When to pause a campaign instead of optimizing it

The hard part is not knowing this. It is that pausing feels like quitting, and there is always one more thing to try.

I have watched owners spend a full quarter on a campaign they already knew was wrong, because stopping would have meant admitting the last three months were a loss. The last three months were a loss either way. The only decision available was whether to add a fourth.

What is the actual test for pausing?

Two questions, and you need honest answers to both.

Do I have enough data to conclude anything? If the campaign has produced a handful of leads over a few weeks, you do not have a verdict. You have noise. Small numbers produce confident wrong conclusions, and pausing on them is as much a mistake as continuing.

Is the gap a margin or a multiple? Calculate cost per paying customer, not per lead. Compare it to your gross margin on an average job. If cost per customer is somewhat above margin, that is tuning territory. If it is several times margin, tuning will not get there.

The second question is where people deceive themselves. They compare cost per lead to job value, which is a flattering comparison twice over: most leads never become jobs, and job value is not margin. Run it on customers and on margin.

Optimization is for closing a gap of tens of percent. Pausing is for a gap of multiples.

What does that test look like in numbers?

Six lines, off your own invoices and one export.

Pull the window. Ninety days of spend, say $6,000.

Count the leads. Say 58. Cost per lead is $103 derived, which for context sits near LocaliQ's published 2026 average cost per lead of $90.92 for Home and Home Improvement. Those LocaliQ figures are medians across a campaign sample, LocaliQ sells advertising management, and the 2026 edition does not publish its sample size. The 2025 edition of the series disclosed 16,446 US campaigns. Use them for orientation, never as a target.

Count the customers. From invoices, not from the ads account. Say 11, which is a 19% close rate (derived) on leads.

Divide. $6,000 across 11 customers is $545 per customer (derived).

Price the margin. Average job bills $1,400 at a 38% gross margin, contributing $532.

Read the gap. You are losing $13 per customer. That is a margin problem. Four tightening moves could plausibly close it, and this is a campaign to fix rather than stop.

Now rerun with five customers instead of eleven. Cost per customer is $1,200 derived against the same $532 of margin. That is more than twice your margin, and no combination of negative keywords, bid strategy and ad copy closes a gap that size. That is a campaign to pause.

The same arithmetic tells you something about your sample. Eleven customers and five customers are both small numbers, and the close rates behind them, 19%derived and 8.6%derived, are not reliably distinguishable at that volume. So the multiple, not the close rate, is what you act on. A gap of a few percent inside a small sample is not a finding. A gap of two and a half times is visible through the noise.

What should I try before I pause?

A short, ordered list. If none of it moves the number in thirty days, stop.

Cut the wrong searches. Read your search terms report and add negatives aggressively, because negative keywords save money without costing you anything you wanted and it is the fastest available improvement.

Tighten geography to the area you actually serve, and hours to the hours you answer.

Filter the ad copy so it names your minimum job, your specialty, or your area.

Fix the destination. If the page is generic, point the traffic somewhere specific. Sometimes the fix is more information rather than less, and long pages sometimes convert better than short ones when the purchase is considered.

That is the list. Four moves, thirty days, one decision at the end. What you should not do is keep making small bid adjustments indefinitely, because bid tuning is the least powerful lever available and it feels like work.

How long should I wait before deciding?

Long enough to accumulate customers, and I want to be honest that nobody publishes the right number.

No study with a disclosed sample tells you how many weeks a local campaign needs before a verdict is fair. What exists is arithmetic. If your campaign produces four customers a month, a one-month window cannot separate a bad campaign from an ordinary month. If it produces fifteen, it can. Divide the customers you need for a readable number by your monthly customer count and that is your window, in months.

Two mechanical constraints sit underneath that.

Google's documentation notes that some Smart Bidding strategies rely on a minimum volume of historical conversion data. A campaign that has been running for three weeks with six conversions has not given the system enough to work with, so what you are measuring is partly the absence of learning rather than the quality of the campaign.

And every significant change restarts part of that clock. Changing bid strategy, budget or conversion actions gives the system a new problem to solve. A campaign you have been adjusting weekly has never had a stable observation window, which means the ninety days you thought you were measuring were really nine ten-day experiments.

There is also a market-level check worth doing before you blame yourself. LocaliQ's 2026 data records cost per lead falling across all industries for the first time in five years. If your cost per lead rose over a period when the published direction was down, that is a signal about your account rather than about paid search. Read it as direction, not as your number.

When is pausing the right answer even though the campaign works?

Three situations that catch people out.

You are at capacity. More leads are a cost when you cannot serve them. Pausing during a full stretch protects your reputation and your team. Restart when the schedule opens.

The season ended. Spending into a period when nobody buys is not persistence. Set an end date in advance rather than noticing later.

The campaign works but a different one works better. Budget is finite. A campaign clearing a low bar can still be the wrong home for money that would do more elsewhere. That is a reallocation, not a failure.

There is a fourth that is harder to accept. Sometimes paid search is simply not where your customers decide. Some trades are dominated by proximity and reputation in the map results, and the closest contractor does not always rank first for reasons that have nothing to do with ads. If the map pack is the battleground for your service, ad spend may be renting attention you could earn.

How do I pause without losing what I learned?

Pause deliberately, not by deleting.

Pause the campaign rather than removing it. The history stays, the conversion data stays, and if you restart, you restart with a record instead of from zero. The same applies to your audience segments, which keep collecting while the campaign sits idle and are among the first things lost when an account changes hands.

Write down what you concluded and when. What the cost per customer was, what you tried, what changed and what did not. In six months you will not remember, and the version you remember will be kinder than the truth.

Keep the tracking running. A paused account with working tracking still tells you what your organic traffic does, which is the baseline you will need if you ever restart.

If you want to stay alert to conditions changing without checking manually, simple alerting handles it, and the scripts and automations worth using include monitors that tell you when something moves.

Then reconsider the whole question rather than the campaign. Whether a small business should be running Google Ads at all depends on three conditions, and a failed campaign usually means one of them was never true.

What to do this week

Pull ninety days of spend and count how many paying customers came from it. Not leads. Customers, from your invoices.

Run the six lines above. Write cost per customer and gross margin on the same line and look at the ratio, not the difference.

If the gap is a multiple, set a thirty-day deadline, run the four fixes listed above, change nothing else during those thirty days, and put the decision date in your calendar now. A decision with a date attached is a decision. Without one it becomes another quarter.

If the numbers are too small to read, do not pause and do not celebrate. Extend the observation window and stop making changes that reset it.

Questions about whether to pause or push?

Email me at eric@seod.com with ninety days of ad spend, the number of leads, how many of those leads paid you, and your gross margin on an average job. I will run the six lines and give you a straight answer on pause, fix, or keep going.

I will also tell you if the numbers are too thin to support a decision, which happens more often than owners expect and is worth knowing before you act on them.

Or keep reading more on paid search for local businesses.