If you've read Guide #3 (How to read your Google Ads data without wasting money), it's worth understanding the fundamental difference between the two platforms before reading any further — because the difference changes almost everything about how you should read the numbers.
On Google Ads, someone is actively searching for something. They have a need and they're typing words into a search bar to solve it. Your ad meets them at the moment of intent. On Meta — Facebook and Instagram — nobody is searching for anything. They're scrolling through photos of friends, news, and entertainment. Your ad has to interrupt that scroll and earn their attention with no prompting from them at all.
This single difference explains why the metrics that matter, the typical performance benchmarks, and the failure patterns are completely different between the two platforms. A Google Ads campaign that stops converting is usually a relevance or budget problem. A Meta campaign that stops converting is very often a creative fatigue problem — the audience has simply seen the same ad too many times and stopped paying attention. Section 4 covers this in detail because it is the single most common and most overlooked cause of declining Meta performance.
A quick scope note: this guide covers Meta's Conversion campaigns — the most common objective for businesses trying to generate leads or sales. Awareness and engagement campaigns, and the newer Advantage+ shopping campaigns, have different performance logic and will be covered in later guides if there's demand for them.
Intent versus interruption — why this matters for your numbers: Because Meta interrupts rather than captures intent, typical click-through rates, conversion rates, and costs per result are usually higher (worse) on Meta than on Google Search for the same business. This is not a sign Meta is performing badly — it's the nature of the channel. Compare Meta's performance against its own historical trend, not against your Google Ads numbers.
Meta Ads Manager organises everything into three levels: Campaigns (the overall objective and budget), Ad Sets (the audience, placement, and budget allocation within a campaign), and Ads (the actual creative — the image, video, or carousel someone sees). Performance can be reviewed at any of these three levels, and knowing which level to look at for which question is the first skill to build.
Use the Campaign level to see overall spend, results, and cost per result against your objective. This is your weekly health check.
Use the Ad Set level to compare how different audiences are performing against each other. If you're testing multiple audiences within the same campaign, this is where you see which one is working.
Use the Ad level to compare individual creative — which specific image, video, or piece of copy is performing best. This is also where you'll spot creative fatigue first, because fatigue happens to a specific ad, not to an entire campaign.
Two things to set up before you do anything else:
First, customise your columns. Meta's default view shows a generic set of columns that often doesn't include the metrics that matter most for your business. Click "Columns" in the top right of any reporting table, choose "Customize Columns," and build a saved preset that includes: Results, Cost per Result, CPM, CTR (link click-through rate), Frequency, and Amount Spent. Save this as a custom preset so it's there every time you open the platform.
Second, understand the Attribution setting. By default, Meta attributes a conversion to an ad if someone clicked the ad and converted within 7 days, or viewed the ad (without clicking) and converted within 1 day. This is a generous attribution window compared to GA4, and it means Meta's own reported conversions will almost always be higher than what GA4 shows for the same campaign — similar to the Google Ads versus GA4 discrepancy covered in Guide #3. Cross-reference Meta's numbers with GA4 before making budget decisions.
The view-through trap: Meta counts a conversion if someone simply saw your ad (without clicking) and converted within 1 day. This is called a "view-through conversion." On a platform where people scroll past hundreds of posts a day, attributing a sale to someone having merely seen an ad they didn't engage with is a generous assumption. If a campaign's reported results rely heavily on view-through conversions rather than click-through ones, treat the real-world impact with more scepticism than the dashboard suggests.
1
Campaign performance over time
This is your weekly pulse check. It shows spend, results, and cost per result trended over time rather than as a single blended number.
Where to find it: Ads Manager → select your campaign → toggle the chart view above the table → set the breakdown to "By day."
What to look for: A cost per result that's flat or improving over a 4-week trend is healthy. A cost per result that's been climbing for three or more consecutive weeks is a signal to move to Report 2.
2
Ad-level performance with Frequency
Frequency tells you, on average, how many times each person in your audience has seen your ad. This is the single most useful early-warning metric on Meta and it has no real equivalent on Google Ads, because Google Ads doesn't show the same ad to the same person over and over in the way Meta's algorithm does.
Where to find it: Ads Manager → Ads tab → add the "Frequency" column.
What to look for: A frequency above 3 for a cold (new) audience usually means people have seen your ad enough times that it's losing impact. A frequency above 5–6 for any audience is a strong signal of fatigue. Section 4 covers this in full detail.
3
Breakdown by placement
Meta shows your ads across multiple places — the Facebook feed, Instagram feed, Instagram Stories, Reels, and Audience Network (other apps showing Meta ads). Performance often varies significantly by placement.
Where to find it: Ads Manager → select campaign or ad set → click "Breakdown" → "By Delivery" → "Placement."
What to look for: If one placement has a dramatically higher cost per result than others, you can exclude it. Audience Network in particular often produces low-quality clicks at a low cost — cheap but not valuable. A low cost per result on a placement you've never checked is not automatically good news.
4
Audience overlap
If you're running multiple ad sets targeting different audiences within the same campaign or account, those audiences can overlap — meaning the same person is eligible to see ads from more than one ad set. When this happens, you're effectively bidding against yourself in Meta's ad auction.
Where to find it: Ads Manager → Audiences (left-hand menu, under "All tools") → select two or more saved audiences → click "Actions" → "Show Audience Overlap."
What to look for: Overlap above 20–30% between two active ad sets targeting similar goals is worth addressing — either by combining the ad sets, or by excluding one audience from the other.
5
Cost per result trend by ad set
This compares how different audiences (ad sets) within the same campaign are performing against each other, so you can shift budget toward what's working.
Where to find it: Ads Manager → Ad Sets tab → sort by "Cost per Result," with the date range set to a rolling 4-week window.
What to look for: If one ad set has a meaningfully lower cost per result than others over a sustained period, that's where Meta's algorithm believes your best opportunity lies. Consider increasing its budget and reducing or pausing consistently underperforming ad sets.
Start with the Frequency column: If you only check one number on Meta each week, make it Frequency at the ad level. It's the earliest, cheapest signal that an ad is running out of road — and it's something Google Ads simply doesn't have an equivalent for. Catching rising frequency early lets you refresh creative before performance actually drops, rather than reacting after the damage is visible in cost per result.
Creative fatigue is what happens when your audience has seen the same ad enough times that they've stopped noticing it. Unlike a Google Ads keyword, which keeps performing as long as people keep searching that term, a Meta ad's performance has a natural lifespan — even if nothing else about your campaign changes.
Why this happens: Meta shows your ad repeatedly to the same group of people as part of how its delivery algorithm works. The first few times someone sees an ad, it has the best chance of catching their attention. By the tenth or fifteenth time, it blends into the background — what's sometimes called "banner blindness." Their attention has worn out, not your offer.
The three signals of creative fatigue, in the order they usually appear:
What to do when you spot fatigue: Refresh the creative. This doesn't always mean creating something entirely new — even a different image with the same message and offer is often enough to reset performance, because it gives the algorithm and the audience something new to respond to. Many advertisers maintain 3–5 ad variations running simultaneously specifically to rotate in fresh creative before fatigue sets in, rather than reacting after performance has already dropped.
The "it just stopped working" trap: One of the most common Meta support questions is some version of "this ad was working great and then it just stopped." In the large majority of cases, this is creative fatigue, not a sudden change in audience interest or a platform problem. Before assuming something external has gone wrong, check whether frequency has been climbing for the two to three weeks before the drop. If it has, the ad simply ran its natural course — refresh the creative rather than troubleshooting the targeting or the offer.
Meta's targeting works differently from Google Ads' keyword-based system. Rather than matching specific search terms, you define an audience using a combination of demographics, interests, behaviours, or — increasingly common and often more effective — by uploading a list of your existing customers (a "Custom Audience") or asking Meta to find people similar to them (a "Lookalike Audience").
Custom Audiences are built from data you provide — your customer email list, people who've visited your website (via the Meta Pixel, a tracking snippet similar in purpose to a GA4 tag), or people who've engaged with your Facebook or Instagram content. These tend to perform well because they're built from people who already have some relationship with your business.
Lookalike Audiences ask Meta to find new people who share characteristics with an existing Custom Audience — for example, people similar to your best customers. The quality of a Lookalike Audience depends heavily on the quality and size of the source audience it's built from. A Lookalike built from 50 customers will generally underperform one built from 500.
Broad targeting — letting Meta's algorithm find your audience with minimal restrictions, rather than you defining detailed interests and demographics — has become increasingly effective as Meta's algorithm has improved, particularly when you have strong conversion data feeding back into the system (similar to the principle covered for Google's broad match in Guide #3). For businesses with limited conversion history, narrower, more defined audiences typically perform more predictably while the algorithm learns.
Audience saturation happens when you've been targeting the same audience for an extended period and have essentially shown your ads to everyone in that group who's likely to be interested. This shows up as a combination of rising frequency, rising CPM, and a shrinking number of new people being reached — visible in the "Reach" metric flattening even as spend continues.
Audiences too small to learn from: Meta's delivery algorithm needs a reasonable volume of data to optimise effectively. An audience smaller than roughly 1,000 people, or a campaign generating fewer than 15–20 conversions per week, often won't give the algorithm enough signal to perform well. If a small, highly specific audience is underperforming, the issue may be size rather than relevance — consider broadening it or combining it with a related audience.
As with Google Ads in Guide #3, the goal here is to read multiple signals together rather than reacting to any single metric in isolation.
A campaign is healthy if:
Cost per result is stable or improving on a 4-week trend. Frequency is below 3–4 for the audience size and campaign age. CTR is stable rather than declining. Reach is still growing or holding steady, not flattening. Results are at or above your monthly target.
A campaign needs attention if:
Frequency has climbed steadily for two or more weeks. CTR has dropped by 20% or more from its early performance on the same ad. CPM has been rising for three or more consecutive weeks without a clear seasonal explanation (Meta CPMs do rise seasonally, particularly around major shopping periods). Audience overlap above 30% exists between active ad sets with similar goals.
A campaign is likely wasting money if:
Frequency is above 6 with no creative refresh in the last two to three weeks. Cost per result has more than doubled your target across four or more consecutive weeks. The audience is small (under 1,000 people) and has been running for more than four weeks with poor results. A large proportion of results are view-through conversions (covered in section 2) rather than genuine engagement.
The most common Meta waste pattern is simple: a single ad creative is left running for too long. Unlike Google Ads, where a keyword can perform consistently for months or years, Meta creative has a natural shelf life measured in weeks, not months, for most cold-audience campaigns. The fix isn't more spend — it's a refreshed asset.
The "set it and forget it" trap, Meta edition: As in Google Ads, Meta's automated bidding and delivery systems optimise toward whatever signals they're given — but Meta's creative has a shorter natural lifespan than a Google Ads keyword. An account left completely unmanaged on Meta will fatigue and quietly waste budget faster than an equivalent unmanaged Google Ads account. If you can only commit to reviewing one paid channel weekly, both deserve attention, but Meta's creative fatigue cycle makes it less forgiving of long gaps between check-ins.
Work through this in order before making significant changes.
Check Frequency first
Before anything else, check whether frequency has been climbing for the audience and ad in question. If it has climbed past 3–4 in the weeks leading up to the drop, you very likely have creative fatigue. Refresh the creative before investigating anything else.
Check your conversion tracking
As with Google Ads, confirm the Meta Pixel (or Conversions API, a more reliable server-side tracking method) is still firing correctly. Go to Events Manager (in the main Meta Business Suite menu) and check whether your key conversion events show recent activity. A website update that breaks the Pixel is a common, easily missed cause of an apparent performance drop that's actually a tracking problem.
Check for audience overlap or saturation
Review the Audience Overlap report (section 3) if you're running multiple ad sets. Also check the Reach metric over time — if Reach has flattened while spend has stayed the same or increased, you may be exhausting the available audience.
Check placement performance
Review the placement breakdown (section 3, Report 3). A change in how Meta is distributing your ads across placements — for example, showing more impressions on a lower-performing placement — can shift blended performance without any change you've made.
Check for external factors
Check whether CPMs have risen across the platform generally — this happens seasonally, particularly around major retail periods (Black Friday, the holiday season) when advertiser competition spikes. This is a market-wide cost increase, not a problem specific to your account.
Check recent account changes
Review whether a budget change, a new ad set, or an automated rule was applied recently. Significant budget increases in particular can temporarily disrupt a campaign's delivery as Meta's algorithm re-learns optimal delivery for the new spend level — this is sometimes called exiting and re-entering the "learning phase," and performance can be volatile for several days afterward.
Give changes time before judging them: Whenever you make a significant change to a Meta campaign — new creative, a budget change of more than around 20%, a new audience — expect a few days of volatility while the algorithm re-learns optimal delivery. Judging a change within 24–48 hours is one of the most common reasons advertisers reverse decisions that would have worked given a proper learning period.
Meta requires a different rhythm of attention than Google Ads — less about ongoing keyword and search term hygiene, more about creative rotation and fatigue monitoring.
That's roughly 50–60 hours per year of structured Meta management — comparable to Google Ads, but weighted more heavily toward creative production and rotation rather than search term and keyword hygiene.
There's no universal benchmark — it depends heavily on your industry, your offer, and what you're counting as a "result" (a lead, a purchase, a sign-up). The more useful approach is the same one used throughout this series: compare your current cost per result against your own 4-week trend and your own target, rather than against an industry average you've found online. Industry benchmark figures are frequently outdated, vary enormously by source, and don't account for your specific offer, audience, or average order value. Track your own number over time and treat any sustained, unexplained increase as the signal worth acting on.
The Meta Pixel is a small piece of tracking code, similar in purpose to the GA4 tracking tag, that you add to your website. It tells Meta when someone who saw or clicked your ad took an action on your site — like making a purchase or submitting a form. Without it, Meta has no way to know whether its ads are actually leading to results on your website, and its delivery algorithm has nothing to optimise toward. If you're running any Meta ads aimed at website conversions, the Pixel (or its more modern, more reliable counterpart, the Conversions API) is not optional — it's the foundation everything else in this guide depends on.
For the same reasons covered for Google Ads in Guide #3, plus one additional factor specific to Meta: the view-through attribution window covered in section 2. Meta counts a conversion if someone merely saw your ad (without clicking) and converted within 1 day, in addition to click-based conversions within a 7-day window. GA4 generally only counts the session that directly preceded a conversion. For comparing Meta against your other channels on a level playing field, use GA4. Use Meta's own reporting when you're optimising decisions within Meta specifically, like which ad or audience to scale.
There's no fixed number, but a reasonable starting point for most small businesses is 3 to 5 active ad variations per ad set. This gives Meta's algorithm enough variety to identify what's resonating, while keeping the account manageable to review weekly. Running too many at once (10 or more) spreads your budget thin across each one, slowing down how quickly any single ad accumulates enough data to be judged fairly. Running just one means you have no fallback when it inevitably fatigues.
Advantage+ is Meta's increasingly automated campaign type, where Meta's algorithm handles much of the audience targeting and budget allocation decisions that used to be manual. For businesses with strong historical conversion data feeding the algorithm, Advantage+ campaigns often perform well with less manual audience-building work. For newer accounts or businesses with limited conversion history, a more manually structured campaign with defined audiences can give you clearer signal about what's working while you build up that data. This guide's principles — checking frequency, CTR, and cost per result trends — apply regardless of which campaign type you choose.
Meta's ad review system rejects ads for many reasons, ranging from genuine policy violations to seemingly minor formatting issues that trip an automated review system. Go to Ads Manager, find the rejected ad, and click on the policy notice to see the stated reason. Many rejections can be resolved by editing the specific flagged element (often an image with too much text, certain claims in the copy, or a landing page issue) and resubmitting. If you believe a rejection is a genuine error, there's an appeal option within the same notice. Avoid resubmitting the exact same ad repeatedly without changes — this can trigger additional account-level review flags.
If your budget can't comfortably support multiple ad sets each accumulating enough data to judge fairly (roughly 15–20 conversions per week per ad set, as referenced in section 5), it's better to consolidate into fewer, slightly broader audiences than to spread a small budget across many narrow ones. A single well-performing ad set with adequate data will teach you more, and convert more efficiently, than five underfunded ad sets that never individually reach a meaningful sample size.
Everything in this guide works. Tracking frequency, watching for the early signs of creative fatigue, checking audience overlap, and refreshing creative before performance drops rather than after — these habits are what separate Meta accounts that perform consistently from ones that spike and fade.
The honest reality is that Meta asks for a different kind of ongoing effort than Google Ads. It's less about search term hygiene and more about a steady supply of fresh creative and consistent attention to a metric — frequency — that most advertisers never think to check until performance has already dropped. For a marketing manager already juggling multiple channels, that steady creative production is often the first thing to slip.
Zynop reads your Meta data every week alongside your other channels, flags rising frequency and fatigue signals before they show up as a cost per result problem, and surfaces the specific actions worth taking — refresh this creative, check this overlap, reallocate budget here. No switching between Ads Manager tabs, no manually tracking frequency in a spreadsheet, no missing the early warning because the week got busy.
If you're not ready for that yet, this guide gives you what you need to manage Meta well manually. The next guide in this series moves away from paid channels to cover organic search — how to read Google Search Console properly and understand what's actually happening with your SEO.