Your booking software decides what your ads can measure

When a clinic chooses its practice management software, the conversation is about clinical workflow. Does it do notes well, does it handle invoicing, does the front desk find it easy, does it talk to Medicare. Nobody in that room is thinking about Google Tag Manager or conversion tracking, because at the time there are no ads to track.

Then a year or two later the same clinic starts spending on Google Ads and quietly hits a wall. The reports look strange. The numbers don't reconcile with how many new patients actually walked in. The owner assumes the ads are broken, or the agency is hopeless. Often neither is true. The software is simply keeping the completed booking signal somewhere awkward, and nobody wired it up properly.

I was guilty of underrating this too. Early on I'd build a lovely account and then discover the booking system kept the completion signal in a place that needed real work to read, no matter how good the campaign was. So now it's almost the first thing I check. Which brings us to the uncomfortable part: not all software is equal, and the difference is not whether you can track a booking, it's how much custom work it takes and how clean the attribution stays.

I'll say the quiet part out loud, because a clinic owner deserves to know where their exact setup sits. And the honest version is better news than the scary version. Every major Australian practice management system now has a documented path to track a completed booking. None of them connect natively to Google Ads. Google Tag Manager, usually bridged through GA4, is the connector in nearly every case. What separates them is where the completion signal lives and how much work it takes to read it.

Cliniko and Nookal are the cleanest, as long as you use the hosted booking page rather than an embedded iframe. Splose is clean too, tracked off a confirmation signal that appears on its single page booking flow. Zanda, the system some people still call Power Diary and a few mishear as Xandr, is fully trackable through its own native GA4 event, just not via a tag you inject yourself. Halaxy is the genuinely hard one, where you can load your container but have to build the completion trigger by hand.

Here's the part that surprises people. None of this changes how good your clinic is, how many people book, or even your real conversion rate. It only changes how easily you can see it. A clinic on the cleanest system and a clinic on the hardest one can be performing identically while one needs five minutes of setup and the other needs a developer. The infographic below maps where each system keeps its completion signal. Once you've seen your own software on it, the next question is obvious: why is the booking page a clean signal in one system and a custom build in another?

The whole game is one question: when a patient finishes booking, does the software fire something a tracking code can read, and where does it fire it. That single difference is what separates a five minute setup from a developer job.

Cliniko's hosted booking page is the gold standard. It has a real confirmation page whose URL marks the booking as done, so a simple page view trigger in Google Tag Manager fires the conversion. Nookal is just as clean from its hosted page: it pushes a purchase event to the dataLayer, fires natively to GA4 and the Facebook Pixel, and can even redirect the patient to a page on your own domain after they book. Splose has no confirmation URL because the whole flow lives on one page, but it exposes a booking confirmed signal you can catch with an element visibility trigger.

Zanda works differently again. Its Client Portal natively fires a booking completed event to a GA4 property you connect, you just enter a GA4 Measurement ID in its settings. You mark that event as a key event in GA4, link GA4 to Google Ads, and import it as a conversion. There's no container you inject and no confirmation URL, so it routes through GA4 rather than a tag of your own, but the completed booking is fully and natively trackable. Anyone who told you Zanda can't be tracked was wrong.

Halaxy is the one that earns its reputation. You can paste a Google tag or container ID into its tracking settings, so its hosted booking pages load your container. But there's no confirmation URL, no dataLayer event, nothing native that fires when a booking finishes. To count a finished booking you have to build a custom trigger to detect the completion step inside Halaxy's hosted single page app, and then solve cross domain attribution, because the booking completes on halaxy.com while the click ID sits on your own site. Doable for a developer, not out of the box.

When nobody wires the completion signal up properly, tracking does not go quiet. It does something worse. It counts the wrong thing and reports it with total confidence. The booking page load gets counted. The book now button click gets counted. A returning patient hitting the ad twice gets counted. None of those is a finished booking, but the numbers look healthy, so the owner spends against them.

This is why the where matters so much. On Cliniko hosted the completion signal is a specific confirmation URL, so a page view trigger can only fire when someone genuinely finished. On Halaxy, with no native completion event, it is tempting to fall back on counting page loads, and that is exactly how an account ends up reporting nonsense. The fix is never to count the easy thing. It is to read the real completion signal wherever the software keeps it.

Which raises the cost of getting it wrong, because if you're counting the wrong event, you can't manage the money at all. You're optimising toward a number that was never real.

There's a tidy rule that floats around: embed the booking on your own site and you track everything, link out to the portal and you lose a slice. It sounds clever and it's wrong as a blanket rule. The honest answer depends entirely on the system, because what matters is whether the completion signal is readable, not whether the widget sits on your page.

Embedding works for Cliniko and Splose, because the completion signal is readable right there on the page, a message to the parent page for Cliniko, a confirmed element for Splose. For Nookal it inverts. Nookal advises against iframe embedding, because recent third party cookie and storage restrictions break embedded tracking, so the hosted portal is the clean path and the embedded widget is the unreliable one.

For Halaxy and Zanda the question is moot. Whether you embed the iframe or link out, the booking runs on the vendor's own page either way, so your site's own tag can't see inside it. The only thing that counts is the tag or Measurement ID you set inside the vendor. So the right move is not a universal rule. It's reading what each system actually exposes and pointing the conversion at that.

Even on the cleanest software, one chunk of demand stays invisible: the phone. Nobody completes a phone booking inside a tracking pixel. So I report a proxy. I count how many phone calls lasted longer than 90 seconds, because a call that short is almost never a booking, and a call past 90 seconds has a real chance of being a new patient.

It's not perfect and I don't pretend it is. The owner can apply a sensible assumption, say half of those longer calls booked, and now even the offline slice has an honest estimate attached. That's the difference between a number you trust and a number you wish were true.

Stack it together and a picture forms. Your software sets where the online signal lives, the right trigger reads it cleanly, the 90 second proxy backfills the phone, and what's left is a measurement system you can actually steer by. The only thing missing is the question every owner asks once they understand all this: do I need to switch software?

Here's the reassuring part. For nearly every clinic the answer is no. Every major system on the market can track a completed booking now. Cliniko, Nookal, Splose and Zanda all have a clean or supported path, and the job is to wire it up properly, not to rip out a clinical system your team loves. Even Halaxy is trackable, it just needs a developer to build the completion trigger and sort out cross domain attribution.

The real watch out isn't the brand on your invoice, it's leaving the tracking on defaults or pointed at the wrong event. That's what produces a 148% conversion rate and months of spending against a number that was never real. The work is to read the completion signal where your specific software keeps it. On Cliniko that's a confirmation URL. On Nookal it's a dataLayer event from the hosted portal. On Zanda it's a GA4 event you connect and import. On Halaxy it's a custom build. Different craft for each, same goal.

And I'll be honest about who does that craft. The fiddly cases aren't something a clinic owner should attempt themselves, and I outsource the trickier code to a developer myself, because when I make a mistake in a custom trigger I can't always find it without starting from scratch. That's not a confession of incompetence. It's the reason a clinic shouldn't have to learn Google Tag Manager to know whether their ads are working. You should just get the truth, every month, in numbers you can trust.