An ad can earn clicks without producing profitable orders. Read the steps after the click before deciding what to change: landing-page views, carts, checkouts and purchases. Add-to-cart rate can help locate the drop, but it is still an intermediate event.

A click is a decision to look, not a decision to buy

A click tells you one thing. Somebody scrolling decided the next few seconds were worth spending on you. That's a real signal and it costs real money to get, so it feels like it should mean something about sales.

Curiosity can be enough to earn a click. Buying also depends on the product, price and what happens on the page. A high CTR therefore needs to be read alongside purchase outcomes and the cost of acquiring them.

What happened when I ranked one account's ads by what they returned

I reviewed a quarter of cold-ad data for a UK makeup-brush-cleaner brand. In that analysis, click-through and video hold did not show a useful relationship with return. Add-to-cart rate was the strongest observed predictor, followed by landing-page-view rate.

The individual ads were even blunter than the averages. The two best-returning cold ads in the account had among the worst retention of anything we ran, and the best-retaining ad in the whole account lost money. A large share of the cold spend sat in the highest click-through band, and that same band had the worst cart rate of any band.

That finding made cart rate worth investigating in this account. It did not establish that the click caused the weak return or that another account would show the same pattern.

The same thing held on a second account, month after month

I've watched the same relationship hold on Skydolls, a clothing brand, month by month for more than a year, where add-to-cart rate and conversion rate tracked return almost exactly. Different category, different price point, different creative style, same answer.

The repeated observation made cart rate a useful diagnostic in my reviews. It remains a correlation to interpret alongside purchases and costs. The clothing account article covers the operating context.

Read the funnel in order, top to bottom

Compare the funnel using consistent event definitions and comparable periods. Read each step against the step before it, then connect the purchase volume and order value to spend.

Compare ads only when they have enough delivery for a useful read. Cart rate can identify candidates for investigation; keep purchases and economics as the outcome you are trying to improve.

Check the explanation against the account

Objective, audience, page and tracking checks all belong in the investigation. Which one deserves action depends on the evidence in this account.

On the brush-cleaner account, older buyer groups had higher CTR and lower cart rate. That suggested a mismatch worth testing. The breakdown did not isolate whether creative, audience selection or another difference caused it.

What to do with an ad that buys clicks and sells nothing

Apply the agreed loss limit, then inspect the ad and page together before writing a replacement. If the hook creates an expectation the page cannot meet, changing the body alone may not resolve it.

A food-brand audit also found repeated arguments across the ads and little relationship between watch-through, CTR and return. That supported testing different reasons to buy rather than using those engagement metrics alone to select the next variation.

Run it weekly, in the same order, on purpose

Review comparable windows regularly and track which part of the funnel changed. Keep the observation separate from the explanation you intend to test. The weekly report should record both and the next action.

If purchases have stopped altogether, start with the no-sales checks for tracking and delivery before diagnosing the creative.

My creative strategy work for DTC brands includes research, briefs, review, testing analysis, and iteration. See the creative work and current service details.