Free tool + guide
FB ads analytics is the practice of reading Meta campaign performance to decide what to run next. Ads Manager gives you the numbers for free. The hard part is knowing which of them predict anything, reading them per creative, and turning the answer into a brief rather than a screenshot. Paste your figures into the free tool below and it will tell you where the funnel is actually losing.
Enter the core figures from Ads Manager and the tool derives your full funnel, benchmarks click-through and cost per click against published averages, works out your break-even return from your margin, and gives you one plain-English read on where the money is actually being lost. Everything runs in your browser. Nothing is uploaded and no account is needed.
One campaign, one date range. Spend, impressions, and link clicks are required; the rest sharpen the diagnosis.
Your derived click-through rate, cost per click, CPM, conversion rate, cost per purchase, and return on ad spend, each rated with the benchmark it was judged against, plus one sentence naming whether you have an attention problem, a conversion problem, or a unit-economics problem. Load the sample data to see a full run.
Want to know why these numbers look the way they do? The tool tells you where the funnel is losing. Selzee tells you what to say instead, using your own customer language.
Facebook ads analytics is the collection and interpretation of performance data from Meta advertising, covering both Facebook and Instagram placements. The primary source is Meta Ads Manager reporting, which exposes over a hundred columns across campaign, ad set, and ad level, along with breakdowns by placement, device, and demographic.
It matters because the alternative is spending on guesswork, and guesswork costs more every year: Facebook's average cost per lead climbed 21% year over year to $27.66 in 2025, per WordStream's benchmarks. Since Nielsen's 2017 creative-effectiveness research credits creative, not targeting or media spend, with up to 89% of a digital ad's in-market success, the point of reading the account is to work out which argument to make next. Analytics that end in a screenshot have done half a job.
Ads Manager exposes over a hundred columns. These are the ones that change a decision, grouped by the funnel stage they belong to.
Whether the creative earned attention. Thumbstop, the share of impressions that became a three-second view, is the purest read on the hook, and everything downstream is capped by it. A weak opening cannot be rescued by a bid strategy.
What attention costs you. Rising CPM with steady CTR is auction pressure; rising CPM with falling CTR is your creative losing relevance. The two together diagnose more than either alone.
Whether the landing page and offer hold up the promise the ad made. A healthy click-through with a poor conversion rate is never a creative problem, and rewriting the hook makes it worse by sending more people to the same drop-off.
The numbers the account is judged on. Both are only meaningful against your own margin: a 2x return is excellent on one product and loss-making on another, which is why break-even ROAS belongs on the same screen.
Every metric above should be read per creative. Campaign-level totals answer whether the account is fine and never answer which idea worked, which is the only question that changes what you build next.
The same creative often performs very differently across feed, stories, and Reels, and across the two platforms. Judge before you split, because a blended average can hide one placement carrying the result and another quietly wasting budget.
Reach, impressions, and raw engagement describe activity rather than results. They rise with budget and tell you nothing about whether the argument worked. Frequency read in isolation is the same trap: on its own it says an audience has seen an ad several times, which is only meaningful next to hook rate, where the two together identify fatigue. Video views without a matching hold rate flatter creative that opened well and lost people immediately. The test for any metric in a weekly review is simple: if a number cannot change what you build next week, it does not belong in the review. Most reporting fluff survives because it is easy to pull and reliably goes up.
Ads Manager reports at campaign, ad set, and ad level, and only the last one answers a creative question. Campaign totals tell you the account is healthy or not; ad-level rows tell you which specific idea did it. Set the date range and attribution window before you look, not after, because adjusting the window once you have seen a number is how a losing creative gets kept.
Cold prospecting is judged on click-through and cost per click, because the job is earning attention. Conversion campaigns are judged on cost per acquisition and return on ad spend. Grading a prospecting ad on ROAS punishes it for a job it was not doing, and is the most common way good top-of-funnel creative gets killed early.
One blended number hides three different stories. The same creative frequently performs well in feed and badly in stories, and a weak audience can drag down an ad that is working. Break down all three before concluding anything about the creative itself.
Published benchmarks are averages across industries and objectives, useful as a sanity check and useless as a target. The comparison that matters is this creative against your current control, on the same placement and audience. Custom tagging by angle, format, and offer is what makes that comparison possible at all.
Line up your top and bottom performers and look at what each one argued: the hook, the angle, the format, the objection it answered. Patterns show up fast, and they are the only output of an analytics review that changes what you make next. This is the step where most reviews stop, because it needs creative reasoning rather than a dashboard.
Scale when a creative has cleared a spend or conversion threshold you agreed in advance and is holding above your break-even, not when it looks good on day two. Then end the review with one specific next test, a new hook, a new angle, a new format, per creative. A review that produces observations instead of decisions is a report.
The tool above tells you where the funnel is losing. Selzee tells you what to say instead, by reading the reviews and comments your customers already wrote.
Request a DemoThree tiers, solving three different problems. Most teams need the first, many add the second, and the third is a different job entirely.
| Dimension | Meta Ads Manager | Third-party reporting | Selzee |
|---|---|---|---|
| Cost | Free with an ad account. | Subscription, usually priced on ad spend or seats. | Credit-based subscription. |
| What it is good at | Every number Meta holds, at every level, with breakdowns. | Cross-channel joins, scheduled refreshes, visual creative reporting. | Deciding what the next creative should argue. |
| Creative-level view | Possible, but tedious. Tagging is manual and easy to abandon. | The main selling point: creative tagging and top-performer reports. | Maps results back to the claim each ad made. |
| Answers "why" | No. It reports what happened. | Partly. It shows which creative won, not why it resonated. | Yes, from customer language rather than from the ad account. |
| Integration | Meta only. | Usually Meta, TikTok, Google, and a warehouse or sheet. | Meta and TikTok Ads plus Shopify, Klaviyo, GA4, and review platforms. |
| Choose it when | You are starting, or your reporting need is one platform. | You run several channels and tagging by hand has stopped scaling. | You know which creative lost and not what to make instead. |
The honest sequence is native first, third-party when manual tagging breaks down, and a research layer when the bottleneck stops being measurement and becomes direction. Buying a dashboard before you have agreed which metrics you act on renders the same confusion at higher resolution.
A dashboard tells you a metric moved. It cannot tell you why, and the same symptom has several possible causes. A falling click-through rate could be a weak hook, a mismatched audience, or fatigue on an angle that worked for six weeks. Each has a different fix, and teams routinely pick the wrong one because the data alone does not distinguish between them. The most expensive version of this is treating a messaging problem as a targeting problem, which produces months of audience experiments while the argument stays wrong.
The missing input is customer language. Pair a performance signal with what buyers actually say in reviews, comments, and survey responses, and the diagnosis becomes specific: the objection the ad never answered, the desire it under-sold, the claim buyers do not believe without proof. Sarah Levinger, Founder of Tether Insights, describes the habit that separates teams doing this well: “I wish more brands would use the data they have and start analyzing for emotional insights instead of just transactions.” This is where spreadsheets and generic AI tools fall short: they analyze performance or feedback, and almost never both against each other.
A worked example. A skincare brand has a creative that ran at a 3.4x return for five weeks and has slid to 1.6x. Frequency is up, hook rate is down, and conversion rate is unchanged, so the funnel is intact and the argument has gone stale rather than the offer failing. Selzee reads the comment thread on that ad and the reviews from the same period, and the pattern is clear: the ad leads on results in fourteen days, while the recent reviews and the comments both keep raising the same unanswered question about whether it suits sensitive skin. That is a messaging gap rather than a fatigue problem, and it is invisible in Ads Manager because no column tracks what buyers are asking. The output is a brief: lead on suitability rather than speed, answer the sensitivity objection in the opening line, use the customer phrasing that already appears in the reviews, and test it as a single variable against the current control. Performance signal, customer insight, new angle, brief, test, compressed from a research week into a session.
Fix: agree a minimum spend or conversion threshold per creative before launch and hold to it. Early leaders reverse constantly at low volume.
Fix: never grade a prospecting ad on ROAS or a conversion campaign on reach. Different objectives optimise for different things and are not comparable.
Fix: always read CTR alongside downstream conversion. A high CTR on the wrong promise sends more people to bounce, and it looks like success until the cost per acquisition arrives.
Fix: watch frequency and hook rate together. Rising frequency with falling hook rate is fatigue, and it appears days before cost per acquisition moves.
Fix: with broad targeting and automated delivery, the creative is doing the qualifying. Most dips are messaging problems wearing a targeting costume.
Fix: reach, impressions, and engagement without conversion describe activity rather than result. If a number cannot change what you build next, it does not belong in the review.
Any tool that collects and interprets Meta campaign performance so you can decide what to run next. Meta Ads Manager is the native version and it is free. Third-party tools add cross-channel joins, scheduled refreshes, and creative tagging that Ads Manager makes tedious rather than impossible.
WordStream's 2025 benchmarks put average click-through at 1.71% and cost per click at $0.70 for traffic campaigns across all industries. Treat those as sanity checks rather than targets, since they are all-industry averages. ROAS has no useful universal figure at all: judge it against the break-even implied by your own gross margin, which the tool above calculates.
Weekly for creative decisions, monthly for account trend. Fatigue moves faster than a monthly cadence, so a month-end review reports damage rather than preventing it. Checking daily is worse than weekly, because it invites reacting to noise before any creative has enough volume to judge.
Account level tells you whether the business is healthy. Ad level tells you which specific idea produced that outcome. Only the second changes what you make next, which is why every creative decision should start from an ad-level breakdown rather than a campaign summary.
The data pull, yes. Third-party platforms schedule refreshes and maintain creative tagging automatically, which removes most of the tedium. The judgement about what a pattern means and what to test next is not automatable from performance data alone, because that answer lives in customer language rather than in the ad account.
Look at frequency and hook rate together on the same creative. Rising frequency with falling hook rate, while conversion rate holds steady, is fatigue: the audience has stopped hearing the argument. If conversion rate fell too, the problem is further down and is more likely the offer, the page, or a genuine audience mismatch.
It supplies the "why" that performance data structurally cannot. Analytics identify which creative lost; reviews and comments explain which objection went unanswered or which claim buyers did not believe. Pairing the two turns a diagnosis into a brief, which is the difference between knowing an ad failed and knowing what to build instead.
Analytics tell you what happened. Selzee turns those performance signals, together with your reviews and comments, into ready hooks, briefs, and messaging angles, so every dip or spike becomes a clear next test rather than another guess. Bring last month's results and a review export and we will run it on your own account.
Keep exploring: ad performance analytics, Facebook ads reporting software, the free ad analysis tool, beating Facebook ad fatigue, creative tracking, ad testing tools.