Your conversion tracking broke. Here's how I'd restart your Google Ads.

It happens more often than you’d think.

The tracking breaks over the summer. Nobody notices for a while, because the ads are still running and the clicks are still coming in. Then the new website goes live.

So now you’ve got five campaigns, a couple of months with no reliable conversion data, and a site that looks nothing like the one Google learned from. Every campaign is still spending. Not one of them knows what a customer looks like any more.

Here’s how I’d fix it.

Fix the tracking first. Then test it.

Before anything else. If the conversions aren’t reaching Google, nothing else on this page matters.

Get it fixed, then get it tested. Fill the form in yourself and check it actually comes through.

Google only counts what you tell it to count.

When a campaign gets set up, Google goes looking for people who behave a certain way on your site. Change the site and that behaviour changes with it. What it learned was about a website that doesn’t exist any more.

Which means the conversion goals need going through again. Campaign by campaign.

Plenty of accounts have 11 or 12 conversion actions bundled in together, and on the old site nobody minded. There were only really two things worth doing: ring you, or fill in the form.

A new site has more buttons. Live chat. Book a call. A download. Leave the goals as one big generic lump and Google will go hunting for whichever one’s easiest to get.

So for every campaign, the question’s the same. What does a positive interaction actually look like for this business?

You can weight them, too. A completed form is worth one amount. Someone starting a live chat is worth less, but still more than someone doing nothing.

“Can’t we just count homepage visits as conversions?”

It’s a terrible idea.

Google only recognises a conversion if you tell it that’s what it is. Tell it a homepage visit counts, and it’ll go and find everyone who’ll click an ad and end up on your homepage.

You’ll have 10,000 conversions that mean bugger all.

Five campaigns on five conversions each is nobody learning anything.

Ballpark, on one campaign:

Google doesn’t set a minimum before you can switch a campaign to Maximise Conversions. Its own help page on Target CPA says to judge performance over at least 30 conversions. In my experience it’ll start making sensible decisions on about 10 a month.

Five is nowhere near either. And Maximise Conversions is built to spend your whole daily budget, whether it’s got anything to learn from or not.

So five campaigns all running at once, all starting from nothing, all on five conversions each? None of them get there.

Put four of them on ice.

To go forwards, you need to go backwards a bit.

1. Pick one campaign. The obvious pick is whichever gets the cheapest cost per conversion, even if those aren’t your most valuable leads. It doesn’t have to be. Pick the one that matters most to you.
2. Put it on Maximise Clicks, be specific on the search terms, and give it the budget. Even £100 a day.
3. Run it until it’s had 10 to 15 real conversions in 30 days.
4. Switch it to Maximise Conversions. Now it knows what a converter looks like, it might only need £40 or £50 a day to keep enough conversions coming through. So pull the budget back down.
5. Move that money to the next campaign. Same again.

Do that five times and you end up with five campaigns on £25 to £30 a day, each of which knows exactly who it’s chasing.

“But I don’t want to lose any leads.”

Fair enough. Neither do I. I want you to have as many leads as possible.

But hit the accelerator on all five at once and you’re in danger of not getting anywhere fast enough. You’re better off getting one working really well first.

The other option is to spend enough on every campaign for all five to learn at the same time. That’s roughly five times your ad budget, in what’s basically a testing phase. I wouldn’t advise it.

Then it gets interesting.

Once the campaigns are working, a form fill is only the start.

You can feed what happened next back into Google. This one was a great lead. This one was a piece of shit. That’s offline conversion tracking, and now it’s learning from the enquiries that turned into customers, not just from anyone who filled in a box.

None of that can happen until the bit at the top of this page is fixed. Conversion goals that aren’t set up properly are one of the five problems I find in nearly every account.

How long it takes.

Start to finish, four to six weeks to get the whole thing set up and working properly.

The tracking is what holds everything else up. Once that’s sorted, the campaigns can get on with it.