The essentials
- The measurement traps come first: a pixel with no conversions API, duplicate events, reading platform results without GA4. They distort every decision that follows.
- On Meta: too many ad sets with too little budget, interest targeting that is too narrow, creative that is never refreshed, daily changes that restart the learning phase.
- On LinkedIn: an audience that is too narrow, audience expansion left on by default, a form that is too long, brand content where a sponsored personal post would do better.
- Common to both: no exclusion of existing customers, no dedicated landing page, a test judged after 3 days, a campaign objective chosen for its price.
SME Meta Ads and LinkedIn Ads accounts show, with a few variations, the same flaws. They do not come from a lack of talent but from habits inherited from an earlier era of the platforms, or from reading generic advice too quickly. Here are the twelve mistakes that cost the most, grouped by family, with their typical cost and the fix that does not restart the whole campaign.
The measurement traps
| Mistake | What it costs | Fix |
|---|---|---|
| 1. Pixel alone, with no conversions API (Meta) or with the Insight Tag incorrectly set up (LinkedIn) | 20 to 40% of conversions not seen by the platform, which then optimises on an incomplete signal | Install the server-side conversions API with deduplication by event identifier; see the Meta pixel |
| 2. Duplicate or badly placed conversion event | ROAS overstated by a factor of 2, budget increased on campaigns that are not profitable | Test each event with the test tool, compare with GA4 over 30 days |
| 3. Reading the platform reports only | View-through attribution at 1 day counted as a conversion; double counting between Meta and Google | Cross-check with GA4 and a multi-channel attribution report; method in conversion tracking in SMA |
Structure and audience mistakes
- 4. Fragmenting the budget: ten ad sets at €15 a day never leave the learning phase on Meta. Fix: group them into two or three ad sets with at least €50 a day each; the principles are in structuring a Meta Ads account.
- 5. Targeting too narrowly: on Meta, stacking interests gives an audience of 80,000 people at a CPM of €25; on LinkedIn, stacking job title, industry, company size and seniority produces 4,000 members that are impossible to serve. Fix: a broad audience with Advantage+ on Meta; function and seniority level rather than job title on LinkedIn, with a target of 50,000 to 300,000 members.
- 6. Forgetting the exclusions: existing customers and recent converters see the prospecting ads, which wastes 5 to 15% of the budget and irritates customers. Fix: import the customer list, exclude converters at 30 or 180 days depending on the buying cycle.
- 7. Leaving audience expansion (LinkedIn) or Audience Network placements (Meta) unchecked: delivery spreads to off-target profiles. Fix: turn off expansion at launch; on Meta, keep automatic placements but check the cost per conversion by placement after two weeks.
Creative mistakes
| Mistake | Symptom | Fix |
|---|---|---|
| 8. A single creative, or five versions of the same visual | CTR below 0.8% on Meta, below 0.4% on LinkedIn; no learning about what works | 3 to 5 genuinely different creatives (angle, format, person) |
| 9. Creative that is never refreshed | Frequency above 4, CTR falling continuously after 4 to 6 weeks | A monthly refresh schedule; keep 20% of the budget to test the new ones |
| 10. Brand content on LinkedIn where a personal post would work | Engagement rate below 0.3%, CPM rising | Sponsor the posts of a director or an expert (Thought Leader Ads) |
A pre-publication review checklist is set out in validating your ad creative.
Budget and management mistakes
- 11. Judging too early and changing too often: cutting a campaign after three days or changing the budget by 50% every day restarts Meta's learning phase and prevents any reading. Fix: seven days without intervention, then steps of 20% at most every three days; 14 days minimum before a conclusion, 21 days on LinkedIn.
- 12. Choosing the campaign objective for its cost: switching from "conversions" to "traffic" because the CPC is lower brings visitors who do not convert. The CPC falls, the cost per conversion rises. Fix: optimise towards the action you actually expect; the reasoning is developed in which objectives to choose.
The two off-platform mistakes
They cannot be seen in the ads manager and are nonetheless the most costly:
- No dedicated landing page: sending the traffic to the home page divides the conversion rate by two to three. One page per offer, with the same promise as the ad, a single call to action and a mobile load under 3 seconds.
- No known conversion value: without a margin per sale or a value per lead, it is impossible to know whether a cost per conversion of €40 is good or bad. The calculation is in ROAS in SMA.
Where to start fixing
In order: measurement (mistakes 1 to 3), without which nothing can be judged; then the exclusions and the structure (4 to 7), which are fixed in an hour; then the creative (8 to 10), which requires production; finally the management (11 and 12), which is a matter of discipline. Fixing everything at the same time restarts the learning phase and makes the before / after comparison impossible; space the workstreams a week apart and note the date of each change.
How GreenRed helps
Rather than juggling several tools, GreenRed's SMO / SMA social media module brings these metrics together in a single dashboard, compares them over time and tells you which actions come first. You can try it free, with no card, from the Pricing.
Frequently asked questions
What is the most common mistake on Meta Ads in 2026?
Fragmentation: too many campaigns and ad sets with budgets too small for the algorithm to complete its learning phase. It is often compounded by interest targeting that is too narrow. The fix, grouping them into two or three ad sets with a broad audience, improves the cost per conversion by 20 to 40% in most accounts.
What is the most common mistake on LinkedIn Ads?
An audience that is too narrow combined with audience expansion left on. The result is paradoxical: a very high CPM on the real target and stray delivery to off-target profiles. Fix it by widening the target to at least 50,000 members by function and seniority level, and by turning off expansion.
Should everything be fixed at once or gradually?
Gradually, with a week between each workstream, starting with measurement. Fixing everything at once restarts the learning phase on all the campaigns and makes it impossible to attribute the improvement to a cause. Note the date of each change so you can read the curves afterwards.
How do I know whether my provider is making these mistakes?
Ask for three things: a screenshot of the events test tool showing pixel / API deduplication, the number of active ad sets and their daily budget, and the date of the last creative put live. The answers are enough to spot mistakes 1, 4 and 9, the most common ones. The other signals are listed in the article on badly managed campaigns.