The First 90 Days of Google Ads

Who this is for

A local service business with a working website, somewhere between five hundred and fifteen hundred dollars a month of ad spend, and no in-house marketer. Everything here assumes Google Search rather than Display or Performance Max, because on a budget this size search is where the intent is and the other formats are where the money quietly goes.

Before day one: the two things that must exist

Conversion tracking. If you do not have it, nothing else in this guide can be judged. At minimum you need a phone call conversion, using a Google forwarding number with a thirty-second threshold, and a form submission conversion firing on a real thank-you page. Test both yourself before you spend a dollar.

A landing page per campaign. Not the homepage. One page per service, with the phone number visible without scrolling, the service named in the first heading, and a form that asks for three fields at most.

Weeks 1 and 2: build small and deliberately

Start with two campaigns at most, one per high-value service. Within each, three to five ad groups, each holding a tight set of keywords built on exact and phrase match. Ignore the recommendation to add broad match at this stage; you can add it in month two once you have a negative list to protect it.

Write three responsive search ads per ad group. Pin your main service and your town into headline position one on at least one of them, so that you always know what is being shown. Fill in every extension: sitelinks to your other services, callouts for the things that make you choosable, a call extension, and location if you have a premises.

Set manual CPC or maximise clicks with a bid cap for the first fortnight. Automated bidding needs conversion data it does not have yet.

Budget: spend at a steady daily rate rather than front-loading. Expect week one to look bad. It usually does.

Weeks 3 and 4: the negative keyword pass

This is the highest-value hour of the whole ninety days. Open the search terms report, sort by cost, and read every single term. Add as a negative anything that is:

• a job you do not do,
• a place you do not serve,
• a search for employment, training, DIY instructions or free anything,
• a competitor's brand name, unless you have deliberately chosen to bid on it.

Do this weekly for the first month, then fortnightly. On a typical new account the first pass removes twenty to forty percent of wasted spend.

At the same time, look at which ad groups have had no impressions. That is usually a bid or a relevance problem, and it is worth fixing before it distorts your read of the whole account.

Weeks 5 to 8: let it settle, then change one thing at a time

Once you have around thirty conversions in the account, switch to maximise conversions bidding. Not before; automated bidding with eight conversions will chase noise.

Now start improving in order:

1. Landing pages. Look at conversion rate by page. Anything under two percent for a local service is usually a page problem, not a traffic problem.
2. Ad copy. Check the asset report to see which headlines Google actually serves, and write two more like the winners.
3. Schedule. Pull the day and hour report and reduce bids where nobody answers the phone.
4. Devices. For most trades, mobile does the work. If desktop is converting at half the rate on a quarter of the spend, adjust.

Change one thing per week. Two changes at once means you learn nothing from either.

Weeks 9 to 12: judge it properly

Now the account has enough history to be assessed. Work out three numbers:

• Cost per enquiry. Total spend divided by conversions.
• Cost per customer. Cost per enquiry divided by the share of enquiries that become jobs. Ask whoever answers the phone; they know.
• Value of a customer. Average job value, times margin, times how many times a customer comes back.

If your cost per customer is comfortably below the value of a customer, the account works and the correct move is to raise the budget until that stops being true. If it is above, do not raise the budget. Fix the conversion rate first, then the offer, then decide whether this category is winnable at your spend level.

Things to ignore in the first ninety days

• Impression share. A vanity metric at this budget. You are not trying to own the auction, you are trying to buy the profitable slice of it.
• Quality score, in isolation. Useful as a diagnostic on a specific keyword, useless as an account-level target.
• Display and video. Both have a place. Neither has a place in a fifteen-hundred-dollar search budget that has not yet proven a cost per customer.
• Google's automatic recommendations. Read them, apply perhaps one in ten, and never leave auto-apply switched on.

The ninety-day checkpoint

Three outcomes are normal. The account is profitable and should be scaled. The account is close, and one more cycle on landing pages will decide it. Or the category is genuinely too expensive at your budget, in which case the honest answer is to move the money into local search and content, where the same spend compounds instead of stopping the moment you switch it off.

All three are useful answers. The failure is spending ninety days and not being able to tell which one you got.