Most tour operators can tell you how many bookings they took last month, because the booking system says so. Far fewer can say which of those bookings came from Google Ads, which from a blog post and which from an Instagram story. The reason is usually technical and fixable: the booking happens inside a widget or on a separate domain, and Google Analytics 4 (GA4) never sees the purchase, or sees it as a visit from your own website.
This guide shows how to set up GA4 conversion tracking so that bookings made through a booking widget are credited to the right channel. It applies to FareHarbor, Bokun, Rezdy, Checkfront, TourCMS and similar systems, though the exact menus differ, so check each platform’s current documentation before you change anything.
Why bookings go missing in GA4
GA4 follows a visitor by a cookie set on your domain. When the visitor clicks “Book now” and lands on a checkout hosted on another domain, or inside an embedded frame from another domain, the cookie does not travel with them unless you tell it to. Three things then tend to happen:
- The purchase is never recorded in GA4, so your conversion count is zero or far too low.
- The purchase is recorded, but as a new session that began on the booking platform’s domain, so the source shows as “direct” or as your own website.
- The payment provider sends the guest back to your site after payment, and GA4 treats that return as a new visit referred by the payment provider.
In all three cases the channel report quietly understates paid search and blog traffic and overstates “direct”. That leads to the wrong budget decisions, which is why we treat tracking as the first job before any campaign work.
Decide what counts as a conversion
Do not mark everything as a conversion. A short, clear list works best. GA4 has a set of recommended events with standard names, and using them means the reports and Google Ads imports understand the data. Google lists them in its GA4 recommended events documentation, including view_item, begin_checkout, purchase and generate_lead.
For a tour operator we would normally use this ladder:
- view_item when someone opens a tour page. This is a signal, not a conversion.
- begin_checkout when someone opens the booking widget and chooses a date. Useful for audiences and for finding where guests drop out.
- purchase on the confirmation step, with value and currency. This is the main conversion.
- generate_lead for enquiries, group requests and “ask about this tour” forms. This is the second conversion for operators who sell private or custom trips.
Mark only purchase and generate_lead as key events (the name GA4 now uses for conversions). Keep the others as ordinary events for analysis.
Three ways to get the booking out of the widget
There is no single method that works for every platform. In practice there are three, and the right one depends on what your booking system offers.
1. The platform’s own GA4 or tag integration
Many booking systems let you paste a GA4 measurement ID or a Google Tag Manager container ID into their settings, and some then send ecommerce events for you. This is the easiest route when it exists. Read the platform’s help pages to see which events it sends and whether it handles cross-domain journeys, because we have seen setups that record the purchase but lose the source.
2. Cross-domain measurement
If the checkout sits on a domain you can add to your tag, GA4 can be told to treat both domains as one site. In the data stream’s tag settings there is an option to configure your domains. When you list your own domain and the booking domain, GA4 passes the visitor’s identifier in the link, so the session carries across. Menu names change from time to time, so search Google’s help for “cross-domain measurement” if the screen looks different.
This only works if your tag also runs on the booking domain, or if the link between the two sites is a plain link that GA4 can decorate. A widget loaded in a frame needs extra care, and some platforms do not allow your own tag inside their checkout at all.
3. A confirmation step you control
If neither of the above is possible, the fallback is a confirmation or thank-you page on your own domain. Many booking systems can redirect the guest to a custom URL after payment. You then fire the purchase event on that page, passing the value from the URL or a data layer. It is less precise, because a guest who closes the tab before the redirect is not counted, but it keeps the source intact and is simple to maintain.
The test of a good setup is not that GA4 shows bookings. It is that the bookings in GA4 match the booking system within a believable margin, and that they carry the right source.
Keep the source when the guest comes back from payment
If your checkout sends guests to a payment provider and then back to you, add the payment provider’s domain, and the booking platform’s domain, to the unwanted referrals list in the data stream settings. That stops GA4 starting a new session when the guest returns and overwriting the original source.
Also tag your own campaigns consistently. Use UTM parameters on every paid and email link so that the source and campaign survive even if something else goes wrong. A short, fixed naming system is worth agreeing before you launch campaigns, not after.
Test with a real booking
Do not trust the setup screen. Test it the way a guest would use it.
- Open your site in a private browser window and arrive through a tagged link, for example one with a test campaign name.
- Open a tour page, start the booking widget and pick a date.
- Complete a booking using a low-priced tour or a test mode if your platform has one. If it does not, make a real booking and refund it afterwards.
- Watch GA4’s DebugView (Admin, then DebugView) while you do it. You should see view_item, begin_checkout and purchase arrive in order.
- Next day, check the standard reports. The booking should appear against your test campaign, not under “direct”.
If the purchase appears in DebugView but not in the reports, give it up to a day, as some reports process later. If the source is wrong, the cross-domain or referral settings are the usual cause.
Reconcile with the booking system every month
GA4 and your booking system will never match exactly. Guests use ad blockers, decline cookies, switch devices and cancel. A gap is normal. What matters is that the gap is stable and not huge.
Take one month and compare the number of bookings and the revenue in both systems. As an illustration only, if the booking system shows 200 bookings and GA4 shows 150, you capture three quarters, and you can read the channel split in GA4 as a reasonable sample. If GA4 shows 40, something is broken. Write the ratio down and check it again each month. A sudden change usually means a widget update or a tag that stopped firing.
Consent also affects the numbers. If you run a cookie banner, which is normal for visitors from the EU and UK, GA4 will not see everyone. We are not lawyers, so take legal advice on consent for your markets, but plan for the gap when you read the reports.
Use the data in your ads and reporting
Once purchase is tracked with value, you can link GA4 to Google Ads and import it as a conversion, so that campaigns are judged on bookings and revenue instead of clicks. You can also build audiences from begin_checkout, which feeds the retargeting work we describe in our guide to Meta ads for tour operators in the shoulder season.
The same data tells you where the booking flow leaks. If many guests start checkout but few finish, the fix is not more traffic. Our piece on mobile checkout for tour bookings covers the usual causes. And when you plan next season’s spend, the numbers feed straight into your marketing calendar for 2027.
If you would like someone to look at the whole path from advert to confirmation, we offer a website and booking-flow check, described on our websites and funnels page.
Frequently asked questions
Do I need Google Tag Manager to track bookings?
Not always. If your booking platform has a built-in GA4 field and sends purchase events, you may not. Tag Manager helps when you need a custom confirmation page event or more control, and it keeps changes out of your site’s code.
Why does GA4 show far fewer bookings than my booking system?
Common causes are the widget being on another domain, ad blockers, cookie consent and the guest not reaching the confirmation step. Test a booking end to end and compare monthly totals before assuming the tag is wrong.
Which events should I mark as key events?
Usually purchase and, if you take enquiries, generate_lead. Marking every step as a key event makes the numbers look better and the decisions worse.
Can I track bookings that happen by phone or email?
Not automatically. Ask staff to record how each guest found you, or use a call-tracking number, and add those bookings to your monthly review by hand. It is imperfect, but it stops phone bookings being credited to nobody.
How often should I check the setup?
Once a month for the reconciliation, and again whenever the booking platform, the website theme or the cookie banner changes. Those are the moments tracking tends to break.
Where to start
Start with one tour and one test booking. Find out what the platform sends, fix the source problem, and write down the monthly ratio between GA4 and the booking system. Once that is stable, every other marketing decision gets easier, because you are finally looking at bookings and not just visits.