How We Build Ad Content From a Blank Page

Most ad accounts start with a creative. A designer makes something, a copywriter adds a headline, and it goes live. The first real question, who is this for and why should they believe us, gets answered by the results, weeks later and at full price.

When we build ad content for a client from nothing, we reverse that order. This is what a month of that looked like across our accounts, without the client names.

1. Voice before copy

For a brand starting from zero, the first thing written is not an ad. It is a brand voice guide and the platform bios, so that every headline afterwards sounds like the same company. We confirm the handle, write the bio for each platform, and only then draft copy.

It sounds like admin. It is what stops a fortnight of ads reading as five different businesses.

2. Pick the audience, then park the rest

One campaign we rebuilt this month had two audiences on the table. We narrowed active spend to the primary audience and parked the secondary one as a longer-term investment, with a written reason for each.

That choice is the real work. Ad copy for two audiences at once tends to speak clearly to neither, so we would rather write one strong set than two thin ones.

3. Ground every claim in the product

The copy for that campaign was ranked, and each line was tied to a real product specification the client could verify. No claim went in because it sounded good. If we could not point to the spec sheet, it did not make the list.

We also built the creative brief from the client's own posts, so the visuals matched what their audience already recognised, and we shaped the lead-form options to fit the form template the client's sales team actually uses. A lead form that does not match the follow-up process creates leads nobody can work.

4. Write for the audience you actually have

For an app re-engagement push we wrote two separate sets of headlines, descriptions and primary text. One was for an audience built from the client's own customer export. The other was for a behavioural audience built from app analytics.

These are different people. The first already bought. The second has shown up and done things but may not have. The same line does not work for both.

5. Plan the spend on paper first

For a short paid push, the plan was approved before anything launched: the budget split, the target, the ad units, the organic calendar running alongside, the reporting cadence, and the payment terms. Writing it down means the week-two conversation is about results, not about what we agreed.

What the weekly read catches once it is live

Every week, each account is compared against its trailing four-week average, not against last week. Three findings from the month are worth passing on.

• Falling click-through with a flat cost per thousand impressions is the signature of creative fatigue. The audience is unchanged, the auction is unchanged, and the ad is tiring. Three weeks in a row of decline is the point to act.
• A tracking break can look like an efficiency win. In one account, the purchase event stopped firing for a full reporting week while delivery broadened sharply. Cost figures looked healthier because the conversions had vanished from the count, not because performance had improved.
• Sometimes there is no data, and that is the finding. One account could not be queried at all for a week. We reported "no performance data" instead of guessing from stale numbers.

The order, in short

1. Voice and platform basics
2. One primary audience, with the rest parked
3. Ranked copy, tied to verifiable product facts
4. Separate copy for separate audiences
5. A written plan before any spend
6. A weekly comparison against a four-week baseline

If your ads are live and you cannot say which of those steps was done deliberately, the account is probably being steered by its results rather than by a plan. That is the expensive way to learn.