Track performance marketing metrics across the funnel: reach and clicks, conversions, revenue, and profit or contribution margin. Start with conversion volume, conversion rate, CPA or CAC, revenue, and profit, then add diagnostic metrics that explain movement. The right primary metric depends on the campaign goal: ecommerce may need profitable revenue or contribution margin, while lead generation may need sales-accepted leads or pipeline value. Cheap clicks count only when they produce customers or other valuable outcomes. Your reporting should connect ad spend to business results while accounting for tracking gaps, attribution rules, delayed conversions, refunds, and customer value. A short dashboard that supports decisions is more useful than a crowded one.
1. Which performance marketing metrics should I look at first?
Start with the numbers that answer two questions: did the campaign produce the result you wanted, and did it do so at an acceptable cost? For most paid campaigns, begin with conversion volume, conversion rate, cost per acquisition, revenue, and profit or contribution margin.
Conversion volume tells you how much demand the campaign generated. Conversion rate shows how efficiently people moved from one defined step to the next, such as click to purchase or landing-page visit to lead. Cost per acquisition, or CPA, shows what you paid for each recorded conversion. Revenue shows the sales attributed to the campaign. Profit or contribution margin goes further by accounting for the cost of delivering those sales.
A useful first-pass report might look like this:
- Spend: $10,000
- Purchases: 250
- Purchase conversion rate: 2.5% of 10,000 visits
- CPA: $40
- Attributed revenue: $30,000
- Contribution margin after product and fulfilment costs: $12,000
These numbers will not diagnose every problem, but they quickly show where to investigate. If purchases are down, check traffic quality, landing-page conversion, checkout performance, and tracking. If revenue is stable but profit is falling, inspect discounts, product mix, fulfilment costs, or customer refunds.
Do not begin with every number in the advertising platform. Reach, frequency, link clicks, video views, and engagement can explain performance, but they rarely prove that paid marketing worked. Keep the first set small enough to review weekly and tie it to decisions such as increasing budget, changing creative, fixing a landing page, adjusting targeting, or stopping spend.
2. What is the one metric that tells me whether a campaign is working?
There is no universal winner. The primary metric should match the campaign's business goal and the point at which value becomes real.
For ecommerce, that may be contribution margin or profitable revenue rather than purchases alone. For lead generation, it may be qualified leads, sales-accepted opportunities, pipeline value, or closed-won revenue. For an app, it could be activated users, paid subscribers, or retained users after a defined period. A campaign can look excellent on an early metric while failing on the outcome that pays the bills.
Suppose a campaign is designed to generate software demos. Reporting only form submissions may reward campaigns that attract students, job seekers, competitors, or people outside the service area. A better primary metric might be qualified demos, with cost per qualified demo as the main efficiency measure. If sales data is available, closed-won revenue or expected pipeline value gives a stronger view, although it will arrive later.
Choose the metric by asking:
- What action creates business value?
- Can marketing influence that action within the reporting period?
- Is the action measured consistently across channels?
- What cost or quality threshold makes it worthwhile?
Keep an earlier metric as a leading indicator when useful. A lead-generation team may watch form completions daily, qualified leads weekly, and closed revenue monthly or quarterly. The mistake is treating the early indicator as the final outcome.
Give the primary metric a guardrail. If you optimize for purchase volume, watch profit or average order value. If you optimize for qualified leads, watch lead volume so quality does not become so strict that demand disappears. The point is to manage the trade-off, not pretend that one number describes the whole business.
3. How do I track the full journey from impression to profit?
Track each meaningful step in the path, then connect those steps to a customer or order wherever your systems allow. A typical paid campaign path is impressions, clicks, landing-page visits, leads or purchases, repeat value, and profit.
For an online store, imagine a campaign with $10,000 in spend. It generates 500,000 impressions, 10,000 clicks, 8,000 landing-page visits, and 250 first purchases. The average order value is $120, so attributed first-order revenue is $30,000. If product, payment, shipping, and fulfilment costs total $18,000, the first-order contribution before advertising is $12,000. After the $10,000 ad spend, the campaign has $2,000 in contribution from those orders.
The calculations show different parts of the journey:
- Click-through rate: 10,000 clicks divided by 500,000 impressions, or 2%.
- Visit rate from clicks: 8,000 visits divided by 10,000 clicks, or 80%.
- Purchase rate from visits: 250 purchases divided by 8,000 visits, or 3.125%.
- CPA: $10,000 divided by 250 purchases, or $40.
- First-order ROAS: $30,000 divided by $10,000, or 3.0.
Add repeat purchases and refunds as they mature. If customers acquired by this campaign usually place later orders, customer lifetime value may make the campaign attractive even when first-order contribution is modest. Base that value on observed customer behaviour or a clearly labelled forecast; do not use it to make weak results look better.
Use consistent definitions. Decide whether a conversion means a completed payment or an order created, whether revenue includes tax and shipping, and how refunds are treated. Record campaign, ad, audience, landing page, order, and customer identifiers where privacy rules and consent allow. Without consistent keys and definitions, the funnel becomes a set of plausible numbers rather than one connected account of performance.
4. Which metrics are useful and which are just vanity numbers?
A metric is not a vanity metric simply because it appears early in the funnel. It becomes one when it attracts attention but does not help you make a decision about business performance.
Diagnostic metrics include impressions, reach, frequency, click-through rate, cost per click, landing-page load rate, video completion, and engagement. They help explain delivery and user behaviour. A falling click-through rate may point to tired creative. A high cost per click may signal intense competition or weak relevance. A large gap between clicks and landing-page visits can reveal slow pages, broken links, redirects, or tracking problems. Reach and frequency can show whether you are repeatedly showing ads to a small audience.
Outcome metrics include qualified conversions, sales-accepted leads, purchases, CPA, customer acquisition cost, revenue, contribution margin, retention, and lifetime value. These deserve more weight in budget decisions because they connect activity to value.
The same metric can answer one question well and another poorly. Impressions matter when checking delivery or estimating whether a campaign has enough exposure. CTR helps compare creative within the same audience and placement. CPC helps diagnose traffic costs. None of these proves that traffic was profitable.
Be careful with cheap traffic. A low CPC can come from broad placements or low-intent audiences. If those visitors bounce, submit poor-quality forms, or never purchase, the cheap clicks have only made the report look busy. Compare traffic metrics with downstream rates: click to qualified lead, qualified lead to opportunity, opportunity to sale, or visit to purchase.
Label each dashboard metric as an outcome, leading indicator, or diagnostic. Outcomes guide budget decisions. Leading indicators show movement before revenue arrives. Diagnostics tell you what to investigate. That label keeps a rising engagement rate from quietly becoming the definition of success.
5. How do I calculate CPA, CAC, ROAS, and conversion rate?
These metrics overlap, but they answer different questions.
CPA, or cost per acquisition, is spend divided by recorded acquisitions:
CPA = campaign spend ÷ attributed conversions
If you spend $10,000 and receive 250 purchases, CPA is $40. Define the acquisition first. It might be a purchase, lead, booked appointment, or activated account. A CPA based on all leads is not comparable with a CPA based on qualified leads.
Conversion rate is conversions divided by the relevant opportunity set:
Conversion rate = conversions ÷ clicks, visits, or another defined denominator
If 250 visitors purchase from 8,000 landing-page visits, the purchase conversion rate is 3.125%. State the denominator every time; “conversion rate” alone is incomplete.
ROAS, or return on ad spend, is attributed revenue divided by advertising spend:
ROAS = attributed revenue ÷ ad spend
$30,000 in attributed revenue from $10,000 in ads produces 3.0 ROAS, often written as 3:1. ROAS does not subtract product cost, shipping, discounts, salaries, agency fees, or returns.
CAC, or customer acquisition cost, normally measures the cost of acquiring new customers and often includes broader costs than platform CPA:
CAC = acquisition costs ÷ new customers acquired
If a business spends $10,000 on ads plus $2,000 on agency and creative costs and acquires 200 new customers, blended CAC is $60. Platform CPA could still be $40 if it reports 250 purchases. The metrics differ because one counts all attributed purchases while the other includes additional acquisition costs and focuses on new customers.
Use CPA to manage a defined conversion, ROAS to compare attributed revenue with ad spend, and CAC to understand the broader cost of gaining customers. None replaces profit.
6. Why does a low cost per lead not always mean good marketing?
A low cost per lead helps only when the leads have a reasonable chance of becoming customers. Lead volume without lead quality can give a sales team more work and less revenue.
Connect the campaign to later funnel stages. For each source, track leads generated, leads accepted by sales, qualified opportunities, proposals or appointments, closed customers, revenue, and, where possible, gross margin. Then calculate lead-to-qualified, qualified-to-opportunity, and opportunity-to-sale rates.
Imagine Campaign A generates 500 leads at $10 each, costing $5,000. Only 25 are accepted by sales, and five become customers worth $2,000 each. Campaign B generates 200 leads at $25 each, costing the same $5,000. One hundred are accepted, 40 become opportunities, and 15 become customers worth $2,000 each. Campaign A has the lower CPL, but Campaign B produces three times as many customers and much more revenue.
Pass quality signals back to the ad platform when the platform and your consent setup support that use. A basic form-submission event may be available immediately, while a qualified or closed-sale event arrives later from the CRM. Keep the original source, campaign, ad, and lead identifiers so sales outcomes can be joined to acquisition data.
Quality can vary by geography, product, company size, need, or expected contract value. A technically valid lead may still be commercially poor. Define “qualified” clearly and use the same definition across channels.
Waiting for closed revenue gives you a better target but fewer recent observations. Use early quality signals for short-term optimisation, then validate them against sales outcomes regularly. If cheap leads consistently fail later in the funnel, exclude the audiences, placements, or messages that generate them instead of celebrating the low CPL.
7. Should I optimize for ROAS or profit?
Use profit or contribution margin when you can measure it reliably. ROAS is a useful shortcut for comparing attributed revenue with ad spend, but it does not show how much revenue remains after the costs of fulfilling the sale.
A high-margin product can tolerate a lower ROAS than a low-margin product. Say one product has a 70% contribution margin before advertising and another has a 25% margin. A 2.0 ROAS may leave room for advertising in the first case but lose money in the second, depending on the rest of the cost structure.
Calculate a break-even ROAS from the share of revenue available to cover advertising. If 40% of revenue remains after variable costs, break-even ROAS before fixed overhead is 1 divided by 0.40, or 2.5. This is simplified: taxes, refunds, payment fees, shipping subsidies, and other costs may change the real threshold.
Include costs that are easy to hide: discounts, free shipping, returns, damaged goods, payment processing, fulfilment, agency fees, creative production, and customer support. If agency and creative costs are shared across campaigns, report platform ROAS and a broader blended view so media efficiency is not confused with total acquisition efficiency.
Customer lifetime value can change the decision. A first purchase may be only moderately profitable, while repeat orders make the acquired customer valuable. Compare acquisition cost with observed or conservatively forecast lifetime contribution, and separate new customers from existing-customer revenue. Do not assign lifetime value to every campaign simply because repeat purchases are possible.
ROAS works well as a near-term signal when margins and order values are stable. Profit is stronger for budget allocation when product economics vary. If neither can be measured cleanly, improve the cost and revenue definitions before making fine-grained budget decisions.
8. Why do my ad platform numbers not match analytics or sales data?
Different systems often answer different questions, using different attribution rules and data-collection methods. A mismatch does not automatically mean that one system is broken.
Start with attribution windows. An ad platform may credit a conversion after a click within one period and after a view within another. Your analytics tool may use a different lookback window, last-click rule, data-driven model, or session definition. View-through conversions can add sales that occurred after an ad was seen but not clicked. They may be useful in some campaigns, but separate them from click-through conversions so their influence is visible.
Tracking gaps matter too. Consent choices, browser restrictions, ad blockers, mobile app limits, cross-device behaviour, deleted cookies, and privacy changes can prevent a conversion from being observed or linked to its original ad. Server-side events and modeled conversions may recover some signal, but they do not make the data complete.
Operational errors create further differences. A purchase event can fire twice, a reload can create a duplicate lead, a payment can fail after the event fires, or an order can later be refunded. The sales system may record net revenue while the platform reports gross order value. Reporting delays are common: leads may arrive quickly, while offline sales, qualification, refunds, and repeat orders appear days or weeks later.
Create a measurement note covering each system's conversion definition, attribution window, timezone, revenue basis, refund treatment, and reporting delay. Use the ad platform for delivery and optimisation signals, analytics for traffic and onsite behaviour, and the order or CRM system for authoritative business outcomes where appropriate. Reconcile systems at a consistent level rather than forcing identical totals. The goal is a trustworthy explanation of differences, not artificial agreement.
9. How often should I review these metrics?
Use different rhythms for delivery, optimisation, and business economics. Checking everything every day encourages overreaction, while checking only once a quarter can let simple problems waste budget.
Daily, check for delivery and measurement failures: unexpected spend, campaigns that stopped serving, budget pacing problems, broken links, rejected ads, large tracking changes, payment failures, and unusual conversion drops. These checks catch faults; they do not declare a winner from one day's results.
Weekly, review performance by campaign, audience, creative, placement, geography, and landing page when the sample is large enough. Compare spend, conversion volume, conversion rate, CPA, revenue, and profit or contribution margin. For lead generation, add sales acceptance and quality by source if the data has had time to mature. Make a small number of changes, record them, and allow enough time for results to accumulate.
Monthly, examine blended CAC, new versus returning customers, product mix, discounts, refunds, contribution margin, and channel interaction. Compare platform-reported conversions with orders or CRM outcomes and review whether the attribution window still fits the buying cycle.
Quarterly, or on a cycle that fits your sales process, assess retention, repeat revenue, customer lifetime value, payback period, and budget allocation. Long sales cycles may require cohort analysis rather than calendar-month performance.
Avoid reacting to tiny samples. A campaign with three conversions can show an attractive CPA by chance. Use minimum spend, conversion, or confidence rules that fit your economics, and flag small samples rather than presenting them as stable results. Match the reporting cadence to the speed of the decision and the delay before the outcome becomes reliable.
10. What should my performance marketing dashboard include?
A practical dashboard should fit on one screen or a short report and make the next action clear. Group metrics by the question they answer rather than copying an ad platform's layout.
Spend and pacing: show spend, budget, pacing against plan, and blended CAC or CPA. Use this view to adjust budgets or investigate unexpected delivery.
Delivery: show impressions, reach, frequency, CPM, clicks, CTR, CPC, and landing-page visits. Use it to investigate reach, creative fatigue, media cost, or technical losses. These metrics explain performance; they do not prove profitability.
Funnel health: show visits, leads or purchases, conversion rate at each defined step, qualified leads, sales acceptance, and checkout or form completion where relevant. Use it to fix the stage with the largest meaningful drop, test the message, or improve the landing page.
Business outcomes: show new customers, attributed revenue, average order value, pipeline value, closed revenue, and refund-adjusted revenue as appropriate. Use it to shift attention toward campaigns producing genuine customer or sales value.
Profitability: show contribution margin, profit after ad spend, ROAS, payback period, repeat revenue, and lifetime value or lifetime contribution. Use it to protect or increase investment in campaigns that meet the business's return threshold and reduce spend that cannot reach it.
For each metric, include the period, comparison, source, definition, and owner. Add a notes area for attribution-window changes, tracking incidents, promotions, and major budget changes. A useful weekly dashboard might contain 10 to 15 headline metrics, with drill-downs underneath. If a number never changes a decision, remove it. The dashboard should help you decide what to keep, fix, test, or stop.
Conclusion
Build reporting from the bottom line backwards. Choose the business outcome that defines success, connect it to conversion and cost measures you can act on, and add only the diagnostic metrics needed to explain movement. Start with one reliable weekly view of spend, conversions, CPA or CAC, revenue, and contribution margin. Add lead quality, retention, or lifetime value as the buying cycle gives you enough data. Ignore traffic numbers that do not survive contact with sales, orders, refunds, or profit. A useful report answers three questions: what happened, why did it happen, and what should you change with the next budget decision?
Frequently Asked Questions
What are the most important performance marketing metrics?
Start with conversion volume, conversion rate, CPA or CAC, revenue, and profit or contribution margin. Add qualified leads, pipeline, retention, or lifetime value when they fit your business model and buying cycle. Use reach, clicks, CTR, and CPC as diagnostics rather than proof of profitability.
Is ROAS better than CPA for measuring paid campaigns?
Neither is universally better. CPA measures the cost of a defined conversion, while ROAS compares attributed revenue with ad spend. Use profit or contribution margin when margins, refunds, discounts, or fulfilment costs make revenue alone incomplete.
What is a good conversion rate or CPA?
There is no useful universal benchmark because results depend on the offer, audience, channel, price, sales cycle, and margin. A good CPA lets the acquired customer produce acceptable profit or lifetime contribution, and a good conversion rate supports that target at a sustainable traffic cost.
Why are platform conversions higher than my sales records?
The platform may use a different attribution window, count view-through conversions, duplicate events, include orders that later fail or are refunded, or report modeled conversions where consent or privacy limits direct measurement. Compare definitions, windows, timezones, revenue treatment, and reporting delays before deciding that either source is wrong.