The Google Ads defaults that quietly spend your budget

The Display Network checkbox nobody unchecks

When you create a Search campaign, Google offers to include Display Network and search partners, and the boxes are checked for you. Those clicks are cheaper, which is exactly why they are dangerous on a small budget. They arrive from apps and content pages where nobody was looking for a plumber, and they can absorb a large share of a modest daily budget before the search side has spent anything.

Turn both off in a Search campaign. To see what they were doing, segment your campaign table by network and compare conversions per thousand clicks. In most of the small New Jersey accounts we inherit, the Display side has plenty of clicks and almost no calls.

Location targeting is not what the name suggests

The default location setting is presence or interest, which means people in your area, people regularly in it, and people anywhere in the world who have shown interest in it. That last group is how a Middletown contractor ends up paying for clicks from someone in Arizona reading about the Jersey Shore.

Open the campaign's location settings, choose presence, and save. While you are there, add exclusions for anywhere you will not drive to. Then check the geographic report every few weeks, because a service area drawn by radius will quietly include a bridge toll and forty minutes you did not intend to sell.

Ten minutes a week in the search terms report

The search terms report is the only place you see what people actually typed, as opposed to the keywords you bought. Broad match will find you queries you would never have chosen, some of them good and many of them not.

For trades, the reliable waste patterns are jobs and salary queries, do-it-yourself and how-to searches, parts and manual lookups, and people searching for free work. Build one shared negative keyword list, apply it to every campaign, and add to it weekly. Ten minutes a week here is worth more than any bid adjustment you will make this year.

Conversions that count the wrong thing

If your only conversion action fires when someone loads a thank-you page, and nobody has checked whether phone calls are counted at all, then Smart Bidding is optimizing toward form fills from browsers. It will get very good at that.

Set up call conversions with a minimum call length so a two-second hang-up does not register as a lead. Sixty seconds is a sensible starting point for most trades. Then set the counting method to one rather than every for lead actions, so the same customer calling back four times does not look like four jobs and pull your bids upward on a phantom.

The landing page you send them to

Most small accounts send every ad to the homepage. The homepage is a menu, and someone who searched for one specific thing has to find it again after paying you nothing and costing you a click.

Send each ad group to the page about that service, with the same words the ad used. If the ad says water heater replacement, the page heading should say water heater replacement, and the phone number should be tappable near the top. This is the single change that most often moves a struggling account, and it costs nothing in media.

Ads running when nobody answers the phone

A missed call is the most expensive click in the account, because you paid full price and got nothing at all. Compare your ad schedule against the hours a person actually picks up. If calls after six go to a voicemail box that is full, either staff the phone or stop buying those clicks.

Weekends are where this shows up most. Emergency searches on a Saturday convert well for the companies that answer them and are pure waste for the companies that do not. Check your call recordings or your forwarding number before you decide which one you are.

Testing comes after the leaks are closed

Split testing headlines is the advice everyone gives and it is the wrong first move on a small account. At twenty clicks a day, a headline test takes months to say anything you can trust, and in the meantime the Display Network is still spending your money.

Fix tracking, then targeting, then the landing page, then bidding. That order is boring and it is the order that works. And an honest limit: on a few hundred dollars a month in a competitive trade, you buy data slowly. Expect two to three months before the account has enough history to tell you anything, and be suspicious of anyone who promises a verdict sooner. If you would rather not learn the interface, that is a reasonable thing to hand off, but ask whoever runs it to show you these four settings first.