Meta can be a shop’s most effective and most overrated channel at the same time, depending on who is counting the results. The dashboard has its own windows, its own events and modelling; the shop has paid orders. This guide sets out how Meta counts conversions, when frequency and creatives start working against you and how to judge this channel on real sales.
Counting rules
How Meta counts results: windows, omni and modelling
Before you look at any ROAS from the dashboard, you need to know three mechanisms by name:
| Mechanism | How it works | Effect on the numbers |
|---|---|---|
| Attribution windows | a purchase counts after a click (typically up to 7 days) and after a mere view (typically 1 day) | the dashboard adds purchases without a click; Google counts the same transactions too |
| Omni events | the omni prefix sums events from all conversion locations according to Meta attribution | omni_purchase is often higher than purchase; mixing the two spoils comparisons |
| Modelling | when consent is missing, some conversions are estimated statistically | dashboard figures are partly an estimate, not a count |
Then there is what the dashboard does not know: order statuses (cancelled orders count as successes) and the cost of goods. That is why Meta results are read on two levels: platform metrics for optimising creatives, the shop’s total orders and MER for budget decisions. The mechanics of discrepancies between dashboards and the shop are covered in the guide to conversions and attribution.
Settling results
Meta against shop sales: panel ROAS, GA4 and MER
The calculation is simple once the data is connected. You take Meta spend and ROAS from the platform, put the revenue GA4 credits to Meta alongside as an independent estimate, and add the MER of the whole shop: revenue from paid shop orders divided by total spend across all platforms. An example from a jewellery shop (illustrative data, 30 days):
| Measure | Value | What it really says |
|---|---|---|
| ROAS according to the Meta dashboard | 7.4 | conversions by windows and modelling, without statuses |
| ROAS according to GA4 | 4.8 | revenue GA4 credits to Meta / Meta spend; an estimate from a second system |
| MER of the whole shop | 6.1 | revenue / total spend across all platforms |
The gap between 7.4 and 4.8 does not mean Meta is lying: the two systems count different things. The problem only starts when budgets are scaled on 7.4. A healthy rhythm is a weekly window in which you look at three numbers together: Meta spend, ROAS according to GA4 and the shop’s MER; if spend grows while MER stays flat, the growth exists only on paper.
The Meta dashboard optimises campaigns. The shop’s total sales and MER settle the budget. Confusing these roles costs the most when scaling.
Reach
Frequency: when an ad starts to do harm
Frequency tells you how many times, on average, one person sees your ad. In itself it is not bad; it only does harm when it drifts away from the freshness of the creative. A typical picture from daily data looks like this (illustrative data):
CTR by frequency · same creative · 30 daysjewellery shop · illustrative data
Prospecting and remarketing have different norms: cold audiences need only a low frequency, while remarketing tolerates a higher one. The warning sign is the trend, not the number: frequency rises, CTR falls and cost per result goes up.
Creatives
Creative fatigue: the ad life cycle in daily data
Every creative has a life cycle, and it shows in the data before results collapse. The weekly timeline of a single ad usually looks like this:
Cost per result · one creative · 6 weekslower is better
Rotation works ahead of time: new variants are prepared before the current ones wear out, and swapped in when the downward trend persists despite stable reach. Waiting for complete burnout costs a week or two of expensive results.
Real risk
Scaling the budget on a fatigued creative speeds up its wear. A higher budget raises frequency, frequency drags CTR down, and the algorithm compensates with more expensive impressions. The order is the reverse: fresh creatives and audience first, budget second.
Methodological note: engagement metrics (CTR, frequency, reach) come from Meta and are used to diagnose creatives; they are not converted directly into sales. Whether a campaign adds sales is judged on the shop’s total orders and MER, ideally with the on-off test described in the guide to attribution.
DataOrganizer
How DataOrganizer sets Meta beside GA4 and MER
DataOrganizer pulls the full data hierarchy from Meta (campaigns, ad sets and ads with spend, reach, frequency and granular actions, including omni) and sets it side by side, day by day, with GA4 data, spend on other platforms and the shop’s total paid orders and margin. As a result, the same questions that required three reports above fit into a single chat:
The three largest campaigns: Meta dashboard compared with GA4:
| Campaign | Spend | Platform ROAS | ROAS per GA4 | Frequency |
|---|---|---|---|---|
| Remarketing 14 days | €1,550 | 9.1 | 5.6 | 4.1 |
| Broad prospecting | €2,100 | 5.8 | 4.4 | 1.8 |
| Prospecting, spring creative | €975 | 4.6 | 2.1 | 5.3 |
The third campaign combines the lowest return according to GA4 (2.1) with a frequency of 5.3 and falling CTR: classic creative fatigue. Recommendation: swap the variants before any budget change, and direct scaling to broad prospecting, where ROAS per GA4 of 4.4 holds at low frequency. After the change, watch the shop’s total MER: if it does not move, the growth exists only on paper.
Key takeaway
Two levels, two sets of metrics. CTR, frequency and cost per result diagnose creatives inside Meta; ROAS according to GA4 as an estimate and the shop’s total MER decide budgets. DataOrganizer keeps both levels in one model, so neither poses as the other.
FAQ
Frequently asked questions about Meta Ads for shops
Why is ROAS in the Meta dashboard higher than in reality?
Because the platform counts with its own windows (click and view), models some conversions when consent is missing and does not know order statuses, while Google claims the same transactions in parallel. The reference point is the shop’s paid orders in total and MER, with GA4 as an estimate per channel; platform ROAS remains a tool for optimisation within the platform.
What does omni_purchase mean in Meta data?
The omni prefix sums events from all conversion locations according to Meta’s multichannel attribution. For a shop, the key pairs are purchase/omni_purchase and add_to_cart/omni_add_to_cart. Stick consistently to one event type in comparisons, as mixing them inflates some rows.
What ad frequency on Meta is safe?
There is no single number: prospecting works at a low frequency (around 1–3 per week), while remarketing tolerates a higher one. The warning sign is the trend: frequency rises, CTR falls and cost per result goes up. Response: fresh creatives, a broader audience or a lower budget, in that order.
How can you recognise creative fatigue?
By a simultaneous rise in frequency and fall in CTR, together with a growing cost per result under unchanged settings, persisting week after week. The standard is proactive rotation: new variants ready before the current ones wear out.
How should Meta campaigns be judged on real sales?
On two levels: Meta metrics for optimising creatives and audiences, and the shop’s total orders and MER for budget decisions. When in doubt about how much the channel really adds, the fairest method is an on-off test while observing sales across the whole shop.
Cheat sheet for the whole guide
Three rules for working with Meta Ads
1. Know how the counting works. Attribution windows, omni events and modelling explain why the platform sees more than the shop. This is essential knowledge before any discussion of results.
2. Optimise in the platform, decide budgets on shop totals. CTR, frequency and cost per result guide creative work; ROAS according to GA4 and the shop’s total MER decide budgets. Scaling on platform ROAS is the most expensive habit in this channel.
3. Creatives age faster than strategies. The ad life cycle shows in daily data: rotate ahead of time and raise the budget only on fresh variants. DataOrganizer keeps Meta campaign data, GA4 and the shop’s total sales side by side, day by day, so these decisions take minutes.
DataOrganizer · Meta set against MER
Connect Meta, GA4 and your shop: the campaign hierarchy, frequency and omni actions are set side by side with GA4 and the shop’s total paid orders and MER. A demo account is available straight away, and you only pay for the sources you connect.