Above the fold on your homepage is the most expensive real estate your business owns. It is also, on most sites I audit, occupied by a sentence that could belong to any company in the category — and a stock phrase like “Welcome to our website” is not a headline, it is a doormat.
Here are seven hero structures that consistently outperform vague positioning, when each one is the right choice, and how to pick without guessing.
The eight seconds you actually get
The numbers people quote about attention above the fold vary wildly depending on who measured and how, so treat any single figure with suspicion. What is consistent across the eye-tracking and scroll-depth research is the shape of the behaviour: visitors form a rough judgement very quickly, they read headlines far more than body copy, and the decision they are making is not “should I buy” but “is this for me”.
That reframes the job. Your hero does not need to sell. It needs to pass a filter test in the reader’s head: is this person talking about my situation? Every formula below is a different way of answering yes.
The five parts of a working hero
Regardless of which formula you use, the anatomy is stable:
- Headline — the claim, in your customer’s language.
- Sub-headline — who it is for and how it works, in one sentence, usually 15–25 words.
- Primary CTA — one button, verb-led, naming the next screen.
- Risk reducer — a short line under the button that lowers the cost of clicking.
- Proof shard — one number, one quote or one logo row. Not all three.
Most heroes that fail are missing parts two and four. The headline gets all the attention in workshops; the sub-headline does most of the work in practice.
Formula 1: Outcome-first
Structure: [Achieve desired outcome] without [the thing they dread].
Use it when your customers know they have the problem and are actively shopping. It is the safest, most versatile formula and the right default if you are unsure.
Headline: Fill your clinic’s Tuesday mornings — without discounting.
Sub: Patient reactivation campaigns for allied health practices in Australia and New Zealand. We write them, you approve them, your admin sends them.
CTA: See a sample campaign
Risk reducer: Real campaign, real numbers. No email required.
The “without” clause is what stops this being generic. It names the trade-off your competitors quietly ask people to accept.
Formula 2: Mechanism
Structure: [Familiar outcome], via [distinctive method].
Use it when your category is crowded and the outcome is table stakes. If everyone promises the same result, the differentiator has to be how.
Example: “Website copy built from interviews with your actual customers — not a questionnaire and a guess.” The outcome (website copy) is undifferentiated. The mechanism (customer interviews) is the claim.
This formula fails when the mechanism is not genuinely distinctive, or when your audience does not yet care how it works. Do not lead with mechanism to a first-time buyer who has not accepted the problem.
Formula 3: Contrast
Structure: Not [the thing they expect]. [The thing you actually are].
Use it when you are frequently mistaken for something you are not, and that mistake is costing you the right leads.
Example: “Not another agency retainer. One project, one fixed price, one writer who answers your emails.”
Contrast is high-risk and high-reward. It repels people, which is the point — but if you are not confident about who you want to repel, it will quietly shrink your pipeline. Only use it when you have enough demand to be choosy.
Formula 4: Category-defining
Structure: The [new category name] for [specific audience].
Use it when you genuinely do not fit an existing category, and the closest label actively misleads people about what you do.
The failure mode is inventing a category nobody is searching for and then having to explain it in the sub-headline, which costs you the very seconds you were trying to save. A safe hybrid: use the recognised category in the sub-headline while the headline stakes the new ground. If the reader has to learn a new noun before they understand you, you have spent your budget on vocabulary rather than persuasion.
Formula 5: Proof-led
Structure: [Striking, specific number] [what it refers to].
Use it when you have a number that is genuinely unusual and independently checkable.
Example: “41 clinics reactivated 2,900 dormant patients last year. Here’s the exact sequence they used.”
Three conditions must hold: the number must be specific (2,900, not “thousands”), attributable (whose number, over what period), and relevant to the reader’s own scale. A number from an enterprise client will alienate a sole trader rather than impress them.
Formula 6: Question-led
Structure: [Question only your ideal customer would answer “yes” to]?
Use it when the problem is emotionally loaded and under-discussed, and naming it out loud creates immediate recognition.
Example: “Getting plenty of traffic and almost no enquiries?”
Two cautions. First, never ask a question the reader can answer “no” to and leave. Second, the question must be followed immediately by a statement — a question hero with a question sub-headline reads as evasive, and after eight seconds of interrogation people go elsewhere.
Formula 7: Problem-agitate-relieve
Structure: [Name the problem]. [Show the cost]. [Point at the relief].
Use it when the reader is aware of the symptom but has not connected it to a cause, and where the cost of inaction is concrete.
Compressed into a hero, that becomes a headline naming the problem, a sub-headline quantifying it, and a CTA offering the first step out. “Your best page is your least-visited page. Most service businesses bury their strongest proof three clicks deep, where 4% of visitors ever reach it.”
The ethical line here is real and worth naming. Agitation means helping someone see a cost they already carry. It does not mean inventing a fear they did not have. If the cost you are describing would not survive scrutiny, you are not agitating — you are manufacturing, and it corrodes trust faster than it converts.
How to choose, and how to test
Pick using your audience’s awareness level, which is the variable that actually predicts which formula lands:
| Where your visitor is | Best formulas | Avoid |
|---|---|---|
| Doesn’t know they have the problem | Question-led, Problem-agitate-relieve | Mechanism, Category-defining |
| Knows the problem, not the solutions | Outcome-first, Problem-agitate-relieve | Proof-led |
| Comparing solutions | Mechanism, Contrast, Proof-led | Question-led |
| Comparing providers | Proof-led, Contrast | Problem-agitate-relieve |
| Ready to buy | Outcome-first with a hard CTA | Anything that re-sells the problem |
Then run the three tests before you commit:
- The competitor swap. Paste your headline on a competitor’s site. Still true? Rewrite.
- The five-second test. Show the hero to five people outside your business for five seconds, then hide it and ask what the company does and who for. Three out of five correct is a pass.
- The read-aloud test. Say it out loud to a colleague. If you would not phrase it that way in conversation, neither will your reader in their head.
Finally, on A/B testing heroes: it is the most over-recommended advice in conversion writing. Below about 1,000 sessions a month to the page, you will not reach significance before the seasons change. At that scale, replace testing with five customer interviews. The clarity you get will beat any winner a noisy test could have declared.