werbero.comThe magazine for advertising that works

Online Advertising

Meta Ads: cutting audiences properly

The tighter the audience the better the advertising. That held until about 2021 and is now the most expensive misconception in most accounts.

Jul 8, 2026 2 min read 521 words
Meta Ads: cutting audiences properly

Key points

  • Narrow audiences raise the cost per thousand and stall the learning phase.
  • Your own customer list is the one signal competitors cannot copy.
  • An ad set needs 50 conversions a week or it never finishes learning.

Anyone advertising on Facebook in 2018 learned to cut audiences as narrowly as possible: job, interests, life events, all combined. That technique no longer works, and the reason is not privacy but arithmetic.

Why narrow audiences got expensive

Delivery runs through an auction per ad slot. The smaller the pool of possible recipients, the more often the same person must be addressed to spend the budget. Rising frequency means falling response and rising cost.

Audience sizeTypical weekly frequencyEffect
under 50,0006 to 12fatigue after 10 days
100,000 to 500,0003 to 5tolerable for short bursts
500,000 to 2 m1.5 to 3the norm for continuous running
over 5 munder 1.5only sensible with a strong creative

The ranking of signal quality

Not all audience types are worth the same. This order holds up in practice.

  1. Your own customer list. Uploaded email addresses of existing customers. The best signal, because nobody else has it.
  2. Site visitors of the last 30 days. Reliable, provided consent was collected properly.
  3. Engagement with your content. Video watched to 75 per cent, profile visited, form started.
  4. Statistical lookalikes. Derived from 1 and 2, one to three per cent similarity.
  5. Broad with location and age. Surprisingly often the winner above £3,000 a month.
  6. Detailed interests. As a seasoning, not a foundation.
A business that has never uploaded its customer list is giving away the one advantage a small firm has over a large one: real knowledge of its customers.

The 50-conversion rule

An ad set leaves the learning phase at roughly 50 conversions in seven days. Below that it stays permanently in learning and results swing hard.

Do that sum before you build. At £40 per enquiry you need £2,000 a week per ad set to clear the bar. If your total budget is £1,200 a month, you may run exactly one ad set, not five.

Overlap costs real money

When two ad sets reach the same people you bid against yourself. The system tries to prevent it and does not fully succeed.

Practical rules: always exclude the remarketing audience from the prospecting set; exclude existing customers from acquisition campaigns except for repeat purchases; never run lookalikes of different percentage tiers in parallel, only the larger; and keep to at most three ad sets per campaign, or overlap is unavoidable.

The data protection frame that gets skipped

Uploading a customer list is a processing of personal data. You need a legal basis, normally a consent that names this purpose explicitly. A newsletter consent does not automatically cover advertising delivery on a platform.

Likewise the measurement tool on your site needs consent before it loads. Get either wrong and you risk not only a fine but the loss of the whole data base if challenged.

A build that works in small accounts

For budgets between £500 and £3,000 a month this lean build has proven itself: one campaign, conversion objective, budget at campaign level; ad set A broad, location and age only, no interests; ad set B remarketing of the last 30 days, excluded from A; three to four creatives per ad set, at least one of them vertical.

Anything beyond that is fine-tuning that only pays above roughly £5,000 a month.

Frequently asked questions

How large should an audience be?

For a national market, 500,000 to 2 million people is a good starting range. Below 100,000 the cost per thousand climbs sharply because the same people are hit too often.

Do detailed interests still help?

Less and less. They now act as a starting signal for the algorithm rather than a hard boundary. As sole steering they no longer work.

More from Online Advertising

Keep reading

All articles
Google Ads: what clicks actually cost in 2026

Online Advertising

Google Ads: what clicks actually cost in 2026

Cost per click is the most widely misread number in online marketing. Look only at it and you will pay more per customer than the competitor with the pricier click.

3 min readJul 28, 2026
Performance Max: when the black box pays off

Online Advertising

Performance Max: when the black box pays off

A campaign type that plays across every channel and shows you almost nothing. For some firms it is the strongest lever available; for others an expensive trap.

2 min readJul 24, 2026