Your first 90 days on Google Ads: the strategy to follow

In this guide · 11 parts

1. How long until Google Ads work for a home-services business?
2. What do you check every day, week and month?
3. Month one: collect the data and leave the settings alone
4. Month two: make the first real changes
5. Month three: scale what books jobs
6. What the routine looks like after day 90
7. Why do most contractors quit Google Ads too early?
8. How much should you budget to get through the first 90 days?
9. What’s normal versus a genuine red flag in the first 90 days?
10. How does Clique to Close line up with the first 90 days?
11. FAQ

How long until Google Ads work for a home-services business?

You can get clicks and calls within days of launch. What takes longer is a steady, trustworthy cost per booked job. That usually lands around month three. The first weeks feed Google’s bidding the data it needs, and your own weekly work trims the waste. So the honest answer is: fast for first calls, about 90 days for a stable number you can plan against. Treat it like a garden: month one you sow and collect the data, month two you make the first real changes, month three you harvest by scaling what books jobs.

Month 1 · collectMonth 2 · changeMonth 3 · scale

This page assumes the campaign is already live. If it is not, build it first with our how to run Google Ads walkthrough, then come back here on launch day.

Early results swing hard in both directions. A great first week is not proof it works. A dead first week is not proof it fails. Neither is enough data yet, which is why the plan below tells you what to do in each month and, just as importantly, what to leave alone.

One thing decides whether any of it pays off. Answering speed. Call a new lead within 5 minutes and your odds of qualifying them are 21x higher than at 30 minutes. A campaign that books calls nobody answers is wasted spend, and no amount of optimizing fixes it.

What do you check every day, week and month?

Here is the whole routine for the 90 days. The months below say what to change and when. This says how often to look. Pick one morning a week, like Monday, and keep it.

How often	What to check	Time
Every day	Answer every lead fast, and log each one: where it came from, and whether it booked. A shared spreadsheet is enough.	A minute a lead
Every week	1. Conversions are still recording: no day with clicks and zero conversions (how to test is in conversion tracking). 2. Spend is on track. Google can spend up to twice your daily budget on a busy day, but no more than 30.4 times it in a month.[8] 3. The search terms report: block the junk. 4. Any fake or junk leads, and where they came from.	About 30 minutes
Every month	1. Cost per booked job: the month’s spend divided by the jobs your log says booked. 2. The recommendations tab: apply the repairs, dismiss the rest. 3. From month two, the location report. 4. From month three, impression share and auction insights.	About an hour
Every quarter	The full 15-point audit.	One to two hours

What you leave alone matters just as much. Each month below lists it.

Month one: collect the data and leave the settings alone

Month one has one job. Gather clean data and prove the foundation works. Google’s bidding is in its learning phase: the first week or two after launch, or after any change to the bid strategy, while it works out what to bid. It has almost nothing to go on yet. Expect uneven results and a cost per lead that bounces around. Your work this month is checking, not fixing.

Do this in month one:

• Send a test lead in week one and confirm it records as a conversion. Do this before you look at any other number. See how to set up conversion tracking.
• Answer every call and log which ones turned into a booked job. Tell whoever picks up the phone that these calls are paid for, and have them note where the lead came from and whether it booked. This is the data nobody else has, and by month three it is what you optimize on.
• Read the search terms report once a week. Add negatives only for searches that are obviously not your job: jobs, salary, DIY, a service you do not offer. That is the one change that does not disturb the learning.
• Write down your own numbers. What a booked job is worth to you, and the most a lead can cost before the job stops paying.
• Check that AI Max, text customization and the campaign-level broad match setting are all off. Where to look is below.

Leave these alone in month one:

• The bid strategy and your Target CPA. Changing either restarts the learning and sets you back weeks.
• The daily budget.
• Your keywords. A keyword with 12 clicks and no call has told you nothing yet.
• Your ad copy.

The only trigger to act in month one is a broken foundation. Real clicks arriving with zero conversions recorded means the tracking is wrong, not the campaign, and that gets fixed the day you spot it. So does a campaign that has spent nothing at all after three or four days. Check the ad first, because an ad still under review cannot show. Then check your Target CPA and cost-per-click cap are not far below what clicks cost in your market, because then Google cannot bid at all. If it is spending, just slowly, give it the two weeks the bidding guide asks for, then raise the cap a little. Everything else waits.

That last checklist item is worth a minute of your time. Google is moving advertisers onto AI Max, its newer matching system. Since September 2026, Google has been upgrading Search campaigns to AI Max on its own when they use the campaign-level broad match setting or text customization, the setting once called automatically created assets.[2][4][5] The launch walkthrough leaves AI Max off on a new account, so all three should be off already. If one is on, switch it off now. Otherwise the upgrade can land in the middle of the stretch this plan asks you to hold your nerve through, and you will watch your cost per lead move and wonder what broke. Your keywords and your negative keywords keep working either way.

Month two: make the first real changes

Month two is when you start cutting. The account now has enough data to show where the waste is, and the learning phase has settled enough that careful changes will not knock it over. Work from the report, not a hunch. Small, evidence-led changes. Not a rebuild.

Do this in month two:

• Mine the search terms report properly. Sort by cost, read the top of the list, and add a negative for every search that spent money and could never book a job. See the weekly search terms ritual and the negative keyword master list.
• Prune the location report the same way. Open Matched locations, sort by cost, and find the towns that spend and never book, and the fringe ones where the drive time eats the margin. Then remove or exclude them in the campaign’s location settings, as in location targeting.[6]
• Move money toward what books jobs. A budget belongs to a campaign, not an ad group. With one campaign, pause the keywords that have had a fair run and never booked (the 80-click rule below), so the same budget goes further on the ones that do. With two or more campaigns, shift budget toward the one that books.
• Work out your real cost per booked job: last month’s spend divided by the jobs your log says booked. Compare it with the number you wrote down in month one.

Leave these alone in month two:

• The bid strategy type. Do not swap Target CPA for something else because one week looked bad.
• Any keyword that has not yet had a fair run of clicks.
• The account structure. If it needs rebuilding, that is a month-three decision, not a month-two panic.

The trigger we work to is patience with a threshold. If you expected a booked job from a keyword by 50 clicks and there is still nothing by 80, that is the signal to act. Say “roof replacement cost” has taken 80 clicks of your money and booked nothing. Pause it, or move it to its own ad group with tighter copy. Not before 80. Changes made to feel busy just slow the account down.

If you do adjust your Target CPA, change it once and let it settle before you touch it again. Two changes in two weeks tell you nothing, because you cannot see which one did what.

---

Month three: scale what books jobs

Month three is where it comes together. The learning phase has settled, the waste is trimmed, and the cost per booked job is finally steady enough to trust. Now you put more money behind what works. This is the month the account starts to look like a machine, not a gamble.

Before you spend a dollar more, check the other side of the business. Can your crews take the extra work? Scaling a campaign that books jobs you cannot service turns into long lead times, canceled quotes and bad reviews. That costs you more than the ads ever made. If you are already booked three weeks out, the month-three move is a higher price or another crew, not a bigger budget.

Do this in month three:

• Raise the daily budget on the campaigns and keywords that are booking real jobs, once the cost per booked job has held steady for a few weeks and you have the capacity to service them.
• Ease off the ones that never booked, even if their clicks were cheap. A keyword with cheap clicks that never books is worse than a dear one that fills the calendar.
• Feed your booked jobs back into the account so the bidding optimizes toward work, not just leads. See how to send booked jobs back to Google.
• Check your impression share on your best searches. On the keywords table, click the columns icon, open Competitive metrics, tick Search impr. share and apply.[7] Auction insights shows how to read it.

Leave these alone in month three:

• Anything that is working. Scaling is adding budget, not rewriting.
• The urge to chase a cheaper cost per click. Judge everything on booked jobs.

There is a ceiling to watch for, and it is the trigger for the next decision. Once you hold more than 80% of the impressions for a search, your costs climb. You are near the top of that market. The way up from there is a new service line or a wider area, not more budget on the same searches.

What the routine looks like after day 90

The 90-day plan ends. The weekly work does not. From here the account runs on a rhythm, and the rhythm is short.

Every week, read the search terms report and add negatives. That single habit is where most of the money is found, and it is why a worked account and a neglected one look nothing alike by month six. Every quarter, run a full check on the account. Our audit checklist is the same 15 points we start with.

If someone else runs the account, ask for that search terms report every month. Not a summary, and not a dashboard with a green arrow on it. The actual list of searches you paid for. Read down it and you will know in two minutes whether anyone is doing the weekly work.

Why do most contractors quit Google Ads too early?

Because they judge the channel by weeks two to four. That is almost always the worst stretch of any well-run campaign. The account is still learning, the waste is not yet cut, and the cost per lead looks ugly. An owner who pulls the plug then never sees month three, where the same account would have settled and paid off.

It feels rational in the moment. The spend is real and the results are thin. But quitting at week three is like judging a new hire on their first shift. The data is not in yet. The owners who win are the ones who set that money aside as a data-gathering cost and let the plan run.

That does not mean blind patience. It means patience with a plan and weekly work. A campaign left alone for 90 days is not the same as a campaign worked weekly for 90 days. One drifts. The other improves.

---

How much should you budget to get through the first 90 days?

Budget enough to buy real data across the three months, not just week one. The bidding needs a steady flow of conversions to settle. Around 15 conversions a month (tracked calls and forms) is a slow start, and around 50 steadies it. At the January 2026 blended cost per lead of $104, that many leads costs real money. Funding it is what gets the account out of the learning phase.

For most small contractors that means $1,500 to $3,000 a month, held steady for the full 90 days. Cutting the budget mid-learning to “see if it works” means it never will. The account never gathers enough to prove itself. Fund the whole runway or do not start.

Work the exact number backward from your own jobs and close rate. The budget guide shows the method and the learning-phase floor in detail.

What’s normal versus a genuine red flag in the first 90 days?

Some early pain is normal. Some is a real warning sign. Knowing the difference keeps you from quitting a good account or clinging to a bad setup. Normal is fluctuation and a learning phase. A red flag is a broken foundation or an account nobody is touching. Here is how to tell them apart.

Normal in the first 90 days	Genuine red flag
Cost per lead swings week to week	No conversion tracking, so nothing is measured
A slow, uneven learning phase	No changes made week after week
Some junk searches before you prune them	The search terms report never gets checked
A dead week followed by a strong one	You cannot get a straight answer from whoever runs it

If you see the left column, stay steady. If you see the right column, that is not bad luck, that is a broken setup. A silent account with no tracking will not improve on its own. Fix the foundation yourself, or put it in the hands of someone who will do the weekly work. Our guide to why ads stop working walks through the diagnosis.

How does Clique to Close line up with the first 90 days?

The first 84 days are the build: three 28-day cycles to set the account up and get it working. That is the same runway this guide describes.

That is why we never judge the account on week one. The channel needs the runway. If you want to see where your jobs are going before the clock even starts, run the Google Journey.

Free · no cardSee where your jobs are going.The Google Journey searches Google like a homeowner and shows who gets the call.Run my free Google Journey →

FAQ

How long does it take for Google Ads to become profitable? You can get calls in days, but a stable, profitable cost per booked job usually takes about three months. The first weeks are a learning phase. Judging profit before month three is judging incomplete data.

Why is my cost per lead so high in the first month? Because the account is still learning. In month one the bidding has little data and spends unevenly. Cost per lead almost always starts high and comes down as you prune waste and the learning phase settles. It is normal, as long as tracking is firing and someone is working the account weekly.

Should I make changes in the first week? No. Aggressive changes in week one reset the learning phase and set you back. Watch that calls are coming and tracking is firing, and add negatives for the obvious junk searches. Save the real refining for month two, once there is data to act on.

When should I worry that Google Ads isn’t working? Worry about the setup, not the calendar. No conversion tracking, no weekly changes, no straight answers from whoever runs the account. That is a red flag at any point. Fluctuation and a slow learning phase are not.