Every agency has had this conversation. The business owner looks at the month’s numbers, then at the invoice, and says it plainly: this isn’t working.
Sometimes they are right. Plenty of marketing doesn’t work, and nobody should keep paying for it.
But in our experience, a big share of “it isn’t working” turns out to be “we can’t see that it’s working”. Or, just as often, the reverse: something looked like it was working, and it never was. Either way the decision gets made on bad numbers, and that is the expensive part.
This is what marketing attribution is for. Not the textbook version with seven competing models, but the practical question every owner actually cares about: which of my jobs came from which of my spending?
Below are the five gaps we see most often in Australian service businesses, every one of them from the last few weeks of real account reviews, with the names taken out.
Strip away the jargon and attribution is just bookkeeping for leads. Every enquiry gets a source, every job gets traced back to an enquiry, and at the end of the month you can say “Google Ads produced these jobs, worth this much, and cost this much”.
When that chain is intact, decisions are easy. When a link is missing, people fill the gap with gut feel, and gut feel is heavily weighted toward whatever happened last week.
The chain usually breaks in one of five places.
This is the simplest gap and the most damaging, because it cuts both ways.
If nobody asks, a referral from an old client who happened to Google your business name gets logged as an “online lead”. Your marketing looks better than it is. Do that for a few months and the business ends up making budget decisions on a channel that was quietly getting credit for work it never did.
It also works in reverse. A customer who saw three of your ads, read two blog posts and then rang the number on your van gets logged as “word of mouth”. Your marketing looks worse than it is.
The fix: make the source question a required field. On the web form, in the call script, in the CRM. Not optional, not “if you remember”. Then compare what people say against what your analytics says. Where the two disagree is where the interesting answers are.
A lot of trade and service businesses use job management software for quoting and bookings, and a lot of that software puts its enquiry form on your website inside an embedded frame. It looks like part of your site. To your analytics, it is invisible.
We reviewed a roofing account recently where phone clicks and email clicks were tracked properly, and the quote form was not, because it lived inside an embed from the job management system. On the numbers we could verify, form enquiries were a large slice of the month’s total. None of them were being credited to the ads that produced them.
That means the ad platforms were optimising on part of the picture, and the monthly report was under-counting the channel that was doing a lot of the work.
The fix: wherever possible, use a form you control and pass the data into your job system afterwards, rather than embedding theirs. If that isn’t possible, at minimum track the thank-you page or confirmation step. If the software offers no integration at all, that is worth knowing before you sign up, not after.
One recent example: a booking dashboard showed one figure for the month, and Google Analytics showed more than double. Both were “right” by their own definitions. Neither was useful until someone worked out why they differed.
This happens constantly. Analytics counts a key event when a page loads. The booking system counts a confirmed appointment. The ad platform counts a conversion within its own attribution window. A new staff member adds a second tracking tag and everything starts counting twice.
The fix: decide which system is the source of truth for each number, write it down, and keep a simple change log. Any time someone touches tracking, a form, a tag or a plugin, it gets a line with the date. When the numbers jump, the log usually explains why within a minute.
Lead forms attract junk. Bots, link spammers, people selling SEO services, offshore “partnership” pitches. On some accounts it is a trickle. On others, one bad day can drop dozens of fake enquiries into a month.
If those stay in the lead count, two bad things happen. This month looks better than it was, and next month looks like a collapse when compared against it. We’ve seen a single spam spike make a genuinely improving month look like a decline.
Social lead forms are especially prone to this. One account recently logged well over a hundred enquiries from a paid social campaign, and the majority were junk.
The fix: clean the lead sheet every month before anyone reports on it. Zero out obvious spam, keep a note of what was removed, and add spam protection to the forms that need it. A smaller, honest number beats a big, wrong one every time.
This is the one that catches everyone. Clicks look normal. Click-through rate looks normal. Cost per click is fine. Leads are zero.
In one recent case, the main call to action had been removed at the client’s request while new content was being prepared. Traffic kept arriving and had nowhere obvious to go. In another, a form looked fine but was failing on a blank message field, so enquiries were getting stuck as partial submissions and never reaching anyone.
Neither problem showed up in the traffic numbers. Both showed up as “marketing has stopped working”.
The fix: test every form yourself after any website change, and once a month regardless. Use an email alias (yourname+test@yourdomain) so repeat tests don’t get flagged as spam. And never remove a primary call to action without putting a replacement in the same spot, even a plain “Contact us”.
Once those five gaps are closed, marketing ROI is straightforward:
(Revenue from jobs that came from marketing, minus the full cost of that marketing) divided by the full cost of that marketing.
Two things make or break that sum.
Use revenue, not leads. A hundred enquiries that close at a low rate can be worth less than twenty that close at a high one. If your CRM doesn’t link a job back to its original enquiry, that is the first thing to fix, ahead of any campaign change.
Be honest about the full cost. Ad spend, agency fees, and the staff time spent handling and following up enquiries. Then do the same for the revenue side: only jobs that genuinely came from the channel, not repeat clients and referrals who would have called anyway.
Done properly, this is the most useful number in the business. Done with the five gaps still open, it is a confident-looking figure built on guesswork.
Run this at the start of each month, before the report goes anywhere:
It shows you what is happening. It doesn’t change it.
If the offer is weak, the phone rings out at lunch, or quotes take a week to go out, accurate tracking will simply show you that clearly. Which is still worth knowing, because then you are fixing the right thing instead of cutting the channel that was feeding you.
What is marketing attribution in simple terms? It is working out which marketing activity produced each enquiry and each job, so you can compare what you spent on a channel with what it brought in.
Which attribution model should a small business use? For most small and mid-sized service businesses, the model matters far less than whether the conversions are being captured at all. Get forms, calls and the source question recorded properly first. Model choice is a fine-tuning question once the data is complete.
Do I need call tracking? If a meaningful share of your enquiries arrive by phone, yes. Without it, every call looks like it came from nowhere, and phone-heavy channels like Google Ads will look weaker than they are.
How do I measure marketing ROI if I can’t track every lead? Start with the leads you can track and the source question for the rest. A partially complete picture, labelled honestly, is far more useful than a complete-looking one built on assumptions.
How often should I review attribution? Monthly for the checks above, and immediately after any website, form or tracking change.
If you want a second set of eyes on whether your tracking is telling the truth, that is part of how we run Google Ads management, Brisbane SEO and small business SEO. If paid search is a big part of your spend, our guide to negative keyword lists covers the other half of the problem: the clicks you should never be paying for in the first place. Tradies specifically might want how Brisbane tradies get more jobs online.
