them before I change a pixel.

At tbh, "the number one support message we received was can I pay to reveal who sent me polls?" Gas later sold a subscription that hinted at who sent the compliments.

I ranked about 200 written feedback comments on Careerkit from the last 60 days:

So this week I shipped the boring fix: the download menu now shows how many free downloads are left. Next I'll test whether templates, the thing people actually ask for, should be the paid part.

He also splits apps into two kinds: durable apps, and "paywall-ad arbitrage" apps that live on people forgetting to cancel. That's a good question to ask about your own churn.

"Execute at 100% for the thing you're trying to validate at that specific stage of the product development cycle. And then you can kind of half-ass the rest just so you can get 100% signal on that one part."

This is the rule I'm applying to my own week, not just to Careerkit's funnel. Three products, one founder, and too many things to do. That's founder life, and when everything feels urgent, nothing gets 100%.

So:

If you take one thing away: before you trust any growth number, count your signups twice, in your analytics and in your database. If they don't match, fix that first. Then pick the one thing you'll do at 100%.

Reply with your question. I answer one every week. And if you've ever compared your analytics signups with your database, tell me how big the gap was.

Keep shipping,

Nishant