Most competitor analysis is a scrapbook

Ask ten teams for their competitor analysis and nine will send you a table. Logos down the left, features across the top, green ticks scattered around like confetti. It's tidy, it's colour-coded, and not a single decision comes out of it.

A feature table doesn't tell you how they win

Because they don't win on features. They win on acquisition: where the traffic comes from, what it costs them, which creative keeps running, what happens between the click and the money. None of that fits in a tick box, so it quietly gets left out — and what's left is a scrapbook.

What actually matters

If a competitor research exercise doesn't answer these, it hasn't started:

• Where their traffic comes from — channels, shares, and which ones are growing.
• Which creatives survive. A creative that's been running for forty days is running because it pays. Lifespan is the closest thing to a public profitability signal you'll get.
• The funnel — what happens between the ad and the payment, and where they lose people.
• Offers and pricing — the actual hook, not the pricing page.
• Who sends them traffic — affiliates, partners, the people hijacking their brand queries.
• Where they're thin — the channel they ignore, the geo they're weak in, the queries nobody owns.

Templates and frameworks aren't the problem

Every competitor analysis template you download is fine. Every competitor analysis framework — Porter, SWOT, whatever — is fine. The structure was never the bottleneck. The cells are empty, and filling them honestly is weeks of an analyst's life: digging through ad libraries, matching promo codes to operators, pulling backlinks, reading reviews nobody wants to read.

So the framework gets filled with vibes instead of data, and everyone nods at the meeting.

Tools give you slices, not the picture

Competitive analysis tools each own one slice. One shows ads. Another shows traffic. A third shows reviews. Each wants $99–300 a month, and competitive intelligence quietly turns into a subscription pile that still doesn't answer "so what do we do on Monday."

This is also why ai competitor analysis mostly disappoints: point a model at a competitor without giving it real data, and you get a confident essay about "focusing on differentiation."

How to do a competitor analysis that ends in a decision

1. Take one rival, not five. Depth beats breadth every time.
2. Write the decision first. What are you actually choosing between? If nothing, don't do the research.
3. Get the acquisition picture — channels, creatives, funnel, partners.
4. Name the gap out loud. One sentence: where they're weak and what you take.
5. Tie every claim to a source. If it can't be re-checked, it's an opinion in a data costume.

That's the whole job. It's also exactly what the competitor teardown does automatically — one domain in, the acquisition machine out, in hours instead of weeks. If you need that on repeat rather than once, it lives inside the research subscription.

Bottom line

Competitor analysis isn't a scrapbook of what they have. It's a map of how they get paid — and where that map is thin enough for you to walk through.