Meta advertising: how to diagnose a Facebook and Instagram account
Meta advertising now runs on three levers: the quality of the conversion signals you send, the real diversity of your creative portfolio, and a campaign structure simple enough to concentrate the learning. Manual targeting, long the main lever, has become the least decisive of the four.
- You diagnose an account in this order: signals, structure, creative, and only then audiences.
- The minimum viable budget depends on the number of conversions per week, not on an absolute amount.
- The campaign objective determines who Meta shows your ads to. It is the setting with the heaviest consequences.
- A three-stage funnel on a low-volume account fragments the learning rather than refining it.
On this page
Meta advertising
Meta advertising covers the ads shown across Facebook, Instagram, Messenger and Meta's audience network. It is billed mainly per impression, which sets it apart from search advertising billed per click. It runs on an auction system where the amount you bid matters less than the ad's estimated relevance to the person who sees it.
The diagnostic order for an account
When an account underperforms, the most common reflex is to change the audiences. It is almost always the wrong starting point, and it is what explains the months lost going in circles.
If the account has already been relaunched several times without the trend changing, the creative is rarely the first thing to blame. You diagnose an ad account the way you diagnose a piece of failing equipment: in a fixed order, starting from what feeds the system, not by replacing the most visible parts first.
Here is the order we apply on engagements. It runs from the most decisive to the least decisive, and it has been stable since the ad selection engine moved to neural networks.
| Rank | Lever | Question to ask | Symptom when it fails |
|---|---|---|---|
| 1 | Signals | Is Meta receiving complete conversions with values attached? | Unstable cost per result, gap with actual sales |
| 2 | Structure | Does the learning have enough volume to happen? | Ad sets stuck in the learning phase |
| 3 | Creative | Are there distinct angles or just variants? | Rapid fatigue, rising frequency, stalling reach |
| 4 | Audiences | Is the targeting needlessly throttling delivery? | High cost per thousand, capped volume |
| 5 | Bidding | Does the bid strategy match the objective? | Irregular spend, unspent budget |
An account with incomplete signals cannot be fixed by targeting. The system optimizes toward whatever you declare as the result. If that declaration is blurry, everything else is built on sand.
What to check before adding budget
An account with incomplete conversion signals spends without knowing what it is buying, and no budget increase fixes that. On most SMB accounts, four to six distinct creative angles are enough: beyond that, you produce variants of the same message and pay to distribute them.
- In what order was the account diagnosed, and what was ruled out at each step?
- Which conversion events are verified, and since when have they been reliable?
- How many genuinely different angles are running right now, not how many visuals?
- What share of the budget goes to testing, and what share goes to repeating what works?
- What sales result, not what cost per click, justifies the current structure?
The useful answer names a diagnostic order and says what was eliminated. A hollow answer proposes optimizing the audiences, refreshing the visuals, or raising the budget to let the system learn.
If you have to defend a budget increase internally and want a second opinion before you present it, a 90-minute consultation produces a written summary designed to circulate without us.
Choosing the campaign objective
It is the setting with the heaviest consequences, and the one most often chosen out of habit. The objective decides not only what Meta counts, but who it shows your ads to.
A traffic campaign is delivered to people who click often. A conversion campaign is delivered to people who buy often. These are not the same people, and that is why a traffic campaign generates plenty of visits that never turn into sales.
| Objective | Useful when | Avoid when |
|---|---|---|
| Sales | Conversion tracking is reliable and volume is sufficient | Fewer than a handful of sales per week |
| Leads | The sales cycle is long and qualification happens afterward | Lead quality is not measured downstream |
| Traffic | You want to feed a retargeting audience | You expect sales from the channel |
| Awareness | Brand launch, new market | The budget must show a short-term return |
| Engagement | Social proof to build on a post | No other case really justifies it |
Choosing Traffic because Sales costs too much means changing the measure, not the result. The cost per click drops, the cost per sale does not move, and the gap becomes invisible in the reports.
The minimum viable budget
The question comes up on every first call: how much to invest. The answer is not expressed in dollars, it is expressed in conversions.
Meta recommends roughly fifty conversions per week per ad set for the learning phase to end. Below that, the system keeps exploring without ever stabilizing its delivery, which makes results erratic and conclusions impossible.
So the calculation runs backward. You start from the observed or estimated cost per acquisition, multiply it by fifty, and get the minimum weekly budget for a single ad set. A cost per acquisition of $40 implies roughly $2,000 per week to exit learning on a single ad set.
If that amount exceeds what the business can invest, the answer is not to split further. It is to optimize toward a more frequent event higher up the path, such as add-to-cart or an information request, then measure what follows in your own system.
Structuring without fragmenting
The three-tier funnel structure, awareness then consideration then conversion, is taught everywhere. It is excellent on large accounts and destructive on small ones, for a purely arithmetic reason: it divides the conversion volume by the number of tiers.
A simple rule settles it. You add up the account's weekly conversions, divide by fifty, and get the maximum number of ad sets that can exit learning. If the result is below two, the tiered funnel is out of reach.
One objective per campaign
Mixing two objectives in one campaign makes the results unreadable and forces the system to arbitrate between two definitions of success.
As few ad sets as possible
Each ad set is a separate learning pool. Two ad sets at twenty-five conversions are worth less than a single one at fifty, even though the total is identical.
Budget at the campaign level
Letting Meta split the budget across ad sets produces better results than imposing a fixed split, except when one ad set has to be protected for commercial reasons.
Creative angles in the same ad set
The system compares the angles against each other, not you. Isolating them in separate ad sets removes the very comparison you are trying to get.
The most costly mistakes
| Mistake | What it produces |
|---|---|
| Editing a campaign every couple of days | Learning restarts, results never stabilize |
| Judging a campaign on three days | A decision made on statistical noise |
| Stacking up inherited exclusions | Delivery throttled without anyone knowing why |
| Optimizing for the click on a sales objective | Plenty of traffic, flat sales |
| Comparing Meta data to actual sales with no allowance for a gap | False conclusions in both directions |
The last point deserves a clarification. Meta attributes a sale to an ad seen or clicked within a given window. Your accounting system attributes the sale to the last point of contact. The two figures cannot match, and trying to make them match wastes more time than accepting the gap and following the trend.
A five-point diagnostic before touching the audiences
The method for mapping creative angles is detailed in our article on the Andromeda update. The platform side is covered on the Meta advertising page.
The framework for choosing and concentrating budgets is set out in choosing and concentrating advertising budgets.
Meta advertising is one lever 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 Meta advertising
What budget do you need for Facebook advertising?
The budget is calculated from the cost per acquisition, not an absolute amount. You need roughly fifty conversions per week per ad set to exit the learning phase. A cost per acquisition of $40 therefore implies about $2,000 per week for a single ad set.
Which campaign objective should you choose?
Sales if conversion tracking is reliable and volume is sufficient. Leads if the sales cycle is long and qualification happens downstream. Traffic only to feed a retargeting audience, never to expect sales from the channel.
How many ad sets should you create?
As few as possible. Divide the account's weekly conversions by fifty to get the maximum number of ad sets able to exit learning. Below two, a multi-tier funnel structure is out of reach.
Why do Meta's numbers not match my sales?
Meta attributes a sale to an ad seen or clicked within a given window, while most accounting systems attribute it to the last point of contact. The two measures cannot match. It is better to accept the gap and follow the trend than to try to reconcile them.
Do you need to remake your ads often?
You replace the weakest angle rather than adding to the pile. An angle wears out when the system has finished measuring its useful reach, which you read from frequency rising while reach stalls, not from a fixed schedule.
Does Facebook advertising work in B2B?
Yes, and often at a lower cost per lead than LinkedIn. The condition is to qualify after the first contact rather than weighing down the form, and to accept a volume of less-qualified inquiries at the entry point.
- Meta for Business, about the learning phase, accessed July 2026.
- Meta, Conversions API documentation, accessed July 2026.
- Meta Engineering, Meta Andromeda, accessed July 2026.
- Operational thresholds and number of angles: engagement observation by Falia, revised in 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