Why the angle decides who responds
The creative hub sets out the mechanism. Meta documents that its auction weighs a bid together with estimated action rates and ad quality, and that estimated action rates draw on how people have responded to the ad. An ad that speaks to a situation one kind of person recognises earns a response from that kind of person, delivery learns who they are, and the ad is shown to more of them. The creative has, in the hub's phrase, written the audience definition.
The angle is the part of the creative that does that writing. Two ads for the same product, one about saving time on a weekday morning and one about impressing guests at the weekend, will be shown to different people within the same broad audience, because different people respond to each. That makes the choice of angle a targeting decision as much as a messaging one, and it is the reason a weak angle cannot be rescued by a better hook or a bigger budget: the ad is finding the wrong people efficiently.
It also explains why customer evidence is the first place to look for angles, before the product team's feature list. Customers describe the moment they needed the product in their own words, and other people in that moment recognise the words. A feature described in the company's language describes the product, which nobody in the feed was looking for. The terms on this page fit together as follows: the angle is the reason to care; the concept is the creative idea built on it; the hook is the opening of one ad. The creative testing guide treats concept as the biggest variable and tests it first, and the angle is the raw material for every concept.
Where customers have already told you
Most businesses hold more customer language than they use. The sources differ in what they reveal, and reading several of them is what separates a recurring theme from one person's quirk. Everything below is general research practice; the site's research guides cover each method in depth.
| Source | What it tends to reveal | Watch for |
|---|---|---|
| Reviews, your own and on marketplaces | The outcome people value, in their words, and the surprises, good and bad | Reviews skew to the delighted and the disappointed; the reviews and social proof guide covers collecting them fairly |
| Sales calls and sales notes | The situation that triggered the enquiry and the objections raised before buying | Notes record the salesperson's summary; ask for the customer's phrases |
| Support tickets and emails | Problems in use, misunderstandings and the jobs people hired the product for that you did not expect | Mostly problems; read them for the situation, not only the complaint |
| Returns and cancellation reasons | Where the promise and the experience diverged | Short, coded reasons hide detail; read the free text |
| Post-purchase surveys | Why people chose you and what nearly stopped them, at scale | Leading questions produce the answers you wanted; the survey design guide covers wording and order |
| Customer interviews | The full story of a purchase, from the first trigger to the decision | Small numbers; the customer interviews tutorial covers recruiting and a discussion guide |
| Comments under ads and posts | Objections and questions from people who have not bought yet | Public and unfiltered; the community management guide covers response and moderation |
Comments deserve a particular mention because they come from the people the ads are trying to reach, not from customers. The Meta Ads setup for consideration staffs the comments during a campaign and logs recurring objections for the next round of creative; that log is one of the richest angle sources an account has, and it costs nothing to keep.
Collect excerpts, not summaries. A phrase such as "I could finally sleep on the night shift" carries a situation, a problem and an outcome in a few words, and summarising it as "sleep quality" throws all three away. Record where each excerpt came from and when, and keep personal details out of the working file.
Coding what people say: situation, problem, outcome, objection, proof
Coding means tagging each excerpt with the parts that matter for an ad. Five codes cover most of what customers say.
- Situation: when and where the need arises. "On the train", "when the baby wakes", "before a big presentation".
- Problem: what goes wrong in that situation without the product, or with the alternative they used.
- Outcome: the progress the person wants, functional and emotional. Often the most quotable part.
- Objection: what nearly stopped them buying, or what they doubted. Price, comfort, whether it would work for someone like them, effort.
- Proof: what convinced them, such as a demonstration, a guarantee, a friend's recommendation or a detail they noticed.
The first three map onto a job statement from the jobs to be done framework: when a situation arises, I want to make progress, so that I can reach an outcome. Writing angles that way keeps the product out of the statement and the customer's moment in it. The value proposition canvas adds the other side: for each pain and gain the excerpts reveal, what in the offer relieves or creates it, and where nothing does.
Code in a simple sheet, one excerpt per row, with a column for each code and one for the source. Then group rows that describe the same situation and outcome. Each group is a candidate angle. Keep the objections and proof codes attached to it, because an angle without proof is a claim the business cannot support, and an angle whose main objection is never answered will earn clicks that do not convert. Code for the person too where the evidence allows: a buyer personaA research-based profile of a typical customer type: goals, problems and buying behaviour. is useful only if it is built from rows like these rather than from guesses.
Building an angle bank (table)
The angle bank is the coded evidence turned into a list the creative team can use. Each row is one angle, written as a job statement, with the evidence and the constraints beside it.
| Column | What it holds |
|---|---|
| Angle | The job statement: situation, progress, outcome, in the customer's words |
| Who | The kind of person who described it, as the evidence shows |
| Verbatim lines | Two or three phrases worth hearing in an ad, with their sources |
| Main objection | What these people doubted, and how the business answers it |
| Proof held | The demonstration, guarantee, review or data that supports the claim the angle implies |
| Evidence | How many excerpts, from how many independent sources |
| Standards check | Any claim, before-and-after or personal-attribute risk, noted for review |
| Status | New, in test, proven, retired, with the test reference |
Score each angle before choosing. A simple three-level score on a handful of criteria is enough: recurrence (does it appear across independent sources, not only in one channel?), specificity (can you picture the moment?), proof (can the business show it, not merely say it?), breadth (does it describe many of the people the account wants, or a small niche?), distinctiveness (would a competitor's ad say the same?), and risk (does it need claims the standards make difficult?). The scores are judgements, not measurements, and their job is to make the choice explicit, so that the team can see why one angle was tested before another.
Choosing which angles to test first
Each angle that reaches a test becomes a HypothesisA testable statement of the change to make, the effect expected on a named metric and the evidence-based reason for expecting it.: people in this situation will respond to this outcome, and the people delivery finds through it will buy. The creative testing guide explains why the biggest uncertain variable goes first, and on a new angle bank that variable is the angle itself. Test angles as concepts before testing hooks within them, because a better opening on the wrong angle answers a question you cannot use.
Pick angles that differ meaningfully. Two angles about saving time in slightly different situations will produce two similar ads and an inconclusive result; an angle about the morning rush against one about weekend guests will tell you something. Include the account's current best concept as a control where the budget allows, so that the result says whether the new angle beats what already works rather than only which of two new ideas is less bad. Run the comparison with the method the testing guide recommends for deciding between concepts, the A/B test tool or separate ad sets with equal budgets, judged on the result that matches the objective rather than on click-through.
Once an angle has proved itself as a concept, the hooks within it become the next test. The ad creative hook test sets out a design for comparing openings on the same body, and the customer's verbatim lines from the angle bank are the first candidates.
From angle to brief and hooks
The creative hub's performance brief names one person, one situation and one claim, plus the proof, the offer, the formats and the metric. An angle bank row supplies almost all of it: the who, the situation, the outcome as the claim, the proof held and the objection the ad must answer. The creative brief template gives the rest of the structure. A Creative BriefA short document telling the people making creative work what it must achieve, for whom, with which message and within which limits. written from an angle row has one advantage over one written from a product sheet: the creator can read the customers' own lines and hear how the moment sounds.
Hooks come from the verbatim column. The hooks and user-generated content guide lists hook patterns, and several map directly onto codes: a problem-first hook opens on the problem code, a social proof hook on what other customers said, a result-first hook on the outcome. Paraphrase customer language into the ad freely; quoting an identifiable customer, using their review as on-screen text or their video as the ad needs their written permission, which the hooks guide covers under rights. Creator content briefed from an angle bank lets the creator say what customers already say, which is the quality the hooks guide describes in user-generated content: a peer speaking from experience rather than a presenter reading copy, and a natural carrier for social proofEvidence that other people have chosen, used or approved something, used to reduce a buyer's doubt..
Staying inside the Advertising Standards
Customer evidence produces strong lines, and some of them are lines an advertiser may not use. The site's pages describe the relevant parts of Meta's Advertising Standards as follows; the standards themselves are the current reference and they change.
- Claims must be substantiated. The hooks guide notes that the standards prohibit misleading claims and deceptive practices. A customer who says the product "fixed" something has made a claim the brand must be able to support, or leave out.
- Testimonials must be genuine. Endorsements have to reflect real experience, so a line from the angle bank can be used as the experience of a real customer who agreed to it, not reworded into a promise.
- Health and appearance rules restrict some imagery. The hooks guide notes that before-and-after imagery and content implying unrealistic results are restricted under the personal health and appearance rules.
- Ads must not assert or imply personal attributes. The healthcare marketing guide explains that the standards prohibit ads that assert or imply personal attributes, with medical conditions among them. An angle about a situation is safer written as a description of the product and who it is for than as a question telling the reader what they have.
Use the standards check column in the bank to flag these risks before a brief is written, and keep the substantiation for any claim that survives. An angle that can only be expressed through a claim the business cannot prove is not an angle the account can use, however often customers repeat it.
An earplug maker's angle bank (illustrative)
Hushmarten Earplugs codes its customers into six angles
Hushmarten Earplugs, an invented maker of reusable earplugs, had run the same "premium comfort" concept for most of a year. The marketing lead coded 420 reviews, 160 support emails and 230 comments from the account's ads over one quarter, one excerpt per row, and grouped them into six angles.
| Angle | Excerpts | Independent sources | Proof held | Standards note |
|---|---|---|---|---|
| Sleeping through a partner's snoring | 138 | Reviews, emails, comments | Comfort guarantee, fit demonstration | Describe the situation, not a condition |
| Sleeping in the day after a night shift | 74 | Reviews, emails | Reviews from shift workers, with permission | None |
| Concerts without ringing ears afterwards | 61 | Reviews, comments | Filter demonstration | Avoid hearing-damage claims without evidence |
| Focus in an open-plan office | 52 | Reviews | None beyond reviews | None |
| Comfort for side sleepers | 47 | Reviews, emails | Fit demonstration | None |
| Travel and flights | 29 | Reviews | None | None |
The team scored recurrence, specificity, proof and breadth, and chose two angles for a concept test against the incumbent: the partner's snoring and the night shift. Both were filmed as creator videos with the same offer and landing page. Over three weeks the test returned a cost per purchase of £19 for the snoring angle, £23 for the night-shift angle and £27 for the incumbent, with the snoring angle's buyers also showing a lower return rate in the shop's data. Comments under the night-shift ads kept asking whether the earplugs block alarms, an objection nobody had briefed for. The next quarter's briefs went to three hooks on the snoring angle, a night-shift version that answers the alarm question in the first seconds, and a first test of the concert angle once the filter claim had been checked.
Common mistakes
- Inventing angles in a meeting. Customers have usually said the better line already; read what they wrote before writing for them.
- Summarising instead of quoting. "Sleep quality" loses the situation and the words that make people recognise themselves.
- Reading one source. Reviews alone over-represent the delighted; add support, sales and comments.
- Angles without proof. An outcome the business cannot show becomes a claim it cannot support.
- Testing hooks before angles. A better opening on the wrong angle finds the wrong people faster.
- Testing near-identical angles. Two versions of the same idea give an inconclusive answer.
- Quoting customers without permission. A review on your site is not a licence for paid use.
- Writing the reader's condition into the ad. Describe the product and who it is for; do not tell the reader what they have.
- Letting the bank go stale. New comments, tickets and reviews arrive every week; code them on a schedule and retire angles that tests have settled.
