Meta Ads Manager can show you hundreds of columns, and almost every advertiser has had the same experience: yesterday's results look different today, the purchases in Ads Manager don't match Shopify or GA4, and the "results" column means something different in every campaign. Reading the data correctly is the difference between scaling a profitable campaign and switching off one that was quietly working.
This guide explains how Meta reporting is structured, which metrics matter at each level of the funnel, how attribution affects the numbers, why figures keep changing for weeks, how to use breakdowns without fooling yourself, and how to build a simple weekly report.
If you are just getting started with campaigns, our beginner's guide to Facebook ads covers setup. This post assumes you already have campaigns running.
Key Takeaways
- Always report at the level where decisions are made: campaign for budget, ad set for audience, ad for creative.
- Conversion numbers depend on the attribution setting; Meta's API defaults to a 7-day click window if none is specified, while Ads Manager uses the ad set's setting.
- Meta insights refresh about every 15 minutes and can keep changing for up to 28 days, so recent days are provisional.
- Breakdowns (age, placement, device, region) are useful but have limits; unique metrics such as reach do not work with hourly breakdowns.
- Meta and GA4 will never match exactly. Track the trend in both, and use incrementality tests to settle big questions.
How Ads Manager Data Is Structured
Meta accounts have three levels: campaigns (objective and, with campaign budget, spend), ad sets (audience, placements, schedule, optimisation event and attribution setting) and ads (creative). Totals at a higher level are de-duplicated, so the reach of a campaign is usually lower than the sum of its ad sets' reach because some people saw ads from more than one ad set.
Meta's documentation adds another subtlety: stats from deleted and archived objects still roll up into parent totals, but filters default to active objects. That is why a campaign total can be higher than the sum of the ads you see listed.
The Metrics That Matter
| Funnel stage | Metric | What it tells you | Watch out for |
|---|---|---|---|
| Delivery | Amount spent, impressions, reach, frequency | How much you spent and how many people saw ads, how often | Rising frequency with falling CTR signals ad fatigue |
| Cost efficiency | CPM (cost per 1,000 impressions) | How expensive it is to reach your audience | Seasonal spikes (e.g. Q4) affect CPM regardless of your ads |
| Attention | Hook rate (3-second video plays ÷ impressions), ThruPlays | Whether creative stops the scroll | Custom calculation; define it consistently |
| Interest | Link clicks, CTR (link click-through rate), outbound clicks | Whether the ad drives action | "Clicks (all)" includes likes and profile clicks; use link clicks |
| Landing | Landing page views, cost per landing page view | Whether clicks become loaded pages | Large gap vs link clicks suggests slow pages or tracking issues |
| Conversion | Results, cost per result, conversion rate | Leads, purchases or sign-ups attributed to ads | "Results" depends on each ad set's optimisation event |
| Value | Purchase conversion value, ROAS | Revenue attributed per unit of spend | Attributed, not incremental; compare with your own sales data |
Meta's API documentation makes the clicks distinction explicit: link clicks and all clicks are counted differently. For more on choosing headline numbers, see our guide to PPC KPIs, and for typical cost drivers, our guide to Facebook ads costs.
Attribution: Why Conversion Numbers Change
Every conversion in Ads Manager is credited according to an attribution setting – a window after someone clicks (and, optionally, views) an ad during which a conversion counts. Ads Manager reports use the setting chosen on each ad set. Meta's Marketing API behaves differently by default:
“The default behavior of the API is different from the default behavior in Ads Manager.”
— Meta for Developers, Insights API
Meta's documentation explains that if no attribution window is requested, the API uses 7-day click, and that setting the unified attribution option brings API results in line with Ads Manager. This matters if you use a dashboard tool such as Looker Studio connectors: check which attribution setting it requests, or your report may disagree with Ads Manager for reasons that have nothing to do with performance.
Practical rules:
- Use the same attribution setting across ad sets you compare.
- Use the "compare attribution settings" columns to see how many conversions came from clicks vs views.
- Be cautious with view-through conversions for prospecting campaigns; they can flatter results.
- Make sure the Pixel and Conversions API are sending reliable events; our Meta Pixel and Conversions API setup guide covers deduplication.
Data Freshness: Don't Judge Yesterday Too Soon
“Insights refresh every 15 minutes and do not change after 28 days of being reported.”
— Meta for Developers, Insights API Best Practices
In practice, conversions continue to be attributed back to the day of the ad interaction as people convert later within the window. A campaign that shows a poor ROAS yesterday may look fine in a week. For daily checks, focus on delivery and cost metrics; judge conversion metrics over at least 7 days, and lock monthly reports only once the numbers have settled. Note too that insights are reported in the ad account's time zone, which can differ from your store's.
Using Breakdowns Without Misleading Yourself
The Breakdown menu splits results by age, gender, placement, device, region, time of day and more. It answers questions such as "is Reels cheaper than Feed for us?" or "which regions convert best?"
| Breakdown | Good question to ask | Limitation |
|---|---|---|
| Age and gender | Which demographics convert at lowest cost? | Meta's delivery system already optimises; excluding segments can raise costs |
| Placement | Which placements need dedicated creative? | Low-cost placements may deliver low-quality clicks |
| Device | Is mobile conversion rate much lower than desktop? | Cross-device journeys: people often see on mobile and buy on desktop |
| Region / DMA | Where should we raise or lower budgets? | Some breakdowns are restricted for off-Meta conversion actions |
| Time of day | When do conversions happen? | Hourly breakdowns don't support reach, frequency or unique metrics |
Meta's documentation confirms that "Hourly breakdowns do not support unique fields," and that only certain breakdown combinations are allowed. Use breakdowns to generate hypotheses, then test them with a proper split test rather than cutting segments on a hunch. Our guide to Meta creative testing explains how.
How to Build a Weekly Meta Ads Report
- Create a saved column preset. Spend, impressions, reach, frequency, CPM, link clicks, CTR (link), landing page views, results, cost per result, conversion value and ROAS.
- Set the date range to the last full 7 days and compare with the previous period.
- Review at campaign level first. Where did spend go, and what did it return?
- Drill into ad sets. Compare cost per result across audiences; note any in the learning phase.
- Review ads. Rank creatives by cost per result and hook rate; flag fatigue (rising frequency, falling CTR).
- Cross-check with your own data. Compare purchases or leads with your store, CRM or GA4 trend lines.
- Write three actions. For example: shift budget, launch new creative, refresh an audience.
Accounts using automated campaign types need slightly different reporting; see our guide to Meta Advantage+ campaigns.
Why Meta and GA4 Don't Match
- Attribution models differ. Meta credits conversions after views and clicks within its windows; GA4 uses its own attribution model and only sees sessions on your site.
- Cross-device and in-app browsing. Meta can connect logged-in users across devices; GA4 often cannot.
- Consent and ad blockers. Each tool loses different data.
- Time zones and dates. Meta reports conversions against the ad interaction date in the ad account's time zone.
Neither number is "the truth". Use both directionally, rely on your own sales data for totals, and run incrementality tests when you need to know what ads really add.
Common Mistakes
- Judging yesterday's ROAS. Late conversions will still arrive.
- Using "Clicks (all)" as traffic. Use link clicks and landing page views.
- Comparing ad sets with different optimisation events or attribution settings.
- Cutting audiences from a single breakdown. Test before excluding.
- Ignoring frequency. Fatigue shows up there before it shows up in cost.
- Reporting attributed revenue as profit. Factor in margins, returns and incrementality.
Related Guides
- How to Build Marketing Reports Clients Actually Read
- How to Set Up Google Analytics 4 on WordPress
- Share of Search: A Leading Indicator of Market Share
Frequently Asked Questions
Why do my Meta Ads results change after a few days?
Conversions are attributed back to the day someone interacted with the ad, so they keep arriving within the attribution window. Meta says insights stop changing 28 days after being reported.
What is the difference between clicks and link clicks?
Clicks (all) includes any click on the ad, such as likes, comments and profile clicks. Link clicks only count clicks that lead to your chosen destination.
Why doesn't Meta Ads Manager match GA4?
They use different attribution models, data sources and counting rules. Meta includes view-through and cross-device conversions; GA4 only sees sessions on your site that it can track.
Which attribution setting should I use?
Use one setting consistently for the campaigns you compare. Many advertisers use click-based windows for prospecting to avoid over-crediting views. Check that your reporting tools request the same setting.
How often should I report on Meta ads?
Check delivery daily, review performance weekly over the last full 7 days, and finalise monthly reports once the 28-day update period has passed or clearly label recent data as provisional.
Turn Ads Manager Data Into Better Decisions
Reading Meta reports well means knowing which numbers are provisional, which are attributed rather than incremental, and which level each decision belongs to. If you want an experienced team to audit your Meta account, fix tracking and build reporting you can trust, our PPC management service can help. Request a free quote to get started.
