Google Ads Is Recording Conversions. Why Are You Still Losing Money?

You open the account and see conversions coming in. The cost per conversion looks manageable. Yet the sales team says the enquiries are poor, or the shop's bank balance tells a different story.

That gap deserves investigation. Changing keywords or bids immediately can leave the underlying problem untouched.

First, find out what the conversion actually means

Open the campaign's conversion goals and inspect the individual actions behind the total. Write down what triggers each one, where the data comes from and whether it represents a purchase, a lead or an earlier interaction.

A contact-button click, a submitted form and a paid order describe different stages. Combining them in one total makes it difficult to judge customer acquisition.

Google distinguishes primary and secondary conversion actions. Primary actions can be used for bidding when the campaign uses their associated goal; secondary actions generally provide observation rather than bidding signals. There is an important exception: actions included in a custom goal are used for bidding when that custom goal is assigned, even if an action is marked secondary. Check the campaign's actual goal configuration, not just the label beside an action. Google explains these settings.

For a sales campaign, ask whether its bidding goals reflect the outcome you want. Earlier interactions can be useful to report, but selecting them as bidding signals is a decision that needs a reason. Google's conversion-goal documentation explains how account-default and campaign-specific goals apply.

Test one real journey before trusting the total

Follow the action yourself, using a controlled test lead or order that you can identify later. Confirm that a form records success only after submission succeeds. Check that a purchase sends the correct amount, currency and unique transaction ID. Revisit the confirmation page and look for repeat firing.

For purchases, Google can deduplicate two events with the same transaction ID within the same conversion action. The ID must be unique to that transaction. Reusing a fixed ID can undercount sales. This does not mean two separate purchase conversion actions will automatically be deduplicated against each other. Google's transaction-ID guide covers the implementation.

Also inspect whether the same business outcome is measured through multiple actions, such as a Google Ads purchase tag and an imported Analytics purchase event. Establish which action should inform bidding and which, if any, is needed for comparison.

The counting setting is a separate choice. “One” counts one conversion per ad interaction for that action; it does not identify one unique person across your business. “Every” can count multiple purchases after an interaction. Choose the setting that fits the action and use order IDs to prevent duplicate purchase records. Google's counting options explain the distinction.

A cheap lead can be an expensive customer

Consider this hypothetical service business. These figures illustrate the calculation; they are not industry benchmarks or AdPilot results.

The reported cost per enquiry is £30. The advertising cost per customer is £400. Three customers contribute £900 before advertising, so the campaign leaves a £300 shortfall after its £1,200 ad spend, before fixed overheads.

Here, contribution means revenue less the direct and variable costs of delivering the work. If those costs are missing, the £300 estimate needs correcting too.

Keep the enquiries together as a group and allow time for follow-up. Record which were unsuitable, which were lost and which became customers. A low close rate could reflect poor targeting, an uncompetitive offer or missed sales follow-up. The number alone does not tell you which.

Google supports qualified-lead and converted-lead goals for deeper outcomes. A qualified lead has been assessed beyond the initial enquiry; a converted lead has reached a business-defined later stage. If you return these outcomes to Google, define those stages consistently and verify the records being imported. Do not relabel every submitted form as qualified.

A 3× return on ad spend can still lose money

Now consider a hypothetical online shop with £3,000 in net sales attributed to £1,000 of ad spend. Its revenue ROAS is 3×, or 300%.

Suppose £2,250 goes on product costs, fulfilment, payment fees and other variable costs. That leaves £750 before advertising and a £250 shortfall after advertising, before fixed overheads. For this example, revenue excludes tax and is already net of refunds; costs do not include refunds a second time.

With a 25% contribution margin before advertising, revenue ROAS must reach 4× just to cover advertising in this simplified example: 1 ÷ 0.25 = 4. Fixed overheads and a desired profit would require more. This is an accounting threshold, not a recommended bidding target or a promise that Google can achieve it.

Use a consistent revenue basis when comparing the store and the ad account. Check discounts, tax, cancellations and returns. Google provides conversion adjustments to restate values or retract eligible conversions after the original event. A refund in the shop does not prove the ad report has been updated.

If the business accepts an initial loss to acquire repeat customers, document the expected repeat contribution and payback period. Base that expectation on observed customer behaviour, not hoped-for lifetime value. Attributed sales also do not establish how many sales the advertising caused; incrementality is a separate question.

What to inspect across the six campaign types

The financial test stays the same, but the evidence available and the appropriate goal vary. Use this table as a starting point, not as a claim that every format should be judged by immediate sales.

Choosing the campaign format is part of setup, but it does not settle the measurement question. Google's campaign-type guide describes the different jobs campaigns can perform. Our campaign-building guide covers the structure that sits underneath the ads.

Give delayed conversions time to arrive

Recent spend may be complete while the conversions associated with it are still arriving. Google explains that conversion delay can make recent performance appear weaker than older periods.

Use a review window that reflects your actual sales cycle. Compare the same group of leads or orders, and check the reporting date basis before reconciling systems. Do not demand exact daily agreement between reports with different attribution rules.

Waiting for delayed sales is different from ignoring a verified fault. A broken form, incorrect purchase value or unsuitable bidding goal needs attention when you find it.

Decide what to change from the evidence

If measurement is wrong, repair it before using those figures to justify more spend. If genuine leads are unsuitable, inspect the searches, targeting, offer and qualification process. If customers buy but contribution is inadequate, examine product mix, costs and acquisition economics before chasing more conversion volume.

AdPilot's Optimize walkthrough shows how account data can identify campaign work such as wasted search terms and weak ad copy. That can help you act on a finding. It does not replace the order records, margins and sales outcomes needed to judge profit. For irrelevant searches, use our negative-keyword guide; for ongoing checks, use the Google Ads audit checklist.

The first useful improvement may be a clearer definition of success. Verify what is being counted, connect it to the customer outcome and calculate what remains after costs. Then choose the campaign change that addresses the problem you found.

Common questions

Does a Google Ads conversion mean a sale?

Only if that conversion action measures a sale. Inspect the action and its trigger; the total may include enquiries or other configured events.

Is a higher ROAS always better for the business?

ROAS depends on the values reported. Revenue ROAS does not subtract fulfilment, product or advertising costs. Compare it with contribution, total volume and the business's objectives.

Should I pause a campaign as soon as it appears unprofitable?

First establish whether the result reflects a tracking fault, incomplete sales outcomes or a genuine shortfall. Confirmed unaffordable spending needs prompt action; a fresh report alone may not show the eventual result. Set an affordable review budget and a clear decision rule before continuing an experiment.