Five static ads that each test one claim

A static ad is one image and one sentence. It is a useful first test once a product works and a stranger can open a page, because nothing in the file can hide a vague claim behind pacing or a voiceover.

Five is a set size that creates contrast. It is not a sample that guarantees a winning ad. If all five sentences make the same promise, you learn which color you prefer and almost nothing about the message.

Five claims, and what may sit under each one

Every line has to come from the public page or from the product. A screenshot of your own interface is evidence. A customer you do not have is not. If there is no measured result, leave the result out. Keep the same rule for generated assets: no invented quotes or numbers.

1. **Problem.** Name the situation in words the buyer already uses. Type on a plain background is enough.
2. **Product.** Crop a real screen to the moment the product does the job.
3. **Outcome.** Show the after-state the product actually produces.
4. **Objection.** Answer one doubt with a fact already on the page: a limit, a setup step, or what the product will not do.
5. **Fit.** Name a situation that should lean in, and one that should keep scrolling.

**Hypothetical: a support-inbox digest.** The tool lists unanswered questions from yesterday. The problem line is "Yesterday's questions are still in the inbox." The product image is that list. The outcome line is "One list of what still needs an answer," because that is what the screen shows. The objection line is "It only reads the inbox you connect," if that limit is true. The fit line is "For a founder who still closes support." "Teams save ten hours a week" does not belong, because nothing you can show proves it.

One sheet, one destination, one action

Open the page and write three source notes: the job, the screen that proves it, and one limit. Draft the five lines from those notes before you decorate anything. Build the images from your screenshots, your type, and your colors.

Public ad libraries are for noticing structures. The Meta Ad Library and the Google Ads Transparency Center can show that an ad exists. They do not show profit, audience, spend, or conversions. Do not download someone else's creative and swap the logo. Do not lift their sentences.

Track the draft in one sheet:

The shared URL matters. If each ad opens a different page, you are running two experiments at once. The page headline should be able to sit beside each line without describing another product. Landing page best practices is a concrete pass for that kind of clarity.

Choose the action before any money moves: a signup, a qualified reply, or a paid start. Clicks from the wrong people make a problem line look successful. There is no universal number of days that makes an ad test trustworthy. Compare the action you named, and only after each concept has enough of those actions to compare. An agent can fill the sheet. Creating an ad account, uploading, or spending is a separate decision you authorize on purpose. How to launch ads through an agent keeps that boundary in view.

Keep the files readable at phone size. One sentence, high contrast, the same palette on all five, and enough empty space that the sentence survives a small preview.

Prompt

Draft five static ads from the facts below. Concepts: problem, product, outcome, objection, and fit.

For each concept give: an on-image line of 12 words or fewer, an image description that uses only my screenshots or a plain text card, primary text under 40 words, and the click action.

Do not invent testimonials, customer counts, savings, awards, or quotes. Do not echo a competitor's lines. If a line is not supported by my facts, drop it and name the missing fact. Do not create accounts, upload ads, or spend money.

Facts:
- Buyer situation:
- Job the product does:
- What my screenshot shows:
- A limit already stated in public:
- Destination URL:
- Colors and words already on my page:

The drafts are finished when a stranger can tell the five lines apart, each line points at a fact you can show, the sheet has one destination and one action, and every image is yours.

Then paste that destination into the free scan. The score is a static read of the public page and the main gap a cold visitor hits. It is not a forecast of which ad will win. It tells you whether the page can carry the promise you are about to pay to repeat.