Your competitors' three-star reviews are the most honest brief you'll get this year

The middle of the rating scale is where buyers explain exactly what they wanted and didn't get. Here's how to read it in thirty minutes and turn it into headlines, promises and objections worth answering.

Marina ComisiFounder, Paxton Studio

5 min read

Most online reviews are written by people who want to "brag or moan." That phrase comes from a 2009 study by Nan Hu, Jie Zhang and Paul Pavlou, published in Communications of the ACM. Looking at Amazon, they found ratings piling up at five stars, with a smaller bump at one star and very little in between. Between 72% and 78% of the ratings for books, DVDs and videos they studied were four stars or more.

The people in the middle are the interesting ones. A three-star reviewer bought the thing, used it, liked parts of it, and still took the time to explain what was missing. That's rare, and it's precious, because it's exactly the brief you'd want before writing a page: what buyers hoped for, in their words, and where the current options let them down.

And when the reviews are about your competitors, nobody's feelings get hurt and it costs nothing. You just need thirty minutes and a spreadsheet.

01Three stars is where people explain themselves

Five-star reviews are lovely to read and hard to use. "Great tool, the team loves it" tells you someone is happy, and very little about why they bought or what almost stopped them.

One-star reviews tend to be about something else: a delivery that went wrong, a billing dispute, a bad day. As Ali Noyan puts it on Medium, they mostly criticize without revealing much about the reader's actual problem.

Three-star reviews usually follow a shape you'll start to recognize after five minutes: I wanted this, I got part of it, and it would have been perfect if… That sentence holds the promise that attracted them, the gap they hit and the thing they'd still pay for. Three pieces of a positioning brief, written by a buyer, for free.

Joanna Wiebe of Copyhackers, who has taught review mining for years, makes the same call: focus on the two- to four-star reviews, because they're the most candid. I'd start at three and widen to two and four if a competitor doesn't have many.

02The thirty-minute method

Pick three to five competitors. The ones your buyers actually compare you with, which may not be the ones you'd name. If you're unsure, ask ChatGPT for the best options in your category and take the first few names it gives.

Find where their buyers leave reviews. For software, that's usually G2, Capterra or Trustpilot. For a local business, Google reviews. For a product, Amazon or the app stores. For services, Google and the directories your industry uses. Filter to three stars, and add two- and four-star reviews if there aren't enough.

Read for about twenty minutes, and only read recent ones. The last twelve to eighteen months is plenty. Older reviews describe a product that may not exist anymore.

Copy, don't summarize. Wiebe's instruction is to clip any "language that stops you," and the word that matters is clip. Paste the exact sentence, with its typos and odd phrasing, along with the competitor, the star rating and the date. The moment you rephrase it in your own words, you lose the part that makes it useful.

Stop when you start predicting. Somewhere around the fifteenth review, you'll catch yourself guessing the complaint before you reach it. That's the signal you've found the pattern. Spend the remaining ten minutes sorting.

03Sort every clip into four columns

Once you have twenty or thirty clips, put each one in this grid. The example row is invented to show the shape, so replace it with your own.

What they hoped forWhat fell short"It would be perfect if…"Their exact words
Set up reporting in an afternoonTook two weeks and a consultantIt came with templates for our industry"we bought it to save time and spent a month configuring it"

Each column feeds a different part of your marketing.

"What they hoped for" gives you headlines. It's the promise that made them buy, in the words they used to describe it. If the same hope shows up across three competitors, you're looking at the main reason people enter your category. That's what your first screen should talk about.

"What fell short" gives you your proof points, if you can back them. Where buyers hit a wall with others, show how you handle it, with something concrete: a setup time, a screenshot, a guarantee. Only use the ones you actually deliver. A gap you claim to fill and don't becomes your own three-star review a few months later.

"It would be perfect if…" gives you your roadmap and your offer. Some of these are features. Some are services, onboarding, or a different pricing model. Share this column with whoever builds the product.

"Their exact words" gives you your copy. When a phrase shows up twice from different people, it goes in a headline, an ad or the first line of a sales email, word for word.

One rule while you use all this: don't name the competitors or quote their reviews against them in public. The point is to understand what buyers want and say it better. Pointing at someone else's bad reviews tends to make you look smaller.

04Know what the method can't tell you

Reviews are a shortcut, and like every shortcut they skip a few things.

  • Reviewers aren't all buyers. The same 2009 study points out that people with moderate opinions are the least likely to write anything. Three-star reviews are rarer than the others, so give each one weight, and don't treat twenty of them as a survey.
  • Some reviews are fake. Wiebe says it plainly: the method needs judgment. Skip anything that reads like a press release or a revenge post.
  • Your competitors' customers aren't exactly yours. If you sell to a different size of company or a different budget, filter for reviewers who look like your buyers before drawing conclusions.
  • Reviews tell you what, rarely why. When a pattern looks important, confirm it with two or three conversations with your own customers before you rebuild a page around it.

Run it again every six months or so. Products change, competitors fix things, and new complaints appear.

One last detail from the research. Northwestern's Spiegel Research Center found in 2015 that purchase likelihood peaks for products rated between 4.2 and 4.5 stars and drops as ratings get closer to a perfect 5.0, which shoppers tend to find too good to be true. Buyers already trust the slightly imperfect middle. It's about time marketers read it too.

Marina Comisi

Founder of Paxton Studio. Strategy first, then the hands to make it happen, for startups and scale-ups.

If you'd like help turning thirty clips into a page that sounds like your buyers, say hi. We reply like humans.

Tell us where you're stuck.

We'll tell you honestly if we can help.

Marina replies personally. No sales team, no obligation.

New · Playbook No. 02New playbookThe Anti-Robot Prompt PackDownload it →
Write me