Blog · Board and management reporting · Sales teams
Facebook ad KPIs connect campaign spending, delivery, conversion events, and attributed purchase value. Use a worked campaign export to calculate the ratios, check the totals, and interpret attribution without claiming incremental sales.
Facebook ad KPIs connect campaign spending to delivery, conversions, and attributed purchase value. Compare campaigns only after aligning the conversion event, placement scope, attribution setting, reporting dates, and currency. Then use cost and revenue measures to identify the question the team should investigate.
A purchase, a lead, and a landing-page arrival are different outcomes. Write the event name beside the result count, and define what success means for that event. The KPI glossary explains why a business target belongs beside the measurement.
For a website purchase campaign, use this compact scorecard:
| KPI | Formula | What it helps explain |
|---|---|---|
| CPM | Spend / impressions x 1,000 | Cost of delivering impressions |
| Link CTR | Link clicks / impressions | Response to an advertiser destination |
| Cost per link click | Spend / link clicks | Cost of that response |
| Landing-page arrival rate | Landing-page views / relevant link clicks | How often reported clicks produce a recorded page arrival |
| Cost per purchase | Spend / attributed purchases | Advertising cost per credited purchase event |
| Purchase ROAS | Attributed purchase value / spend | Purchase value credited per advertising dollar |
Link clicks require care. Meta has explained that they include destinations within Facebook as well as external websites. For the arrival-rate calculation, restrict the export to ads whose relevant links lead to the website being measured; otherwise the numerator and denominator describe different destinations.
The arrival rate is a business calculation defined here, not a promise that each recorded page view maps to a unique person. It can help flag a change in arrival measurement or the website experience, but it cannot diagnose the cause on its own.
Bring campaign or ad-set IDs, names, dates, placement scope, currency, spend, impressions, link clicks, landing-page views, purchase events, purchase value, and attribution settings. Keep account totals separate from campaign rows. Combining summary rows and detail rows counts the same activity twice.
If the job concerns Facebook placements, filter to that scope before totaling. A cross-placement export can answer a broader paid-social question, but it should carry that label. Record the export date because recent conversion figures can still develop as the selected observation window matures.
Preserve the exact conversion and attribution labels supplied. Meta's March 2026 measurement announcement describes changing click-through attribution for website and in-store conversions to link clicks and separating other engagement credit. A historical comparison therefore needs a definition check as well as matching column names.
Fictional US company Cedar Office Goods sells workplace supplies online. Its September 2026 export is restricted to Facebook website-purchase ads, frozen on October 23, 2026. Both rows use the same supplied attribution setting, and the relevant link clicks all lead to its website. These are illustrative inputs, not industry benchmarks.
| Audience campaign | Spend (USD) | Impressions | Link clicks | Landing-page views | Attributed purchases | Attributed purchase value (USD) |
|---|---|---|---|---|---|---|
| Prospecting | 12,000 | 600,000 | 9,000 | 7,200 | 240 | 36,000 |
| Returning audience | 6,000 | 150,000 | 4,500 | 3,600 | 180 | 27,000 |
| Total | 18,000 | 750,000 | 13,500 | 10,800 | 420 | 63,000 |
Calculate ratios from these inputs. Dollar ratios and percentages are rounded to two decimal places; ROAS is rounded to two decimal places. Use unrounded values in subsequent calculations.
| Audience campaign | CPM (USD) | Link CTR | Cost per link click (USD) | Arrival rate | Cost per purchase (USD) | Purchase ROAS |
|---|---|---|---|---|---|---|
| Prospecting | 20.00 | 1.50% | 1.33 | 80.00% | 50.00 | 3.00x |
| Returning audience | 40.00 | 3.00% | 1.33 | 80.00% | 33.33 | 4.50x |
| Total | 24.00 | 1.80% | 1.33 | 80.00% | 42.86 | 3.50x |
Prospecting CPM is 12,000 / 600,000 x 1,000 = $20.00. Its cost per purchase is 12,000 / 240 = $50.00, and its purchase ROAS is 36,000 / 12,000 = 3.00x. Returning-audience ads cost more per impression but have a higher link response and lower cost per attributed purchase.
These findings describe the export. They do not establish that the returning audience contains more new customers, or that increasing its budget would preserve its current ratios.
The total cost per purchase must divide total spend by total purchases: 18,000 / 420 = $42.86. The total ROAS is 63,000 / 18,000 = 3.50x. Averaging campaign ratios would give each campaign equal weight despite different spending.
Check a second path through the counts. Total arrival rate is 10,800 / 13,500 = 80.00%, and purchases divided by page views are 420 / 10,800 = 3.89%. The latter is a descriptive ratio of reported events; because attribution can include other interaction types, do not label it a verified probability that a page visitor bought.
Tie spend back to the selected advertising rows, and purchases and values back to the supplied event summaries. Separately compare order records after agreeing dates, currency, returns, and event definitions. Keep any unexplained difference visible rather than forcing attributed purchase value to equal accounting revenue.
The paid-social manager owns campaign scope and attribution labels. The ecommerce team owns order definitions and website observations; finance owns the comparison with revenue and margin. Use dashboard selection and KPI checks when deciding how these definitions will stay visible in recurring reporting.
A rising CPM calls for a review of delivery conditions and audience mix. A falling arrival rate calls for a check of destination scope and measurement before a creative conclusion. A rising cost per purchase with stable traffic costs directs attention to the conversion outcome and its reporting basis.
Attributed sales still need separate evidence for causality. Meta Blueprint treats incrementality studies as a validation layer for reporting. Keep that question distinct from recomputing export ratios. For the separate paid-search workflow, see Google ad KPIs.
Covirage is not affiliated with Meta or Facebook.
Bring a sample of your own campaign export or Google Sheet; Covirage maps the columns, and its tools compute and check the figures. The external AI model explains the results with every figure traced to its rows. Create a free Covirage account, which takes about a minute, then set up your business and bring your file.
See revenue driver analysis for the supporting approach.
Start with the campaign's intended outcome, such as purchases or qualified leads, and its cost. Use attributed revenue and margin information to assess economics. Delivery and click measures explain the result but do not independently establish sales effectiveness.
No. Purchases can include repeat orders from existing customers, while customer acquisition cost needs verified new customers. An advertising-only calculation also excludes other acquisition expenses. Label both the denominator and the included costs before comparing the measures.
Higher attributed ROAS means more reported purchase value per advertising dollar under the selected rules. It can reflect audience mix, repeat buyers, different margins, or attribution settings. A conclusion about incremental sales requires evidence about what would have happened without the advertising.
Do not assume audiences are mutually exclusive. The same person may appear in multiple campaigns, so summed campaign reach can count that person repeatedly. Use a deduplicated total for the same reporting scope when calculating aggregate frequency.