Playbook · Creative signals
Your user reviews are the best ad brief you already have.
The short version
- The best ad copy is rarely invented. It is quoted. Your App Store and Play Store reviews already hold the exact words that make a hook land.
- Most teams write ad scripts from a brand deck. The people you are trying to reach do not talk like a brand deck.
- In our own five-week creative test, the ad built from real review language scored 0.76 on Google's ABCD framework against 0.49 for a script-first version, winning on Connection and Attention.
- The move: mine reviews for the recurring pain, the turning point, and the phrase users repeat, then build the hook from that.
Where ad scripts usually come from
Ask most teams where an ad script starts and you get some mix of the same three things: the brand deck, the positioning doc, and a folder of competitor ads worth copying. All three are useful. None of them is how an actual customer describes the problem out loud.
That is the quiet gap. The brief is written in the language of the company, then handed to a creative team, then filmed, then run. By the time it reaches a viewer it has passed through three rooms that all speak fluent marketing. The viewer speaks a different language, and they decide in about one second whether this ad is about them.
What a review actually gives you
A review is the one place your customer talks about your product with no brief, no incentive, and no house style. They tell you the moment the problem became unbearable. They tell you the objection they had before they trusted you. They tell you, in their own words, what changed after. And the store ranks the most useful ones to the top for free.
That is not feedback. That is a casting call for hooks. Every strong first line you could write is already sitting in there, phrased better than a copywriter would dare, because a real person wrote it about their real Tuesday.
reviews how customers really talk → the recurring phrase said three times, three ways → the hook quoted, not invented
The three things to pull from a review
1. The pain, in their words
Not the category ("productivity"), the specific complaint ("I kept losing the thought before I could type it"). That sentence is a hook. It works because the viewer has had that exact thought and has never seen it written down. Recognition is faster than persuasion.
2. The turning point
Good reviews have a before and an after, and a moment between them. "I almost deleted it, then I tried the one feature everyone mentioned." That structure is a thirty-second ad on its own. You are not writing a story, you are transcribing one.
3. The phrase they repeat
Read enough reviews and the same line keeps surfacing in different handwriting. That repetition is the signal. If three strangers reach for the same phrase to describe the relief, that is the line you say on camera, word for word.
Proof: what happened when we built from reviews
We put this to a test. Same app, same offer, same audience, two ways to build the video. One was a script-first version, the kind a single AI tool produces from a clean prompt. The other was built from real signals first, starting with the actual user reviews, then wrapping a testimonial around the language already in them.
Both were scored blind on ABCD, the Attention, Branding, Connection and Direction framework YouTube's own research ties to creative lift.
| How the ad was built | ABCD score | Won on |
|---|---|---|
| Built from real reviews | 0.76 | Attention, Connection |
| Script-first, single tool | 0.49 | Branding |
The review-built ad did not win because it looked better. It won on Connection, the letter that measures whether a person feels the ad is about them, and on Attention, the first three seconds. Both came directly from using words customers had already written.
The winning ad was not better made. It was better informed. It knew what real users said before a single frame existed.
How to do this without reading 4,000 reviews
The manual version works and is worth doing once, so you feel the language yourself. After that it is a clustering job: group reviews by theme, rank each theme by how often it comes up, and flag the verbatim phrases that repeat across strangers. What falls out is a short, ranked list of pains and relief-lines, which is a creative brief that no one had to invent.
This is the part we treat as a signal inside the pipeline. Reviews feed the concept the same way competitor hooks and category patterns do, so the opening line is chosen from evidence rather than taste. The brief stops being a document someone writes and becomes something you read off your own customers.
Frequently asked questions
Where do the best ad hooks come from?
They are usually quoted, not invented. The strongest hooks come from the way real customers describe the problem in their own words, most often in reviews, support tickets and app store feedback. A hook lands when a viewer recognizes their own frustration in the first line.
How many reviews do you need to build an ad brief?
Not thousands. A useful brief tends to appear once the same pain and the same relief phrase start to repeat, often within 50 to 100 reviews for a focused product. Frequency matters more than total count.
Do review-based ads really perform better?
In our five-week test, the ad built from real review language scored 0.76 on the ABCD framework against 0.49 for a script-first version of the same app, winning on Connection and Attention. One test is a signal, not a law, but language people already use reads as a person while language a brand wishes they used reads as an ad.
What if my product has mostly short or low-quality reviews?
Short reviews still carry the pain word and the trigger moment, which is most of what a hook needs. If store reviews are thin, widen the pool to support chats, cancellation reasons, and comments under competitor ads. The method is the same: find the recurring phrase and build the opening line from it.
Methodology note: the 0.76 vs 0.49 figures come from an internal head-to-head run over five weeks on a single voice-to-text app, scored blind on the ABCD creative framework. The client is anonymized and the competing approach is described by category. A single test is a signal, not a law. We publish the method so you can weigh the result for yourself.
Turn your reviews into your first market signal
Add Orcool to Claude, ChatGPT, or Cursor. Paste one URL, sign in with a magic link, and your own reviews, competitor hooks and category patterns come back as a tool call. Free for two weeks, no card.
Connect the MCP