Click-through rate, cost per click and CPM tell you how an ad performs inside its platform. They do not tell you whether that platform is making you money. Here is what each metric is good for, why platform-reported ROAS overlaps, and how to measure return per platform from your own revenue.
These metrics are real and useful. They are just about the ad, not about the business. Each one stops somewhere before the money.
Cost per thousand impressions. It measures how expensive the audience is to reach. A low CPM means cheap attention, which says nothing about whether that attention is from people who need your service.
Click-through rate. It tells you whether the ad is interesting to the people who saw it, which is a good creative and targeting signal. Curiosity clicks and wrong-fit clicks count the same as buyers.
Cost per click. It depends on competition for the audience or keyword. Cheap clicks can mean an underpriced opportunity or an audience nobody else wants, and the CPC alone cannot tell you which.
The share of clicks that became a tracked conversion. It is closer to the business, but only as good as the conversion definition — a form fill, a button tap and a booked job all look alike here.
Use click metrics to diagnose ads within a platform: which headline earns attention, which audience is overpriced, which landing page loses people. Do not use them to compare platforms against each other. A search click and a social video click are bought in different auctions from people in different frames of mind, and their prices are not measuring the same thing.
Most ad platforms will show you a return on ad spend if you give them conversion values. That number is calculated under the platform’s own rules, and those rules lean toward giving the platform credit.
Each platform credits a conversion to its own ads if its rules are met. A customer who saw a Meta ad, clicked a Google ad and later clicked a Microsoft ad can be counted as a conversion by all three. Add up their reports and you get more sales than you made.
Depending on settings and campaign type, platforms can count a conversion from someone who saw an ad without clicking it. Meta's default attribution setting, for example, has included a one-day view window alongside a seven-day click window.
Click windows vary by platform and by conversion action, and in some cases run as long as 30 days or more. Two platforms with different windows are not measuring the same period, even in the same date range.
Where consent or browser limits hide a conversion, platforms estimate it and fold the estimate into reported totals. Those numbers are statistical rather than individual, so there is no lead record to check them against.
Platform revenue is whatever value the conversion was given. If every form fill is valued at an average job price, the reported ROAS assumes every lead closed, which no service business achieves.
The platform rarely knows whether the lead answered the phone, got a quote or paid an invoice, unless you send that outcome back. Its revenue is often the value of the inquiry, not the sale.
None of this makes platform ROAS useless. It is what the platform’s bidding learns from, and it is broadly useful for ranking campaigns within that platform. It just cannot be summed across platforms or compared between them. The specific ways each platform overcounts are covered in Google Ads Says 30 Conversions. Your CRM Says 9. and Facebook Says 40 Leads. You Got 12.
The fix is to compute return in one place, with one set of rules, using sales you can see. Each sale is credited to one platform, so the platforms add up to your real total.
Capture UTMs and click IDs on arrival, keep them in first-party cookies, and save them with the lead — forms, calls and lead-form ads alike. Each lead should carry exactly one platform under the rule you choose, such as first touch.
Connect won jobs from your CRM, invoicing or field-service software to the lead record by phone, email or ID. The revenue inherits the lead's platform and campaign.
Use the spend each platform billed, not its reported conversions. Spend is the one platform number that is not modeled.
Revenue from a platform's leads divided by that platform's spend is its ROAS. Spend divided by customers won is its cost per customer. Run the same math for every platform, and the comparison is fair.
Compare months that are old enough for most of their leads to have closed or gone cold. A platform whose customers take six weeks to decide will look worse than it is in last week's numbers.
An illustrative example, not a benchmark. Say Platform A delivers clicks at $1.50 and Platform B at $6.00. Over a month, each gets $3,000. A produces 2,000 clicks, 60 leads and 4 customers worth $9,000. B produces 500 clicks, 25 leads and 8 customers worth $22,000. A’s clicks were four times cheaper, but B’s cost per customer was half of A’s and its revenue per dollar was more than double. Only the revenue side shows that.
A click is priced by how many advertisers want the same person at the same moment. On search, someone typing “emergency plumber near me” is close to buying, and every plumber in town is bidding for them, so the click is expensive. On a social feed, the same person scrolling at night is not looking for anything, there is less competition for their attention, and the click is cheap. The difference in price is mostly a difference in intent, and intent is what turns a lead into a customer.
That is not a rule that search always wins. Social platforms can be excellent at creating demand for things people do not search for yet — a remodel, a cosmetic treatment, a seasonal service — and some of those customers come back later through search or by typing your name. That is exactly why a fair comparison needs both a first-touch and a last-touch view of the same revenue: one shows who introduced the customer, the other who closed them. Click metrics cannot show either.
The practical way to read the two layers together is top-down. Start with platform-level ROI from your own revenue to decide how the budget is split between platforms. Then, inside each platform, use its own reporting — CTR, CPC, conversion rate, its ROAS — to decide which campaigns, audiences and ads deserve that platform’s share. When a platform’s own numbers look great but its revenue does not, trust the revenue and look at what the platform is being told counts as a conversion.
That last point closes the loop. If you report closed sales and their values back to the platforms — through Google’s offline conversion import or Enhanced Conversions for leads, and Meta’s Conversions API — their bidding optimizes toward customers rather than clicks, and the gap between their numbers and yours tends to narrow. For more on choosing between networks, see Comparing Ad Networks by Revenue, and run your own figures through the ROAS calculator and the cost-per-customer calculator.
AuriaTrack tracks Google, Meta, Microsoft Advertising and Outbrain campaigns, syncs spend daily, and matches it against the leads and closed deals your business actually generates. The tracking script keeps UTMs and click IDs in first-party cookies with a write-once first-touch record; won deals sync in from GoHighLevel, Square, ServiceTitan, Housecall Pro, Jobber or a CSV and are matched to the lead that started them. The result is ROAS at the campaign and platform level, a side-by-side platform comparison, and cost per closed deal — all from revenue you can trace to a real customer. Real revenue then goes back to the platforms through server-side Meta Conversions API and Google Enhanced Conversions.
Book a demo and we'll line up each platform's spend against the closed revenue its leads produced.