Five-star reviews are marketing. One-star reviews are venting. The three-star review is the one a competitor's prospect should actually read — and it's the one most competitive research skips past on the way to somewhere louder.
Why three stars is the useful range
A five-star review is often written in a honeymoon period, or nudged by an in-app prompt right after a good support interaction — genuine, but not built to be diagnostic. A one-star review is usually written hot, right after a cancellation or an outage, and it tends to generalize: "terrible company," "avoid at all costs," the kind of language that feels satisfying to write and tells you almost nothing you can act on. The three-star review sits between the two, and the tone gives it away — someone who is still a customer, still finds the product useful enough to keep paying for it, and specific enough about what's wrong that you can tell exactly which feature or which interaction prompted the review. That specificity is the point. A reviewer awarding three stars is doing cost-benefit math in public, and the cost side of that math is a list of real, current gaps.
What to pull out of them
Four things are worth reading for, specifically:
- Recurring themes, not single complaints. One review calling onboarding confusing is an opinion. Ten reviews across eight months independently describing the same friction is a product gap.
- Feature gaps stated as workarounds. A line like "we ended up exporting to a spreadsheet to do this" is a customer describing, in detail, a feature that doesn't exist yet.
- Support and pricing friction. Renewal surprises, response-time complaints, a sudden mention of a price increase clustered in the same quarter — these are worth dating as well as reading.
- The customer's exact words. Not your paraphrase — the phrase itself. "Support used to be fast" lands differently coming from a rep repeating a real customer's language than from a generic claim about response times.
From review to battlecard row
Raw complaints aren't battlecard material until they're sorted into the two places a rep will actually look for them. A recurring, current gap becomes a line in "why we win" — but only if it's something your product handles better, not just something the competitor's product handles badly. Those are different claims, and conflating them is how a battlecard ends up making a promise your own product can't back up. A friction point a prospect is likely to raise unprompted — a pricing complaint, a support delay, a specific missing integration — becomes an objection-handling row instead: the prospect states or implies the concern, and the rep has a specific, sourced response ready rather than an improvised one. The full anatomy of what else belongs on a card like that is worth a separate look at how to build battlecards reps actually use.
Reviews are only one half of this picture. They describe what a competitor's existing customers are unhappy about — not why a prospect who evaluated both of you chose the other option. That's a related but different question, and it's what win-loss analysis is built to answer, from people who compared you directly instead of strangers on a review site.
Where this goes wrong
Reading reviews well means discounting some of what you read as much as extracting the rest.
Weight recency. A complaint about a missing feature from two years ago might describe a product that has since shipped exactly that feature — check before it goes into a card, not after a prospect corrects your rep on a call. Don't treat a handful of reviews as a trend, either: a couple of isolated mentions of slow support is an anecdote; the same complaint recurring steadily across a year, from reviewers who otherwise sound nothing alike, is a pattern. Watch for reviews that don't read like customers at all — a burst of five-star reviews with similar phrasing right after a funding announcement, or a wave of one-star reviews the same week a competitor launched something, are both worth discounting rather than citing. And before a specific complaint goes on a battlecard, check whether it's still true. A bug gets fixed, a gap gets closed; citing a specific past problem instead of an ongoing pattern is the fastest way to lose a rep's trust in the card the first time a prospect mentions the fix.
Canopy pulls new reviews from G2, Capterra, Trustpilot, and Product Hunt into the daily briefing as they post, dated and attributed to the source, so a recurring three-star complaint is something you can point to with a timestamp instead of something a rep half-remembers from a review they read once. The reading and the judgment about what's a pattern versus an outlier is still yours to make.