The essentials
- The three most serious signals: no conversion measured (or false conversions), a campaign objective that does not match the intended action, a frequency above 5 per week.
- Where to look: the Events Manager, the results columns by ad set and by creative, the frequency, the budget allocation, and the CRM for lead quality.
- The thirty-minute test: twelve yes/no questions, applicable to your account or to the one run by a provider.
- The fix: most signals are settled within a week; measurement first, then the objective, then the exclusions and the creative.
A badly run SMA campaign is not always visible at first glance: the dashboard shows clicks, reach, sometimes “results”. The signs of bad management are elsewhere: in the measurement, the objective, the frequency, the exclusions, the quality of the leads and the nature of the report. This article lists twelve signals, indicates where to read them and what to fix. It serves to audit your own account or to judge a provider; the contractual traps are in the SMO provider traps.
The twelve signals
| # | Signal | Where to see it | What it reveals | Fix |
|---|---|---|---|---|
| 1 | No conversion measured, or conversions with no value, or double what GA4 says | Events Manager, Results columns, GA4 | Pixel missing, badly installed or duplicated; no conversions API | Rebuild the measurement before anything else; see conversion tracking in SMA |
| 2 | “Traffic” or “engagement” objective for a campaign meant to generate leads or sales | Campaign level | The algorithm finds clickers, not customers | Recreate it as Leads or Sales; see social ads objectives |
| 3 | Frequency above 5 per week (or 10 over 30 days) in acquisition | Frequency column by ad set | Audience too small or budget too high; wear-out, rising cost | Widen, cap, refresh the creative |
| 4 | A single creative per ad set, unchanged for more than 6 weeks | Ad level, dates | Creative fatigue, no testing | 3 to 6 variants, monthly refresh; see high-performing images and videos |
| 5 | Customers and converters not excluded from the acquisition campaigns | Ad set exclusions | Budget spent retargeting buyers; artificially good ROAS | Cascading exclusions |
| 6 | Daily changes to budgets, bids or audiences | Change history | Learning restarted constantly, unstable results | A weekly cycle, no change of more than 20% at once |
| 7 | Ad sets that never leave the learning phase | Delivery column (“limited learning”) | Too many ad sets for the budget, event too rare | Group them, choose a more frequent event, raise the budget per ad set |
| 8 | Cheap leads but never reachable or off target | CRM, feedback from the sales team | Native forms without qualification, Audience Network, targeting too broad | Qualification questions, exclusion of the weak placements, dedicated page |
| 9 | No UTM on the links, Meta traffic invisible or classed as “direct” in GA4 | Ad URLs, GA4 | Impossible to link the campaigns to the site and to the conversions | Systematic UTMs; see sources, mediums and UTMs |
| 10 | Report limited to reach, impressions, clicks and CPM | The monthly report | Vanity metrics; nobody knows what the campaign brings in | Cost per qualified lead, cost per purchase, ROAS, share of new customers; see the SMA KPIs |
| 11 | Budget concentrated on retargeting (more than 40%) or on the brand | Budget allocation by campaign | The account recycles the same people and no longer wins new ones | Rebalance towards acquisition with lookalike audiences and Advantage+ |
| 12 | Ad account, pixel or page owned by the provider | Business portfolio settings | Loss of the assets and of the history at the end of the contract | Immediate transfer of ownership; see Meta Business Suite |
The three most serious signals, in detail
False or missing measurement
This is the root signal: every other setting relies on the measured conversions. Check in the Events Manager that the key events (Lead, Purchase) come through with their value, by browser and by server, without duplicates, and that their volume is consistent with GA4 or the CRM to within 10 or 20%. A gap of double in either direction points to a duplicated pixel, an event fired on the button click rather than on successful submission, or a missing conversions API.
The unsuitable objective
A “traffic” campaign that should be selling shows a flattering cost per click and an invisible cost per purchase because nobody looks at it. The test: which result is the campaign optimising (Results column) and is it the action you want? If the answer is “clicks” or “engagement” for a lead or sales campaign, the campaign has to be recreated.
Excessive frequency
Above 5 exposures per week in acquisition, the cost per result rises, negative comments appear and the brand image suffers. The cause is almost always an audience too small for the budget; the fix consists in widening it (lookalike audiences, automation), capping it where the platform allows, and refreshing the creative faster.
The thirty-minute test
- Open the Events Manager: do the key events come through with a value, by browser and by server? (Signal 1)
- List the campaigns with their objective and their optimised result. (2)
- Add the Frequency, Cost per result and Delivery columns, over 30 days, by ad set. (3, 7)
- Open five ads at random: creation date, number of variants per ad set. (4)
- Open the exclusions of two acquisition ad sets. (5)
- Check the change history over 30 days. (6)
- Ask the sales team about the quality of the last ten leads from social media. (8)
- Click on three ads and check the landing URL for the UTMs. (9)
- Reread the last report received. (10)
- Budget allocation by campaign: acquisition, retargeting, brand. (11)
- Portfolio settings: owners of the assets. (12)
Count the signals present: beyond three, the campaign calls for a full overhaul, in the order measurement, objective, exclusions, creative. The complete audit approach is described in running a social ads audit.
What is not a signal
- A CPM or a CPC higher than the market average: it depends on the objective, the target and the season; only the final cost per result counts.
- A Meta ROAS different from GA4: the attribution models differ; a gap of 20 to 50% is normal; see data-driven attribution.
- Daily variations in cost per result: normal noise; judge over a rolling 7 days.
- A seasonal drop: compare with the previous year.
How GreenRed helps
Rather than juggling several tools, GreenRed's social media tracking 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 first sign of a badly run social ads campaign?
Conversion measurement that is missing, false or inconsistent with GA4: without it, the algorithm optimises on noise and the report cannot say what the campaign brings in. Check in the Events Manager that the key events come through with their value, by browser and by server, without duplicates.
What frequency is too high in social advertising?
Above 5 exposures per week per person in acquisition (around 10 over 30 days), the cost per result rises and the brand image suffers. In retargeting, 3 to 5 per week is acceptable over a short period. A high frequency signals an audience too small for the budget or creative that has not been refreshed.
How do you judge a social ads provider's management?
Apply the twelve signals to their account: measurement, objectives, frequency, creative, exclusions, stability of the settings, lead quality in the CRM, UTMs, the nature of the report, budget allocation, ownership of the assets. Beyond three signals present, the management is failing; ownership of the assets (signal 12) has to be fixed before any discussion.
Is a high cost per click a sign of bad management?
No: CPC depends on the objective (a sales campaign has more expensive clicks than a traffic one), on the target and on the season. A well-run campaign can have a high CPC and an excellent cost per purchase. Judge on the final cost per result and on ROAS, never on CPC or CPM.