Measuring Google Ads when the booking happens at the venue

The most common broken thing I find in a venue’s ad account is not the targeting or the creative. It is that the account has been told the wrong thing is valuable, and it has spent months getting better at producing it.

Why does the standard setup mislead venue businesses?

A typical account counts a form submission or a phone call as the conversion. The bidding algorithm then does exactly what it was asked: it finds the cheapest people who will submit forms.

Those two populations are not the same. In accounts I have audited, the gap between a form fill and a paid booking routinely runs 4:1 or worse, and the ratio is not stable across campaigns. One campaign produces enquiries from people who are comparing three venues for a corporate offsite in March. Another produces enquiries from people who will not travel more than 6km and will book this Saturday. The account reports them identically.

Over a quarter, the bidding drifts toward whichever audience generates enquiries most cheaply. That is usually the worse audience, because unserious enquiries cost less to produce. The report looks like improving efficiency. The venue looks like it is getting quieter.

How do you close the loop?

Three mechanical steps, none of which require expensive software.

Capture the click identifier. When someone arrives from an ad, Google appends a GCLID to the URL. Capture it, store it in a hidden field on your enquiry form, and carry it into whatever system holds the booking. If you use a booking platform with a custom-field option, that field is where it lives. This one step is what makes everything else possible, and it is usually the step that has been skipped.

Decide what each outcome is worth. Not every booking is equal, and averaging them away loses the signal. A birthday party worth ₹18,000 and a two-ticket walk-in worth ₹900 should not both be counted as “1 conversion”. Import values, not counts. If your margins differ sharply by package, import margin rather than revenue.

Upload on a schedule. Weekly is sufficient for most venues. Export bookings with their GCLID, date and value, and upload them to Google Ads as offline conversions. From then on, the bidding model is optimising toward the thing you actually sell.

What breaks when you do this?

Three things, and it is better to expect them.

Your cost per conversion will look far worse overnight, because you have replaced a cheap proxy metric with an expensive real one. Tell whoever reads the report before you make the change, not after. I have watched a perfectly correct migration get reversed in week two because nobody warned the owner.

Your conversion volume drops, sometimes below the threshold where smart bidding is comfortable. The fix is usually to keep a mid-funnel action as the bidding target while reporting on the real one, rather than bidding directly on a conversion that happens 15 times a month.

And your reporting goes out of sync with reality by the length of your booking window. A click on the 3rd that becomes a visit on the 28th does not appear in your numbers until the end of the month. For school trips and corporate groups this lag is long enough that weekly reporting becomes actively misleading, which is why the weekly report format we use for venues separates committed revenue from in-window revenue.

What does a working setup look like in practice?

A venue doing this properly has four things in place, and none of them are exotic.

The booking system stores a GCLID field against every record, populated from a hidden form field. A weekly export pulls bookings from the last 14 days with their click identifier, booking date and value. That file goes into Google Ads as an offline conversion import, either manually or through a scheduled Google Sheets link, which takes about ten minutes to set up once.

The ad account then runs two conversion actions: a mid-funnel one used for bidding, and the imported booking used for reporting and for value-based bidding once volume supports it. The monthly review compares blended spend against total venue revenue, so the gap between what the account claims and what the till says stays visible.

That is the whole system. The hard part is not the configuration. It is holding the line in month two, when the reported cost per conversion has tripled and someone senior asks why the ads stopped working.

What should you look at instead of cost per lead?

Four numbers, reviewed monthly rather than daily.

Cost per confirmed booking, with value attached, is the headline. Booking-to-enquiry rate by campaign tells you which campaigns produce serious people, and it is often the fastest way to spot a campaign that should be paused regardless of its cost per lead. Average booking value by campaign exposes the cases where an expensive campaign is quietly the profitable one. And blended media spend against total venue revenue, at the month level, is the sanity check that catches everything the click-level data misses, including the walk-ins that paid media clearly influenced but can never prove.

None of this requires a new platform. It requires that a click identifier survives the journey from the ad to the booking record, and that someone uploads a file once a week. The reason it is rare is not that it is difficult. It is that the broken version produces a nicer-looking report, and nobody wants to be the person who makes the number go up.

If you are starting from scratch on an account, do this before you touch creative or targeting. Optimising against a wrong objective just gets you to the wrong place faster, and every week spent doing it is a week of the model learning something you will have to unlearn. The thinking behind which campaigns to run at all sits in paid media, but it only pays off once the account knows what a good outcome looks like.