How I reverse-engineered 3 competitors' growth strategies in 10 minutes (and what I found)
Every time I'm evaluating a new market, I used to spend half a day stitching together tabs — SimilarWeb, LinkedIn, Twitter, App Store reviews, G2, the Wayback Machine — trying to build a coherent picture of what a competitor was actually doing. Most of the time I'd end up with a messy doc full of half-observations and no clear takeaway.
Then I started being more systematic about it. Not because I had a fancy tool, but because I realized I was looking at the wrong signals.
Here's the framework I now use, followed by what I actually found when I ran it on three real products.
The problem with standard competitive analysis
Most competitive analysis frameworks ask you to fill in a grid: features, pricing, target audience, positioning. That's fine for a strategy deck, but it doesn't tell you how a company is growing right now.
What I care about is:
- Where is their traffic actually coming from?
- What content is working for them?
- Are they betting on SEO, paid, or community?
- What are their users complaining about — and is that a gap I can fill?
- What changed in the last 90 days?
These five signals are what I call the growth fingerprint. Every company has one. Most founders never look at it.
The 5-signal framework
Signal 1: Traffic source mix
Is a product SEO-dominant, paid-dominant, or community-driven? This shapes everything — their content strategy, their burn rate, their defensibility. A company doing 80% organic is playing a different game than one doing 60% direct (which usually means strong word-of-mouth or a viral loop somewhere).
Signal 2: Content velocity and topic clusters
How often are they publishing? What topics cluster around their highest-performing pages? This tells you where they think their audience lives and what problems they're optimizing for in search.
Signal 3: Review sentiment over time
G2, Capterra, App Store reviews — not just the star rating, but the pattern in recent reviews vs. older ones. A product that was loved 18 months ago but is getting mediocre reviews now has a retention problem. A product getting better reviews over time is iterating well.
Signal 4: Job postings
What a company is hiring for right now is one of the most underrated competitive signals. Heavy engineering hiring = platform investment. Sales-heavy hiring = moving upmarket or into enterprise. Designer surge = rebrand or product overhaul incoming.
Signal 5: Social and community activity
Not follower counts. Engagement quality. Are people in their community genuinely helping each other, or is it a ghost town with scheduled posts? Twitter/X replies, Reddit mentions, Discord activity — these signal whether they have real user love or just a marketing presence.
Running it on three real products
I'll walk through what I found when I applied this framework to Cursor, Linear, and Notion — three products at very different growth stages.
Cursor
Cursor's growth fingerprint is almost entirely word-of-mouth + developer community. Traffic is in the millions monthly, but if you look at the source mix, direct and referral dominate. They're not running aggressive paid campaigns. They're not publishing a blog at high velocity.
What they are doing: getting developers to talk. Almost every senior engineer I know has either tried it or been asked about it by a colleague. The reviews are overwhelmingly positive and skew heavily toward "this changed how I work." That's the signature of product-led growth in its purest form.
The gap I noticed: almost no content for teams or managers. All the surface area is individual developer productivity. If you're building something in this space for team workflows, that's the opening.
Linear
Linear's fingerprint is opinionated positioning + design community pull. Their traffic is meaningful but not enormous. What's interesting is where it comes from: a disproportionate amount is referral from design and product forums, Hacker News threads, and Twitter/X. Not from SEO-optimized blog posts about "project management software."
They write almost nothing. Their blog publishes maybe once a quarter. But when they do publish, it gets hundreds of comments because the content is actually good — it takes a position.
Job postings have been relatively flat, which suggests they're not in a scaling-at-all-costs mode. They're building deliberately.
Review sentiment: consistently strong on UX, occasionally frustrated on customization limits. That's intentional — they've traded flexibility for opinionation. The people who don't like it are not their target user.
Notion
Notion is the clearest example of a community-and-template flywheel. Their organic search footprint is massive, but a lot of it is driven by user-generated content — templates, tutorials, community posts — rather than first-party blog content.
The signal I found most interesting: review sentiment has actually softened in the last year or so. Not collapsed — but the "this is magical" energy from 2020-2021 has been replaced by more measured reviews. "It's good but it's slow" or "I miss how simple it used to be" appear more frequently. That's the complexity tax of becoming a platform.
If I were building in the all-in-one workspace category, I'd be paying close attention to that shift.
How I actually run this analysis
For each of these products, I used Analook to pull the initial data pass — it aggregates 15+ sources into a single 60-second report, which gives me the traffic fingerprint, content signals, and review sentiment in one place. That gets me to a working hypothesis fast. Then I go deeper manually on whatever signal looks most interesting.
The combination of "fast aggregate view" + "manual deep dive on one thing" is what makes this actually useful rather than surface-level. The tool handles the data gathering; I handle the interpretation.
What the patterns actually tell you
After running this framework on 50+ products across different categories, a few things become clear:
Moats show up in the traffic mix. Companies with 60%+ organic search have built something genuinely hard to displace on distribution. Companies with 60%+ paid are either in a high-LTV market or they're dependent on spend — which means they're vulnerable to anyone with better unit economics.
Review trajectory matters more than review score. A 4.2 that's improving is more interesting than a 4.6 that's declining. The trajectory tells you whether the product team is listening.
Silence in job postings is a signal too. A company that was hiring aggressively and has gone quiet either got to efficiency or hit a wall. Worth figuring out which.
Community is the leading indicator, not the lagging one. By the time a product's community goes cold, the growth has already stalled. The community slowdown comes first.
The 10-minute workflow
Here's what the actual session looks like:
- Run a quick Analook report — traffic sources, top content, review snapshot (~2 min)
- Scan their job board for the last 90 days (~2 min)
- Search Twitter/X and Reddit for recent organic mentions — not tagged posts, but people talking about them unprompted (~3 min)
- Check one or two recent reviews on G2 or Capterra for qualitative texture (~2 min)
- Write a one-paragraph hypothesis: this company is growing via X, their current gap is Y, the thing I'd watch for is Z (~1 min)
That last step — forcing yourself to write a hypothesis — is where the analysis actually happens. The data is just the input.
If you try this on a competitor in your space and find something interesting, I'd genuinely like to hear what you found. The patterns are different enough across categories that comparing notes is useful.
If you want to run the first step faster, analook.com is what I use — free to try.
Top comments (0)