Customer reviews in ads: turn reviews into ads that sell

How do you use customer reviews in ads?

Using customer reviews in ads works best in three stages. First, mine the reviews for why people bought, what nearly stopped them and the exact phrases they used. Next, count how many different people said each thing, so that you lead with what's common rather than what's loudest. Then turn the top themes into ad angles written in the buyer's own words. Before any review appears in an ad, check that you're allowed to use it and that the ad gives a fair picture.

Reviews are the best advertising research most businesses already own. When a customer describes the exact moment your product helped, they've handed you a hook, a problem and a result in one line. A copywriter rarely matches that level of detail, and readers recognise it instantly.

This guide shows where to find customer words, how to mine them, the fairness rules for using reviews in marketing and how to turn themes into ads. You also get two prompts, a worked example based on a real brand's public reviews and a mining sheet you can reuse.

Where to find reviews and customer words

Gather more than you think you need, since twenty pieces is a start and a hundred is better. Mix happy and unhappy customers, because complaints reveal the doubts your ads must answer.

• Reviews on your own site, along with your product pages and any marketplace listings.
• Public review sites like Trustpilot or Google reviews, which you can read for themes, although you should check each site's terms before reusing any text.
• Support emails and chat logs, which are full of doubts and "does it...?" questions from people close to buying.
• Post-purchase surveys that ask "What nearly stopped you buying?" and "What would you tell a friend about it?".
• Comments on your social posts and ads, where questions are really objections raised in public.
• Sales call notes, if you sell by phone or video call.

Paste everything into one document exactly as written, and remove names, email addresses and order numbers before you give the pile to your AI.

How to mine reviews for buying reasons, doubts and phrases

Mining means sorting what people said into a few useful buckets and then counting how often each one comes up. Here's the method, step by step.

1. Tag each review: note what triggered the purchase, the reason they chose you, any doubt they mention and the result they describe, since one review can carry several tags.
2. Keep the exact words: copy the phrase behind each tag word for word, because rough, specific wording beats a tidy summary.
3. Count by people, not mentions: one customer who says "fast delivery" five times is still one vote, and counting people stops one long review from steering your ads.
4. Rank the themes: the buying reasons with the most people behind them are your best ad angles, while the most common doubts are what your ads and page must answer.
5. Note the odd ones: a theme that only three people mention in vivid words can still make a strong hook to test if you label it a small sample.
6. Split ad problems from product problems: "wrong size arrived" is an operations fix rather than an ad angle, so send those to whoever can fix them.

You are a customer research analyst. Mine the customer words below for ad research. Never invent or polish quotes.

What I sell, in one line: [product and buyer]
Customer words: [paste at least 20 reviews, emails or survey answers, one per line, numbered, with names removed]

Steps:
1. Tag each item with any of: trigger (what made them start looking), reason (why they chose us), doubt (what nearly stopped them or a worry they had), result (what changed), complaint (a problem with the product, delivery or service).
2. Group the tags into themes. For each theme, give:
   - a short plain name
   - how many DIFFERENT people mentioned it (by item number)
   - 2 to 3 exact phrases, with the item numbers
3. Rank the reasons and the doubts by number of people.
4. List complaints separately as "fix list, not ads".
5. Flag any phrase that makes a health, money or results claim we would need proof for.

Output: three tables (Reasons, Doubts, Fix list), then the top 3 themes worth testing in ads and why, in one line each.

Rules: only use what's in the text. If a theme rests on fewer than 3 people, mark it "small sample". Keep every quote exactly as written, typos and all.

Rules for using reviews in marketing fairly

Reviews persuade because people trust them, and that trust is exactly why rules exist. The guidance below comes from the US Federal Trade Commission, so it applies to US advertising, while other countries have their own rules that you should check wherever you advertise. This is general information rather than legal advice.

What the FTC's endorsement guidance says

The FTC's Endorsement Guides Q&A, revised in 2023, covers reviews and testimonials used in ads. In plain words, it sets out five principles:

• A review must be an honest opinion from someone who really used the product.
• Unusual results need context: if you feature a customer with an exceptional result, the ad must say what people can generally expect, unless you can show it's typical.
• Don't skew what you collect: the FTC says asking only your happiest customers for reviews can mislead if it substantially skews how positive the reviews look.
• Don't hide the bad ones: delaying only negative reviews, even by a few days, can give people a biased picture.
• Disclose connections: if a reviewer got free products, money or a discount, disclose it clearly, and never make an incentive depend on the review being positive.

The FTC's rule on fake reviews

The FTC's Consumer Reviews and Testimonials Rule took effect on 21 October 2024. Among other things, it bans fake reviews and buying reviews that must express a positive or negative view. It also bans undisclosed reviews by insiders and some forms of review suppression. The FTC can seek civil penalties for knowing violations.

Practical rules for your ads

• Ask permission before you quote a named customer in an ad, and keep a record of their agreement.
• Quote word for word, since trimming is fine if it doesn't change the meaning but rewording is not.
• Pick quotes that reflect a common experience rather than only the best one.
• Add context to any standout result so readers know what's typical.
• Disclose free products, payments or other connections between you and the reviewer.
• Check each quote against platform ad rules too, because a review that says "cured my eczema" is still a health claim once it's in your ad.

Turn review themes into ad angles

An ad angle is the reason someone buys, expressed in their own words. Each strong theme from your mining can become one angle, and each angle becomes one ad to test. The table below shows how different kinds of themes turn into ads.

Turn my review research into ad angles and review-based ad copy.

My mining results: [paste the Reasons and Doubts tables, with exact phrases and counts]
Product facts, each with a source: [paste fact sheet]
Quotes I have permission to use: [list item numbers, or "none yet"]
Where the ads will run: [Facebook and Instagram, TikTok, Google]

Do this:
1. Pick the 3 strongest angles: two from Reasons, one from Doubts. Say why, using the counts.
2. For each angle, write:
   - a hook in the customers' own words (exact phrase, or close to it if marked as a paraphrase)
   - primary text of 2 to 3 short lines, with one supporting fact
   - a headline of about 40 characters or less
   - a picture or video idea in one line
3. Where an ad quotes a customer, use only quotes on my permission list. Otherwise write the line as our own voice and mark it "paraphrase".
4. List each claim with its source: a fact from my sheet, or a review item number.

Checks: no claim beyond my facts; no result shown as typical unless my facts say it is; no line that says or implies a personal trait of the reader (health, age, money and so on).

Worked example: mining Gymshark's public reviews

Gymshark sells workout clothing and accessories, and its about page says Ben Francis founded it in 2012, in a garage in Birmingham, UK. Its Trustpilot page had more than 42,000 reviews when we looked in November 2026, which is a lot of public customer language to learn from. Gymshark has no link to this article, and we're only describing its public reviews in general terms.

What went in

The input was a sample of recent public reviews from that page. As with most large retailers, the mix includes both praise and complaints, and we paraphrase the themes rather than quoting anyone.

What came out

A real mining run would count how many different reviewers said each thing, and because we haven't counted here, treat the order of the themes as unproven.

An example draft (written for this guide, not Gymshark's ad)

Primary text: "[Exact customer phrase about comfort, with permission]" [One fabric fact from the label.] [Delivery terms, exactly as on the page.]

Headline: [Product name]: [the phrase most reviewers used]

The brackets are the point of this draft. The ad is shaped by what reviewers say, but every quote and fact must come from a real source before it runs. That is the difference between review-based ad copy and copy that only sounds like it.

A review mining sheet you can reuse

Use one row per theme, keep the sheet somewhere your team can share it and update it monthly as new reviews arrive.

Common mistakes, and what to do instead

• Using only five-star reviews: you miss the doubts your ads must answer, so mine the three-star and one-star reviews as well.
• Polishing customer words: a tidy summary like "comfortable and good quality" could describe anything, so keep the strange, specific phrase instead.
• Leading with the loudest review: count the people behind each theme first, then lead with what's common.
• Quoting without permission: ask first, keep a record of the answer and quote exactly.
• Featuring a miracle result as normal: add what people can generally expect, or pick a more typical quote.
• Turning complaints into spin: if customers say delivery is slow, fix it or state the real delivery times, because an ad that promises "fast delivery" only makes the complaints worse.
• Letting your AI make up a quote: tell it to use only the text you paste, and spot-check every quote against the original source.

Where AI Blocks helps

Everything here is free: copy the prompt or download the skill, then use it in your own AI with your own reviews.

• Customer buying reasons finder: paste reviews or support emails to get buying reasons in customers' exact words, ranked by how many people said them.
• Customer buying reason counter: share at least 20 reviews to get reasons ranked by distinct people, with exact quotes.
• Unanswered customer question finder: find the customer questions your product page leaves unanswered.
• Buyer identity segment finder: get buyer groups drawn from their own words, with sample counts for each.
• Customer review ad copy writer: a workflow that runs from reviews and product facts to a buyer brief, an ad angle and fact-based drafts.
• Image ad maker from customer quotes: turn your best review lines into quote image ads, with flags for quotes to check first.
• Customer feedback and brand reviewer: find new buyer language and doubts in fresh feedback each month.

Once you have your angles, our guides to Facebook ad copy and the video ad script template show how to write the ads. After that, plan a fair test to see which angle wins with real buyers.

Questions people ask

Can I use customer reviews in my ads?

Usually yes, as long as you take care. Ask the customer's permission before quoting them, quote exactly, and don't pick reviews in a way that gives a false picture. In the US, follow the FTC's endorsement guidance, and if a review sits on another site, check that site's terms before reusing the text.

Do I need permission to quote a review in an ad?

It's the safe default, because written permission protects both you and the customer and lets you confirm that the quote is genuine. Keep a record of who agreed, when they agreed and which words they approved.

How many reviews do I need to find ad angles?

Twenty gives you a first look, while fifty or more makes the counts more useful. If a theme rests on only a few people, test it as a hunch rather than treating it as proven.

Can I use negative reviews in marketing?

Yes, as research, because complaints show the doubts your ads should answer and the problems you need to fix. You can also quote a mild grumble that proves a strength, such as a complaint that something sold out too fast, under the same permission and fairness rules.

Can AI write fake reviews for my ads?

Don't do it. In the US, the FTC's rule bans fake reviews and testimonials, and the FTC can seek civil penalties for knowing violations. Use AI to sort and count real customer words, never to invent them.