White-label tools · supported by real people
Marketing

When Google Ads look like they're working — and aren't

A campaign can look healthy in every chart Google shows you and still not be producing a single job. Impressions up, clicks up, spend up, click-through rate respectable — and a phone that doesn't ring. Here's how that gap opens up, and the numbers that expose it.

The metrics that reward you for activity

Impressions, clicks and click-through rate all measure how well the ad performs as an ad. None of them measure whether the business got anything. It's entirely possible to optimise all three while the account gets worse, because the cheapest clicks are usually the least valuable ones.

The classic version: a broad keyword pulls in a huge volume of cheap, vaguely-related searches. Click-through rate looks fine. Cost per click drops, which feels like a win. Meanwhile the people who were actually going to buy are a rounding error in the total.

Four numbers that tell you the truth

Cost per conversion

Not cost per click. What did one genuine enquiry cost? If you don't know, nothing else in the account can be judged, because you can't tell an expensive click from a cheap customer.

Conversion rate by search term

The search terms report — not the keywords report — shows what people actually typed. This is where you find that half your budget went to searches containing the word "free", or "jobs", or a competitor's brand name.

Impression share lost to budget

If you're capped out on budget for the terms that do convert, adding budget is the whole optimisation. Plenty of accounts are simultaneously overspending on rubbish terms and underfunded on good ones.

Phone calls, honestly counted

For most trades and local services the conversion is a call, and calls are the most commonly mismeasured thing in an account. If call tracking isn't set up, the account is being judged on form fills alone while the real result is invisible.

The things worth changing first

  1. Read the search terms report. Not the keywords — the terms. Add negatives for anything that clearly isn't a customer. This is usually the single biggest improvement available and it costs nothing.
  2. Match the landing page to the ad. An ad for emergency plumbing that lands on a homepage is asking the visitor to do the sorting. They won't.
  3. Set up conversion tracking properly, including calls, before changing anything else. Optimising an account you can't measure is guesswork with a budget attached.
  4. Narrow before you broaden. Tight match types and a smaller keyword set will usually beat a broad campaign for a local business, because a local business doesn't need volume — it needs the right dozen people.

The uncomfortable question

Sometimes the honest answer is that the ads are working fine and the website is the problem. If the campaign is sending relevant people to a page that loads slowly, doesn't say what the business does, or hides the phone number, no amount of bid tuning fixes that. It's worth ruling out the site before spending another month optimising the traffic pointed at it.

Want an honest, independent read on your marketing?

Free instant tools →
← Back to all articles