Meta ads metrics fall into five groups: delivery (impressions, reach, frequency, CPM), clicks (link clicks, CTR, CPC, landing page views), video (3-second plays and ThruPlays), conversions (results and cost per result) and value (purchase ROAS). To judge a creative, five of them carry the decision: link CTR, hook rate for video, cost per result, purchase ROAS and frequency, read per creative and against your own account's history. Below: Meta's definition and formula for each, what to change when one moves, and the engage-through attribution setting that changes which conversions count.
Meta's definitions and formulas below come from its Business Help Center and its Glossary of Ad Terms, read on November 6, 2026. Hook rate, hold rate and conversion rate are not Meta metrics; they are custom metrics you build, and they are marked that way.
Meta ads metrics list: formulas at a glance
| Metric | Formula | What it tells you |
|---|---|---|
| Impressions | Times your ads were on screen | Volume, and the denominator for most rates |
| Reach | Accounts that saw the ad at least once (estimated) | How many people you reached |
| Frequency | Impressions / reach | How often each person saw it |
| CPM | Amount spent / impressions × 1,000 | The price of attention |
| Link clicks | Clicks on links to your chosen destination | Traffic intent |
| CTR (link click-through rate) | Link clicks / impressions | Whether the ad earns the click |
| CTR (all) | Clicks (all) / impressions | Any interaction, likes and profile clicks included |
| CPC (cost per link click) | Amount spent / link clicks | The price of a click |
| Outbound clicks | Clicks to destinations off Meta | Traffic that left Meta |
| Landing page views | Clicks that loaded your page or app | Traffic that arrived |
| 3-second video plays | Plays of 3 seconds, or 97% of shorter videos | Whether the opening stopped the scroll |
| ThruPlays | Plays to completion or to 15 seconds | Whether viewers stayed |
| Hook rate (custom) | 3-second video plays / impressions | Opening strength per impression |
| Hold rate (custom) | ThruPlays / 3-second video plays | Body strength per hooked viewer |
| Results | Outcomes counted for your objective | What the campaign was built to get |
| Cost per result | Amount spent / results | What each outcome costs |
| Conversion rate (custom) | Results / link clicks | Whether the click turns into the outcome |
| Purchase ROAS | Purchase conversion value / amount spent | Revenue per dollar spent |
Which Meta ads metrics matter most?
The ones tied to the decision in front of you. Cost per result says whether an ad works, or purchase ROAS when the goal is sales. Why it works or fails sits upstream, because cost per result breaks into three parts:
cost per result = CPM / (1,000 × link CTR × conversion rate)
That is the price of attention, the share of people who click, and the share of clickers who convert. A worked example, with illustrative numbers rather than benchmarks: a $12 CPM, a 1.2% link CTR and a 3% conversion rate give a cost per result of $33.33. Drop link CTR to 0.9% and change nothing else, and cost per result rises to $44.44.
Each part points at a different fix. A creative problem shows up in link CTR, and in hook rate for video. A landing page or offer problem shows up in conversion rate. An auction or audience problem shows up in CPM. Frequency is the context for all three.
In Peachblue, that read is the composite score: every creative ranks by percentile on CTR, ROAS, CPA and spend against your own account, then lands in a tier from Top Performer to Underperformer.
What are Meta's delivery metrics?
Impressions, reach, frequency and CPM describe how often your ad was shown, to how many people, and at what price. On their own they say little about the creative, but they frame every other number.
- Impressions count the times your ads were on screen, once per instance: scroll away and back and it is still one, on screen at two different times in a day is two. A video does not have to start playing for the impression to count (Meta).
- Reach is the number of Meta Accounts that saw your ads at least once. Meta estimates it from sampled data.
- Frequency is impressions divided by reach. Meta says it can average 1 to 2 per ad set, or much more, depending on budget, audience size and schedule, and that performance dropping as frequency rises is a sign of ad fatigue. Meta publishes no threshold.
- CPM is amount spent divided by impressions, times 1,000. Meta's example: $50 for 10,000 impressions is a $5 CPM.
When CPM rises and link CTR holds, the auction or the audience got more expensive; the creative did not get worse. When frequency climbs while link CTR falls, the creative is wearing out. The creative fatigue guide covers that diagnosis, including Meta's Creative limited and Creative fatigue statuses.
In Peachblue, reach, frequency and CPM sync daily for every ad, and the fatigue board flags a creative whose CTR decays against its own best rolling week, with the dollars at stake beside each flag.
What is the difference between CTR (all) and link CTR?
CTR (all) counts every click on the ad; link CTR counts only clicks on links to your destination. Meta's clicks (all) include link clicks plus clicks to your Page profile or profile picture, post reactions, comments, shares, and taps that expand media to full screen. CTR (all) is clicks (all) divided by impressions; CTR (link click-through rate) is link clicks divided by impressions (Meta).
Judge creative on link CTR. A reaction-heavy ad can post a strong CTR (all) and send almost no traffic. CPC splits the same way: CPC (cost per link click) is amount spent divided by link clicks, while CPC (all) divides by every click.
Two more click metrics catch what link clicks miss:
- Outbound clicks count only clicks that leave Meta, to websites, app stores or deep links. A call-to-action click on a single image ad counts as both a link click and an outbound click; a website click inside an Instant Experience counts only as outbound (Meta).
- Landing page views count clicks that successfully loaded your page or app, measured through the Meta Pixel or the in-app browser. Meta names the usual causes of the gap: the page did not fully load, the app did not open, or the ad has several destinations (Meta).
A custom landing page view rate (landing page views divided by link clicks) separates a creative that earns clicks from a page that loses them.
In Peachblue, CTR means link CTR: link clicks over impressions, added up across every ad a creative ran in before dividing. It is one of the four inputs to the composite score.
Which Meta video metrics measure the hook?
3-second video plays and ThruPlays. Meta counts a 3-second play when the video plays for at least 3 seconds, or 97% of a shorter video, with replays excluded. A ThruPlay is a play to completion or to at least 15 seconds. Video plays at 25, 50, 75, 95 and 100% show how far viewers got, including people who skipped ahead, and video average play time includes replays, so it can run longer than the video itself.
Hook rate and hold rate are not Meta metrics. You build them:
hook rate = 3-second video plays / impressions
hold rate = ThruPlays / 3-second video plays
Hook rate scores the opening; hold rate scores the body. The hook rate and hold rate guide works an example and covers the TikTok and YouTube equivalents, and the hook rate calculator runs the math on your own totals. Published "good hook rate" figures contradict each other (20 to 25% in some sources, 30 to 45% or 60% and up in others), so judge each video against your own account's trailing average.
In Peachblue, hook style is one of the 31 dimensions tagged on every video, so you see which openings your top performers share, not only how well each one hooked.
What do results, cost per result and purchase ROAS measure?
Results count the outcome your campaign is built for, set by its objective and settings; cost per result is amount spent divided by results; purchase ROAS is purchase conversion value divided by amount spent. When the result is an action off the ad, like a purchase on your site, Meta counts it only if it happens within your attribution window (Meta).
Three details change how you read them:
- Purchases are broader than your pixel. Meta counts purchases from the Meta Pixel and Conversions API, purchases on Meta itself (Marketplace, Instagram Shopping, checkout), and purchases it detects in messaging conversations. Website purchase ROAS keeps only Pixel and Conversions API website purchases, the closer match to your store's own numbers.
- Some results are modeled. Where events cannot be measured directly, Meta may fill in with statistical modeling, and for iOS 14+ ads cost per result uses results reported by Apple's SKAdNetwork.
- Results arrive late and settle slowly. Meta says results can take up to 48 hours to process. An ad set in the learning phase, which usually ends after about 50 results in the week after the last significant edit, delivers less steadily and usually at a higher CPA (Meta).
Conversion rate is the custom metric that closes the chain: results divided by link clicks, or purchases divided by landing page views to take page load out of it.
The ROAS you need is set by your margin, not by a benchmark. ROAS, calculated honestly shows the breakeven math, and blended vs creative ROAS explains why the account average hides which creative earns it.
In Peachblue, results follow the campaign's outcome, purchases when an ad sells and leads when it collects them, so CPA works for lead generation as well as sales. ROAS is purchase value over spend, and both feed the composite score beside CTR and spend.
How does Meta's attribution setting change your numbers?
It decides which conversions count. Meta's default is 7-day click, 1-day view and 1-day engagement (Meta). Under the standard model, Meta's help page lists three settings for website and in-store conversions (Meta):
| Setting | Counts conversions within | After |
|---|---|---|
| Click-through | 1 or 7 days | A link click on the ad |
| View-through | 1 day | An impression of the ad |
| Engage-through | 1 day | Any other click on the ad, or a 5-second video play (97% of shorter videos) |
Engage-through is the piece to watch. Click-through now means a link click only; a purchase after someone reacted to the ad or watched five seconds of the video counts as engage-through, inside one day. Meta notes that some accounts still use prior versions of click-through and engage-through while the change rolls out, so two accounts, or one account across months, may not be counting the same way. Compare periods only under the same setting, and run Meta's Compare Attribution Settings before you switch.
Meta also offers incremental attribution, which optimizes for and reports the conversions its models predict the ad caused, and custom attribution from your own analytics tool. Results from ad sets on different models cannot be compared in the Campaign Overview table, Meta warns.
In Peachblue, each daily sync reads the previous seven days again, so conversions Meta attributes late still land on the right creative.
How do you build custom metrics in Ads Manager?
In Ads Manager, open Columns, choose Customize columns, then Create custom metric. Build the formula from existing metrics and operators, name it, pick a format (Numeric, Percentage or Currency), and choose who can see it (Meta). Five worth saving:
| Custom metric | Formula | Format |
|---|---|---|
| Hook rate | 3-second video plays / impressions | Percentage |
| Hold rate | ThruPlays / 3-second video plays | Percentage |
| Landing page view rate | Landing page views / link clicks | Percentage |
| Conversion rate | Results / link clicks | Percentage |
| Average order value | Purchases conversion value / purchases | Currency |
Test a formula privately before you share it: once a custom metric is shared with everyone on the ad account, Meta does not let you edit or delete the formula. Custom metrics only do basic math on existing metrics; they do not change the underlying numbers.
The limit is that a custom metric is calculated per row. A video running in six ads gets six hook rates on six thin samples, and reading it as one creative takes the next step.
How do you read Meta ads metrics by creative?
Group by creative, add before you divide, judge against your own account, and wait for enough data. Ad-level rows mislead because one video usually runs in several ads, each on a thin sample.
- Group ads into creatives. Meta's Ads Reporting has an Ad creative breakdown, which Meta is introducing gradually; it reports results, reach, impressions, cost per result and amount spent per creative, and leaves out dynamic creative ads (Meta). In Peachblue, perceptual fingerprinting recognizes the same creative across ads and platforms, even after re-crops, resizes and re-compression, so a re-upload is not a new creative.
- Add, then divide. Sum impressions, link clicks, results and value across the group, then compute the rates. Two ads running one video, one at 30,000 impressions and a 1.0% link CTR and the other at 300 impressions and 5%, average to 3%; the creative's real link CTR is 1.04%. In Peachblue, the By Creative view does the adding before it calculates CTR, CPA and ROAS.
- Judge against your own account. Published benchmarks mix objectives, placements and verticals. In Peachblue, the composite score ranks every creative against your own account and tiers it Top Performer, Above Average, Average or Underperformer.
- Wait for enough data. Meta withholds ad relevance diagnostics under 500 impressions, and its quality, engagement rate and conversion rate rankings compare your ad with ads competing for the same audience, not with your own history (Meta). In Peachblue, the Creative Library's Low-data view keeps thin creatives apart from the ones ready for a verdict.
- Watch each winner against its own best week. A winner's CTR sliding below its own peak is the signal to watch; Meta's own fatigue statuses wait for cost per result to climb above your past ads. In Peachblue, the fatigue board measures exactly that and puts a dollar figure on every flag; creatives with no spend are never flagged.
- Count what the misses cost. Spend on confirmed losers has no column in Ads Manager. In Peachblue, Creative Economics reports wasted spend and hit rate, the share of launches that become winners, and the free creative testing budget calculator runs the same math on your own inputs.
What makes Peachblue better than Ads Manager for Meta metrics?
Ads Manager reports the metrics; Peachblue turns them into decisions: which creative gets the next dollar, which one is wearing out, what the misses cost, and what to make next. Ads Manager is free and the only home of Meta's own diagnostics, so keep it open. Peachblue is a creative intelligence platform for brands and agencies. It analyzes every ad you run on Meta, TikTok, Google Ads, and Amazon DSP, explains why your winners win, and shows you what to do next.
| You ask | Ads Manager gives | Peachblue gives |
|---|---|---|
| Which creative is winning? | Rows by ad; an Ad creative breakdown, still rolling out | One row per creative, re-uploads grouped, ranked against your own account |
| Is a winner wearing out? | Frequency, plus Creative limited or Creative fatigue where eligible | CTR decay against the creative's own best week, in dollars |
| What are the misses costing? | Spend by ad | Wasted spend and hit rate |
| Why does this ad win? | Rankings against competing ads | 31 tagged dimensions and a written analysis per creative |
| What should we make next? | Inspiration themes, in testing | A Next Creative Brief from your own winners |
| Which platforms? | Meta | Meta, TikTok, Google Ads and Amazon DSP |
Performance. Today refreshes as each platform syncs: new winners, drains and fatigue flags, each with its dollars. Every Monday the weekly report lands by email and in Slack with spend and ROAS against the week before, waste, hit rate and the one thing to act on. The next dollar goes to the creative that earned it, and every cut comes with its cost.
Creative. Every creative is tagged across 31 dimensions (format, hook style, emotional tone, CTA type and more) with a written analysis and copy suggestions, and the Intelligence page finds the patterns and archetypes your winners share. Link CTR tells you an ad won; the tags tell you why.
The loop. The Next Creative Brief turns those patterns into the next round. It carries the winning recipe, concepts to make next, proven hooks from your own headline performance, do and don't rules with cohort counts, and labeled reference ads, plus a prompt block ready for an image or video generator, and it shares by link. The new batch is scored on the same scale, and hit rate shows whether the loop is producing winners or just volume. The same data answers in Claude through the MCP server.
Meta and TikTok are on every plan from Starter at $79 a month; Intelligence, the Next Creative Brief and the MCP server start on Pro (see plans).
One boundary, stated plainly: Peachblue reads the numbers Meta reports. It does not re-attribute conversions, and it does not create, edit or pause ads. It does not sync 3-second plays or ThruPlays, so hook rate and hold rate stay a custom metric in Ads Manager or the free calculator.
Where to go deeper on each metric
- Hook rate, hold rate and thumbstop: formulas and a worked example, and the calculator.
- ROAS: the breakeven math and blended vs creative ROAS.
- Frequency and fatigue: creative fatigue in Meta ads, and why near-identical ads count as one creative under Meta Andromeda.
- Hit rate and testing spend: creative hit rate and the creative testing budget guide.
- The category: what creative analytics is, and how to run this analysis in Claude for Meta ads.