Paid Traffic Metrics: What to Track and What to Ignore

An ad with a 3% CTR can quietly eat your margin while another one, stuck at 1%, pays the company's bills at the end of the month. It sounds like a contradiction, but it's arithmetic: people who click a lot don't always buy, and the ad auction doesn't charge for purchase intent. It charges per click.

Every manager running Meta Ads or Google Ads knows the scene: the report shows eight indicators, each pointing in a different direction, and the meeting ends with the same old question. So, is the campaign doing well or not?

The answer is almost never to look at more metrics. It's to put the ones you already have in order.

The essential paid traffic metrics are five: CPM (cost per thousand impressions), CPC (cost per click), CTR (click-through rate), CPL (cost per lead), and ROAS (return on ad spend). The practical rule: pick one of them as the primary metric based on the campaign objective and use the others only for diagnosis.

What each paid traffic metric measures, with formula and example

To keep things concrete, we'll run all five calculations on the same campaign: a fashion retailer invested $3,000 in one month and got 150,000 impressions, 1,200 clicks, 60 leads, and 12 sales with an average order value of $1,000.

CPM: cost per thousand impressions

Formula: spend ÷ (impressions ÷ 1,000). In our example: $3,000 ÷ 150 = a $20 CPM. CPM tells you what it costs to show up, nothing more. It rises when competition for the same audience heats up (November is the classic case) and drops when the platform considers your creative relevant. It's a context metric: it explains cost swings, but it should rarely be a target.

CPC: cost per click

Formula: spend ÷ clicks. In our example: $3,000 ÷ 1,200 = a $2.50 CPC. Here lies a common trap: lowering CPC by broadening the audience or running clickbait creative. The click gets cheaper and the sale gets more expensive, because visitor quality collapses. CPC can only be judged alongside what happens after the click.

CTR: click-through rate

Formula: (clicks ÷ impressions) × 100. In our example: 1,200 ÷ 150,000 = a 0.8% CTR. CTR measures whether the ad grabs the right audience's attention, and it's a great creative thermometer. But optimizing CTR as a final goal produces ads that overpromise. An ad that qualifies ("starting at $300") lowers CTR on purpose and improves everything else.

CPL: cost per lead

Formula: spend ÷ leads. In our example: $3,000 ÷ 60 = a $50 CPL. It's the primary metric for anyone selling through a sales team. The risk is comparing it across channels without looking at quality: a dental clinic had an $18 CPL on Instagram's native lead form and a $45 CPL on Google. Moving the budget looked obvious. Except 8% of the form leads booked an appointment, versus 35% of the Google leads. In the math that matters, cost per booked appointment was $225 on Instagram and $129 on Google. The cheap CPL was the expensive one.

ROAS: return on ad spend

Formula: revenue generated ÷ spend. In our example: $12,000 ÷ $3,000 = a ROAS of 4, that is, $4 in revenue for every $1 invested. Is a ROAS of 4 good? It depends on margin. With a 30% gross margin, breakeven ROAS is 1 ÷ 0.30 = 3.33; the example campaign is profitable, but barely. With a 60% margin, breakeven drops to 1.67 and the same ROAS of 4 is excellent. That's why ROAS is analyzed in tandem with margin and CAC. If you want to go deeper on that math, see how to calculate CAC.

Which metric should you prioritize for each campaign objective?

Here's the most expensive mistake I see in audited accounts: holding a campaign to a metric that doesn't match its objective. An awareness campaign judged by CPL will always look like a disaster. A sales campaign judged by CPM will always look expensive. A hierarchy fixes this.

Campaign objective	Primary metric	Diagnostic metrics	Safe to ignore
Awareness / reach	CPM	Reach, frequency	CPL, ROAS
Qualified traffic	CPC	CTR, landing page bounce rate	CPM in isolation
Lead generation	CPL	Page conversion, booking rate	CTR as a target
Sales / e-commerce	ROAS	CPA, average order value, margin	Raw click counts

Notice the pattern: the primary metric sits as close as possible to the money the objective is chasing. The diagnostic ones exist to explain why the primary one got worse. If CPL went up, the CTR + page conversion pair points to the culprit: CTR dropped, the problem is creative or audience; CTR stable and conversion dropped, the problem is the page.

The metrics you can ignore (almost always)

Raw clicks and impressions look impressive in meetings and support no decisions. "We got 40,000 clicks" doesn't say whether they were cheap, whether they came from the right audience, or whether they turned into revenue. The same goes for likes and shares on a performance campaign: they're pleasant byproducts, not results.

CTR as a target deserves its own paragraph, because it's the most common vice. Back to the arithmetic from the opening: a clickbait ad with a 3% CTR and a $1.80 CPC, converting 1% of visitors, produces a $180 cost per conversion. An honest ad with a 1.2% CTR and a $2.60 CPC, converting 6%, produces a $43 cost per conversion. The ad that's "worse" on CTR costs four times less per result. In practice, optimizing CTR in isolation rarely works because the auction rewards clicks, not customers.

Frequency is also overrated out of context: it matters in high-budget reach campaigns, but in smaller accounts the weekly variation is noise. Check it once a month and move on.

Do benchmarks help or hurt?

They help calibrate expectations; they hurt when they become targets. Two recent numbers show the size of the game. IAB Brasil's annual study found that Brazil's digital advertising market reached R$ 42.7 billion in 2025, up 12.7% year over year. More budget fighting over the same auctions pushes CPM and CPC up, and that's not your media buyer's fault.

Meanwhile, WordStream's 2026 study of 13,474 search campaigns recorded an average CTR of 6.64%, an average CPC of $5.42, and an average conversion rate of 8.18% on Google Ads; for the first time in five years, average cost per lead actually dropped. These are useful references to know whether you're an outlier, but the honest comparison is always against your own history and your own margin. That formula breaks when industry, ticket size, and region change, and they always do.

One habit worth more than any benchmark: crossing your media metrics with what happens afterwards, in the CRM and in finance. It's the difference between knowing your CPL and knowing which channel brings customers who pay and stay. If you want to structure that view, start with this guide to digital marketing data analysis.

Frequently asked questions

What is a good ROAS?

There's no universal number. The minimum acceptable ROAS is 1 divided by your gross margin: with a 40% margin, any ROAS below 2.5 loses money. Healthy e-commerce operations usually run between 3 and 6, but the right target depends on margin, repeat purchases, and operating costs.

What's the difference between CPL and CPA?

CPL measures cost per lead, a contact who showed interest. CPA measures cost per acquisition, usually the completed sale. In businesses with a sales team, CPL is the intermediate metric and CPA is the final bill; a low CPL with a high CPA signals poor lead quality.

Which is better: a low CPC or a high CTR?

Neither on its own. The pair only makes sense next to cost per conversion. A high CPC with high conversion is usually a better deal than a cheap CPC that attracts window shoppers. Compare ads by cost per final result, not by cost per click.

How often should I review paid traffic metrics?

Daily, only to catch serious anomalies like a paused campaign or runaway spend. Optimization decisions need 7 to 14 day windows, because platform algorithms need volume to stabilize. Strategic analysis, with margin and CAC on the table, works best on a monthly cycle.

Is a high CPM always bad?

No. A high CPM can mean the audience is contested precisely because it's valuable, like B2B purchase decision-makers. If the final ROAS or CPL closes the math, a high CPM is just the price of admission. It becomes a red flag when it rises with no improvement in the result metrics.

A good metric is one that changes a decision

If a number can't move budget, creative, or audience, it's report decoration. The discipline that separates profitable accounts from confusing ones fits in three steps: set each campaign's primary metric by its objective, limit diagnosis to two or three supporting metrics, and revisit the hierarchy when the objective changes. Fewer dashboards, more decisions.

And if the hard part is precisely pulling Meta Ads, Google Ads, GA4, and your CRM together to see that hierarchy in one place, Sherlok does it for you: connect your accounts, ask in plain language ("which campaign had the lowest cost per sale this month?"), and get the analysis ready, with the numbers that actually pay the bills.