The Ad MeterAll writing

Where Google Ads budget actually leaks

· 9 min read

Every plumber knows the leak test. Shut off every tap in the building, go out to the meter, and watch the low-flow indicator. If it is still turning, water is going somewhere you are not using it.

A Google Ads account works the same way. The spend runs continuously whether or not anything at the other end converts. The difference is that a water meter is in your front garden and the ad account is behind a login most owners open twice a year. This post is the shut-off test: five specific places money escapes, what causes each one, and how to look for it yourself.

1. The gap between your keywords and the actual searches

This is the single most common leak, and it exists because of a distinction most account owners never have explained to them: the keywords you add are not the searches you pay for.

A keyword is an instruction about which searches you are willing to match. How loosely it matches depends on its match type. Broad match will show your ad for searches Google considers related, which can be a long way from the words you typed. Phrase match requires the meaning of your phrase to be present. Exact match is the tightest, and even it matches close variants — plurals, misspellings, reorderings, and terms Google treats as the same intent.

So an account bidding on "emergency plumber" on broad match can end up paying for "how to fix a leaking tap", "plumbing apprenticeship", or "plumber salary". Those are not customers. They are someone doing homework, looking for a job, or fixing it themselves. Each click costs the same as a real one.

How to check it

Open the Search Terms report — not the Keywords tab, the Search Terms report. Keywords show what you asked for. Search terms show what you actually bought. Set the date range to the last 90 days, sort by cost descending, and read the top fifty. You are looking for terms where you would not want to pay for that click.

If reading that report is uncomfortable, that discomfort is the finding. Most accounts have never had it read line by line.

2. A negative keyword list nobody maintains

Negative keywords are the instruction for searches you never want to match. They are the shut-off valve. In most accounts they are either empty or were written once during setup and never touched again.

The problem with a one-time negative list is that the ways people search change, and Google's matching changes with them. A list written two years ago does not know about the terms costing you money last month. Negative keywords are not a setup task. They are maintenance, and they are the direct output of reading the search terms report regularly.

There are also negatives worth adding that have nothing to do with irrelevance. Terms containing "free", "cheap", "DIY", "how to", "salary", "jobs", "course", and "training" are usually people who will not buy a service. So are competitor brand names, unless you have deliberately decided to bid on them and have thought about what that costs.

How to check it

Find your negative keyword lists and look at the date they were last modified. Then compare that list against the search terms report from the last quarter. The terms that appear in the report and not in the list are the gap.

3. Conversion tracking that counts the wrong thing

This one is worse than a leak, because it corrupts every decision made downstream. If the account is optimising toward a number that does not represent real business, it will confidently spend more money getting you more of the wrong thing.

The failure modes are specific and common:

  • The conversion fires on a page view rather than a form submission, so every visitor who reaches the contact page counts as a lead.
  • Phone calls are the main way customers actually get in touch, and calls are not tracked at all — so the campaigns driving calls look like they do nothing.
  • The same conversion counts more than once per person, inflating the total.
  • Soft actions like newsletter signups or PDF downloads are counted alongside genuine enquiries, with no distinction in value.
  • Conversion tracking was installed, then the website was rebuilt, and nobody reinstalled it.

A campaign using automated bidding is trying to produce more of whatever you told it a conversion is. If that definition is wrong, automation makes the problem bigger and faster.

How to check it

Look at the conversion actions in the account and ask, for each one, whether a real person contacting your business is required for it to fire. Then submit your own form and place your own call and confirm both are recorded. This takes twenty minutes and is the highest-value check on this list.

4. One campaign doing three jobs

Budgets are set at campaign level. Everything inside a campaign competes for the same pot. So when one campaign contains several different services, the one with the highest search volume quietly eats the budget — regardless of which one is worth the most to you.

For a plumbing business, "blocked drain" might have far more volume than "bathroom installation". If they share a campaign, drains will consume the budget while the higher-value work gets whatever is left. Nothing is broken. It is doing exactly what you configured. It is just not what you wanted.

The same applies to geography and to intent. Emergency searches and research searches behave completely differently, and they should not be sharing a budget or a bidding strategy.

How to check it

List your campaigns and, for each, write down what job it is meant to do and how much of the account's spend it took last month. Where one campaign has several jobs, look at the spend split across its ad groups. If your most profitable service is not getting a proportionate share, you have found it.

5. The ad and the page do not answer the search

You pay for the click whether or not the page delivers on it. Someone searching for emergency drain clearing at 11pm, who clicks an ad about emergency drain clearing and lands on a generic homepage with a contact form at the bottom, has cost you money and left.

Relevance is also not only a conversion issue. Google assesses expected click-through rate, ad relevance and landing page experience, and those assessments influence what you pay and how often you show. Poor alignment costs twice: fewer conversions from the clicks you get, and worse economics on every click.

How to check it

Take your five highest-spending keywords. For each one, search it yourself, look at your own ad, click it, and then ask honestly whether the page that loads answers the thing you just searched for within a few seconds and offers an obvious way to get in touch. On a phone, not a desktop, because that is where the traffic is.

Doing the full test

None of the above requires an agency. It requires an afternoon and a willingness to read reports that are not flattering. In order:

  1. Read the last 90 days of the search terms report, sorted by cost.
  2. Confirm every conversion action requires a real human enquiry to fire.
  3. Check that phone calls are tracked, if calls are how your customers contact you.
  4. Compare campaign budget split against which services actually make you money.
  5. Click your own top five ads on a phone and judge the landing experience.

If you would rather have someone else read the meter, that is the free audit. Send two fields and you get the findings in writing within 48 hours, whether or not you ever hire us.

Have us read your meter

Two fields, and you get a Google Ads audit, a brand kit and a landing page within 48 hours. Free, and yours to keep whether or not you ever hire us.

Read my meter — free

Also worth reading