Listen to the podcast episode
Or listen now on Apple Podcasts or Spotify.
Key Takeaways
Custom events tell you what bounce rate never could: Quote displayed is the star example: 100 visitors from Facebook and 100 from Google Ads look identical until you see one source generating two to five times more quotes. When you don’t have enough bookings to judge a channel, engagement events — searches, quote displays, photo gallery clicks — show you where real intent lives.
Ignore the out-of-the-box events: The default GA4 list — page_view, first_visit, scroll, a generic engagement event — tells you almost nothing about a vacation rental website. Enhanced measurement adds a form submit, which lands somewhere between bad and mediocre. Events worth reporting on require Google Tag Manager work and probably your web developer’s time. It costs something up front, but the data on the other side is worth it.
Stop treating one-page visits as bad visits: Someone texts a friend a condo link; the friend opens the property page, browses photos, enters dates, gets a quote, and leaves. Universal Analytics called that a bounce and scored it a failure. GA4’s engagement model finally credits that behavior — one more reason to measure the actions that signal booking intent instead of counting pageviews.
Don’t pull month-end numbers on day one: GA4 processing runs 24 to 48 hours behind, and conversions keep trickling in for days after that. Conrad has watched last month’s ROAS read one number on the 1st and 8 to 9 percent higher by the 5th. If a KPI threshold decides whether a campaign pauses, that gap changes the decision — build settling time into any month-end reporting.
Multi-touch attribution is gone, so find a new way to tell that story: The multi-touch funnel report was the best tool for showing a client that 10, 15, even 25 touchpoints happen before a booking — and GA4 has no equivalent, even in custom reports. Data-driven attribution is now a black box: you can’t see whether first or last touch gets the weight. The buyer journey didn’t change; your evidence for it did.
Never assume a free tool will keep your history: Google deleted Universal Analytics data right on schedule — no export button, no archive option, no five-bucks-a-month way to keep it. Clients lost a decade of traffic trends they used for seasonal comparisons, and one asked the obvious question: why isn’t there a button for this? When you don’t pay for the product, your historical data isn’t guaranteed — export what matters while you still can.
Customize e-commerce tracking until it mirrors your PMS: With extra configuration you can push taxes, fees, and discounts into GA4 at the line-item level — one client’s tracking now lands within 1 to 2 percent of the PMS. You can see promo code redemptions, and 57 bookings means 57 cleaning fees you can sum against rent. The UI still assumes physical products, shipping field and all, so expect square-peg-round-hole moments where items sold really means nights booked.
Fix the Facebook channel grouping in every account: Channel groupings are genuinely more accurate in GA4, with far less traffic dumped into unassigned than UA’s worst days of 10 to 20 percent. The recurring miss: m.facebook.com counts as referral instead of social, and you have to correct it account by account. Make that fix and channel groupings become the cleanest way to walk a client through where traffic comes from.
GA4 earns a C-plus, and you should learn it anyway: Paul graded it B-minus, Conrad C-plus — a 79, in his words, with a little more effort getting it to a B. The entire point of the migration was to never do it again, and a year in, Google has been adding to GA4 rather than resetting it. It’s the default analytics platform and it’s built to last, so configuration work you invest now should pay off for years.
What We Cover In This Episode
One year after the Universal Analytics sunset, Conrad and Paul grade GA4 for vacation rental businesses: the custom events and e-commerce data worth building, the reports that vanished forever, and the quirks you have to work around.