Facebook targeting: Meta itself documents why fine-grained targeting hurts
Meta's own documentation states that multiplying ad sets reduces what its system learns about each one, and recommends combining similar ad sets. Splitting one audience into five therefore forces you to produce five times the learning-phase exit threshold. Fine-grained targeting costs more and learns less.
- Meta states that the more ads and ad sets you create, the less its system is able to gather information about each one.
- It adds that by combining similar ad sets, you also combine their learnings. That is a recommendation, not an interpretation.
- The learning-phase exit threshold is about 50 results in the week following the last significant edit. It applies to each ad set separately.
- Five audiences tested in parallel therefore require 250 results per week. Most Quebec budgets never get there.
- What is left to decide is not the audience setting; it is the offer, the destination and what you declare as a conversion.
On this page
Audience fragmentation
Audience fragmentation is the practice of splitting a target audience into several separate ad sets, by interest, age, region or behaviour, in order to compare how they perform. It looks rigorous and it produces the opposite effect: each fragment has to reach, on its own, the volume of results needed to stabilize delivery. On a budget that does not allow for it, no fragment gets there, and the whole account stays in a permanent learning phase.
The arithmetic no one runs
Fine-grained targeting sells well because it looks serious. Split into segments, compare, keep the best: that is the method expected of a campaign manager, and it is the one written into most service proposals in Quebec.
Meta writes the opposite, though, in its own documentation on the learning phase. Two sentences are enough to overturn the method. The first: the more ads and ad sets you create, the less the delivery system is able to gather information about each one. The second: by combining similar ad sets, you also combine their learnings.
The math that follows is brutal. An ad set exits the learning phase after about 50 results in the week following the last significant edit. That threshold applies to each ad set separately. Five audiences tested in parallel therefore need 250 results per week for the test to mean anything. At $40 per result, that is $10,000 per week. Most Quebec budgets fund a fraction of that amount, then read gaps between segments that measure nothing but noise.
The question to ask a vendor is not "how will you target." It is "how many results per week will each ad set produce." If the answer is below the threshold, the proposed structure cannot produce a conclusion, however fine the targeting.
What to settle before approving an account structure
This is a subject where the mistake looks like rigour. A fragmented structure produces detailed tables, segment-by-segment comparisons and a thick report, while guaranteeing that no decision can be made with any certainty.
- How many results does our account produce per week, across all ad sets?
- How many active ad sets do we have, and how many reach the threshold individually?
- How long have our ad sets been in the learning phase, and who checks it?
- What margin do the sales these campaigns actually produce carry?
- If no ad set exits the learning phase within two months, what gets shut down?
The answer that holds up starts from your weekly volume of results and derives the number of ad sets you can fund. An evasive answer proposes detailed segmentation without mentioning volume, promises per-segment insights on a small budget, or presents a per-audience report as proof.
Deciding how many ad sets your budget can keep alive is an arithmetic trade-off that is settled on your own numbers. A 90-minute consultation settles it, with a written summary your team or your agency can execute.
What is really left to decide
If the audience setting is no longer the lever, the question becomes: what does the decision still rest on. Five choices, and none of them is in the targeting panel.
| What gets decided | Why it matters more |
|---|---|
| What you declare as a conversion | The system buys people who resemble those who triggered it |
| The value assigned to each conversion | Without differentiation, an inquiry counts as much as a sale |
| The offer and its price | No targeting makes up for an offer the market refuses |
| The landing page | It decides half the cost per result |
| The number of ad sets funded | It determines whether the account can learn anything at all |
The first row is the one that moves the most money. The real targeting of a modern campaign happens through what you declare as a result, not through checked boxes. The subject is developed in what breaks a conversion setup, and the break-even calculation in the break-even threshold of a campaign.
What stays in-house: the real weekly sales volume, the margin per product or service, and confirmation of sales closed offline. These three numbers decide the number of ad sets you can fund, and no one outside the company holds them. What gets delegated: the account structure, the grouping of ad sets, the setup of conversions and their values, the production of variants and the reading. A business willing to keep two properly funded ad sets alive learns more than one that launches twelve.
Audiences built from your own data
One point escapes the arithmetic and deserves a separate decision: audiences built from your own customer lists or from behaviour on your site, what Meta calls Custom Audiences.
These audiences are often the most effective, because they rest on people who already know you. They also rely on transmitting personal information to a third-party platform, which in Quebec falls under the rules protecting personal information. The Commission d'accès à l'information lists, among the failures that carry a penalty, neglecting to adequately inform people of the collection of their information. The subject is covered in Law 25 and marketing.
This decision is not delegated to a campaign manager. Handing a customer list to an advertising platform is a business choice that engages your liability, not a technical setting. It is validated with legal counsel, and it is documented. This article reports what public bodies publish; it does not constitute legal advice.
The overall logic of the channel is covered in diagnosing a Meta account, test duration in testing a Facebook ad, and the effect of how delivery works on creative in Andromeda.
This channel is compared to the others in concentrating budget on one channel.
Bringing ad spend back to a defensible return is at the heart of the Optimize the profitability of your digital campaigns goal.
Already running a marketing team? See how we plug in as reinforcement on Meta advertising.
Frequently asked questions about Facebook targeting
Should you segment your audience finely?
Rarely, and Meta documents it. Its page on the learning phase states that the more ad sets you create, the less its system gathers information about each one, and recommends combining similar ad sets. Each ad set has to reach the threshold separately, which makes segmentation expensive.
How many ad sets do you need?
As many as your weekly volume of results can push out of the learning phase. The threshold is about 50 results per ad set in the week following the last significant edit. Divide your weekly volume by 50: that is the maximum number of ad sets your budget can keep alive.
What does targeting rest on today?
On what you declare as a conversion and the value you assign to it. The system buys people who resemble those who triggered the declared results. A poorly defined conversion therefore steers the buying more surely than any box checked in the audience panel.
Can you use your customer list to target?
It is possible and often effective, but it is a business decision, not a setting. Transmitting personal information to a third-party platform falls under Quebec's rules protecting personal information, among them the obligation to adequately inform people of the collection. To be validated with legal counsel.
- Meta Business Help Center, "About the learning phase," official documentation on the exit threshold and on the effect of the number of ad sets, accessed July 2026.
- Meta Business Help Center, "How to edit Facebook and Instagram ad campaigns in Meta Ads Manager," on the edits that restart the learning phase, accessed July 2026.
- Commission d'accès à l'information du Québec, Penalties and prosecutions, accessed July 2026.
Gabriel almost always takes your first call and carries out your audit. He builds the strategy starting from your growth goal: where to put your budget, which market to test and how to connect each lead to a real sale in your CRM. He mainly leads engagements for three goals: Optimize the profitability of your digital campaigns, Develop a new market, and Generate demand and growth. With Geneviève, he also works on organic search (SEO), AI visibility (GEO) and conversion rate optimization (CRO). The sales a Google Ads or Meta Ads campaign brings in depend on the page that receives the click. He writes mainly about marketing strategy, paid advertising and measurement.
About Falia →
SEO
Paid advertising
Conversion
AI visibility
Web design