Customer research for ad creative: turning what customers say into hooks.

The best hook in most accounts was written by a customer. It is sitting in a three-star review, a support ticket from someone about to buy, or an answer to a survey question nobody on the team reads. The job is not to invent better lines. It is to find the lines customers already use and put them where a scrolling stranger will hear them.

This is a method, not a mood board. It covers the five research sources worth mining, a coding scheme that turns messy text into countable themes, the process from raw quotes to a hook bank, how many hooks to brief each month, and the ways research-led creative goes wrong. It follows directly from why similar ads compete instead of adding up: research is the cheapest source of genuinely different concepts, because customers want more different things than any brand team assumes.

Five sources and what each is good for

Each source captures a different moment in the customer’s head. Use all five, because each fills gaps the others leave. Together they are the most useful first-party data a creative team has, and most of it is already sitting in tools you pay for.

Product reviews

Reviews carry outcome language: what changed for the customer after buying. Five-star reviews give you desire and delight. Three- and four-star reviews are often better, because they contain the nuance, such as it took a week to get used to, or I wish I had bought the larger size. That nuance makes a hook sound true. Read reviews for your products and for your closest competitors, since a competitor’s one-star reviews are a list of objections your product may already answer.

Post-purchase surveys

A survey on the thank-you page or in the confirmation email reaches people at the moment they remember the decision best. Keep it to one or two open questions. Three that consistently produce hook material: what almost stopped you from buying today, what were you using before this, and how would you describe this to a friend. Avoid multiple choice for this purpose. Options you write yourself return your own language, not theirs.

Support tickets

Pre-purchase tickets are objections in writing. Does it fit, will it arrive by Friday, is it safe for my dog, how is this different from the cheaper one. Tag tickets by whether they were sent before or after purchase, and mine the before pile first. Every repeated pre-purchase question is a hook that answers a doubt before the viewer has to ask it.

Return reasons

Return reasons tell you where the promise and the product disagree. Too small, not as shown, did not do what I expected. They are rarely hooks themselves. They are guardrails. If a common return reason is that the product looked bigger in the ads, a hook that oversells size will raise returns even while it raises clicks. Read return reasons before you approve a hook, not after.

Search terms

Search terms capture the words people use at the moment of need, before your brand is in their head. Two primary places to find them: the search terms report in Google Ads, which shows the actual searches that triggered your ads, and the Performance report in Google Search Console, which shows the queries your site appeared for in organic results. Look for modifiers, such as for sensitive skin, for small apartments, without a subscription. Modifiers are personas and objections in two or three words.

The coding scheme

Raw quotes are not a strategy. Coding them turns hundreds of lines into a short list of themes you can count. Give each quote one primary code from this set, plus the verbatim phrase that carries it.

• →Desire: the outcome they wanted. Sleep through the night, stop rebuying cheap ones, look put together on video calls.
• →Problem: the pain before the purchase. My feet were soaked by mile three, the old one broke in a month.
• →Objection: the reason to hesitate. Too expensive, will it fit, I have tried these before.
• →Trigger: the moment that started the search. A new baby, a house move, training for a race, a gift deadline.
• →Comparison: what they used before or considered instead. The drugstore version, a competitor, doing nothing.
• →Identity: how they describe themselves. Busy parent, picky coffee drinker, someone who hates clutter.

Keep the sheet simple: source, date, product, code, verbatim phrase, and a count column you increment when the same theme shows up again. The count is the point. A theme that appears forty times across four sources is a concept. A single brilliant quote is an anecdote.

The six codes map onto the axes of difference. Desire and problem feed the desire axis, identity and trigger feed persona, and objection and comparison feed angle and proof. Coding your research is therefore also a way to see which kinds of concepts your account has never tried.

From raw text to a hook bank, step by step

1. Pull the last 90 days from each source. For reviews, include your top products and two competitors. For tickets, filter to pre-purchase conversations.
2. Sample if you have to, but sample evenly across sources and star ratings. Two hundred reviews read closely beat two thousand skimmed.
3. Code every quote with one primary code and the verbatim phrase. Resist rewriting the phrase into brand voice at this stage.
4. Count themes. Merge quotes that say the same thing in different words, and keep the most vivid phrasing as the representative line.
5. Rank themes by count and by how many sources they appear in. A theme in reviews, tickets, and search terms is stronger than a theme in one source.
6. Write two or three hook lines per top theme, each keeping the customer’s wording in the first sentence. Then note the format and proof type that fit it.
7. Log every hook in a shared bank with its source, theme, count, and status: briefed, live, won, retired. The bank becomes the team’s memory of what customers said and what the market answered.

One practical point on the first three seconds. Meta’s own guidance for conversion-focused video recommends putting motion and the benefit in the first three seconds and promoting a single benefit per ad. A research-led hook fits that naturally, because it leads with one thing a real customer cared about rather than a list of features.

Worked example: research source to hook

The table below uses a fictional product, a cold-brew coffee concentrate sold on Shopify, and invented customer phrases. It is an illustration of the method, not real research and not a client example.

Notice the return reason row. It did not produce a hook that hides the strength. It produced one that sets the expectation, which should lower returns from exactly the customers who would have sent it back. Notice too that the six rows cover four different codes. That spread is what makes the resulting concepts diverse rather than six takes on one idea.

How many hooks to brief each month

The answer comes from your testing budget, not from how many ideas the team can generate. A hook you cannot fund to a readable result is a hook you cannot learn from. Work backward from the spend you can put behind new concepts.

Here is the arithmetic with illustrative numbers. Say prospecting spend is $60,000 a month and you set aside 20% for testing new concepts, which is $12,000. Say your read rule needs about $1,500 per concept before you judge it. That funds eight concept reads a month. At two or three hooks per concept, delivered as assets inside one ad, you brief sixteen to twenty-four hooks a month. The ad budget calculator helps set the monthly total those numbers sit inside.

Two adjustments. If most of your concepts fail before the read rule, brief fewer concepts and spend more research time on each. If most pass, your read rule may be too loose, or your research is working and you should fund more tests. Either way, the count follows the budget and the hit rate, and how to build a creative testing system that scales covers the read rule itself.

Stores between $250k and $500k a month usually feel this most sharply, because the account can now absorb more concepts than the team can produce. That is the band where creative supply becomes the binding constraint, and a research-fed hook bank is the cheapest way to keep supply up without lowering the bar.

Where research-led hooks go wrong

• →Cherry-picking. One vivid quote gets briefed because the team loved it, while a theme that appeared fifty times gets ignored. Brief from counts, not favorites.
• →Rewriting into brand voice. The quote gets polished until it sounds like every other ad. Keep the customer’s wording in the first line, even when it is plain.
• →Overpromising. A review describing an exceptional result becomes a headline. The FTC’s endorsement guidance says ads featuring specific results are read as what others can expect, and that a line like results not typical does not change that. Use claims you can substantiate.
• →Using reviews without checking terms. Quoting a customer in an ad can raise permission and disclosure questions, especially if the review was incentivized. This is not legal advice. Check your review platform’s terms and ask a lawyer when in doubt.
• →Treating search terms as scripts. Search queries are fragments. They tell you the need and the modifier, not the sentence to say.
• →Researching once. Customer language drifts with seasons, competitors, and your own product changes. Refresh the bank every month, not every year.

What to do this month

1. Add one open question to your post-purchase flow: what almost stopped you from buying today.
2. Export 90 days of reviews, pre-purchase tickets, return reasons, and search terms into one sheet.
3. Code at least 200 quotes with the six codes and count the themes.
4. Brief your top three themes that have no live concept, two or three hooks each, and check each against return reasons before approval. Brief now and launch in early November, or hold new concepts until after Cyber Monday.
5. Set up the hook bank with source, theme, count, and status columns, and review it alongside results at month end.

If you would rather hand the research-to-brief loop to a team that runs it every week, our DTC growth agency service directs creative strategy from customer evidence and analyzes creative performance against it.