How Long Should I Give a Campaign Before Judging It?

Time is a proxy for the thing that actually matters, which is how many outcomes you have collected. A month at a large budget in a busy category is a far better read than three months at a small budget in a quiet one, and the calendar cannot tell you which you have.

Consider two accounts, and treat these figures as illustrative rather than as anything to copy. One produces forty enquiries in a month; if two of them had gone the other way the answer barely moves, and you can trust it. The other produces four; one lucky enquiry doubles the apparent performance and one unlucky week halves it. The second account has not told you anything about your market. It has told you about a coin toss.

If two enquiries either way would change your verdict, you do not have a verdict yet.

The practical move is to estimate, before you start, roughly how many enquiries the budget should produce in a month. If that number is small, you have three options: run for longer, concentrate the budget on a narrower market so the same money produces a denser read, or accept that the decision will be slow and set the review date accordingly. What you cannot do is get a fast, reliable read out of a thin one.

Some of the delay is not you being patient. It is the platform gathering the information it needs to make decisions.

Google Ads Help documents that an automated bidding strategy enters a learning state after significant changes, while it collects enough data to work with. Performance during that period is not representative, and it is a poor basis for a decision. The trap is not the delay itself — it is that every substantial change restarts it. An owner who adjusts the budget weekly can hold an account in a permanent state of learning and then conclude, quite reasonably, that it never worked.

There is a second, slower clock underneath. If your work has a lag between enquiry and booking — a quote to prepare, a decision to make, a job to schedule — then the outcome data that tells you whether the enquiries were any good arrives weeks after the enquiries did. A campaign judged on enquiry volume at four weeks may look excellent and be producing work you will not want by week ten.

Both clocks argue for the same discipline: change one thing at a time, and give each change enough room to be attributed to something.

Two campaigns with the same budget can need very different windows. These are the factors that decide which one you have.

Only one thing shortens the timeline reliably, and it is narrowing. A budget concentrated on fewer services, a smaller area, or the hours you can actually answer the phone produces a denser set of outcomes from the same money, and a denser set is a faster read. Spreading the same budget across everything is what produces the six-month engagement that ends with nobody able to say what happened.

Seasonality deserves its own check. Google Trends is free and settles in five minutes whether the period you are judging sits in a peak or a trough for your category in your state. Judging a campaign during a predictable annual dip, and then rebuilding it, is how businesses take credit for a recovery that was always coming.

The single cheapest discipline available here costs one conversation and it happens before anything is spent.

Four things to write down in advance

1. The number. How many enquiries constitute a readable result, given what the budget should produce.
2. The date. When you will look, and what you will deliberately ignore before then.
3. The measure. Cost per booked job if you can record it, cost per qualified enquiry if you cannot. Decided once, not renegotiated at review.
4. The threshold. What figure counts as continue, what counts as change, and what counts as stop.

Written down in advance, these four turn the review into an arithmetic exercise. Decided at the review, they turn into an argument, and the person who is most tired of the campaign usually wins it.

The related decision — how much to commit in the first place — is worked backwards from the value of a job rather than forwards from what feels affordable, and it is set out in the piece on setting the budget. If the threshold and the budget were decided by different logic, they will disagree at review.

Waiting for the performance verdict does not mean looking at nothing. Several things are readable immediately, and all of them are mechanical rather than statistical.

• Whether ads are actually serving. Approval, spend and impressions in the first days answer this. Nothing is being learned while nothing is showing.
• Whether the enquiry count is working. A test enquiry proves the plumbing on day one, and a campaign optimising against a broken count is worse than one not running.
• Whether the searches make sense. Reading what people actually typed tells you within days whether the campaign is answering the question you meant to answer.
• Whether the page is doing its job. Visits arriving and nobody enquiring is a landing problem, and it is visible long before the cost figures are trustworthy.

What is not readable in week one is cost per booked job, whether the enquiries were any good, or whether the market is worth being in. Those need volume, and volume needs time. Keeping the two lists separate is most of the skill here — early operational checks are diligence, early performance verdicts are guessing with a spreadsheet open. Confirming the count works before the money is committed is the first job of the measurement setup, not something to discover at review.

Discipline is not the same as stubbornness. There are situations where stopping before the review date is correct.

Stop when the spend itself is the emergency and the business cannot carry it. Stop when the enquiries are actively wrong — not merely disappointing, but for work you do not do or in places you do not go — because more time will produce more of the same. Stop when something structural has changed underneath: the phone number is wrong, the page has gone down, the service is no longer available. And stop when you discover the count was broken for the whole period, because you have not been measuring anything and the honest move is to fix it and restart the clock.

What is not a reason to stop: a bad fortnight, a competitor appearing, a week with a public holiday in it, or a single expensive click. Those are the texture of the thing rather than evidence about it.

For contrast, it is worth holding both timescales in mind at once. Google Search Central's own guidance is that SEO changes take four months to a year to show their effect, which is why the organic timeline is measured in quarters. Paid search is the fastest channel a small business can read, and even it needs more than a fortnight. If somebody has promised you a verdict sooner than that, they are promising you a coin toss with a report attached — and the account will still be there when the real answer arrives.

Is one month long enough to judge a Google Ads campaign?

It depends entirely on how many enquiries that month produced. Forty is a read; four is a coin toss. Work out in advance roughly how many the budget should generate, and if the answer is small, either extend the window or concentrate the budget on a narrower market so the same money produces a denser result.

Why did performance get worse right after we made changes?

Google Ads Help documents that automated bidding strategies enter a learning state after significant changes while they gather data, and performance during that period is not representative. The deeper problem is frequency: changing something every week can hold an account in that state permanently.

Should I pause a campaign that is not working yet?

Rarely before the review date you agreed. Pausing destroys the comparison you need to prove what fixed it and restarts any learning the account has done. Stop early when the spend is an emergency, when the enquiries are for work you do not do, or when something structural has broken.

How do I stop myself judging it too early?

Write the number, the date, the measure and the thresholds down before the money goes in. Decided in advance, the review is arithmetic. Decided on the day, it is an argument, and it is usually won by whoever is most tired of the campaign.