Most GA4 content on the internet is still migration content: how to move over from Universal Analytics, what changed, why the new interface feels wrong. Universal Analytics stopped processing data in July 2023, and Google permanently deleted the historical data a year later. There is nothing left to migrate. You’re GA4-native now, whether you like it or not.
So this isn’t a migration guide. It’s the setup we run across dozens of vacation rental manager websites, and — more importantly — how we actually read the numbers once they’re flowing: the booking funnel we watch every month, the metrics we deliberately ignore, how to get revenue tracking your PMS will agree with, and a straight answer on measuring AI traffic.
We’ve covered analytics on The Heads In Beds Show for years. The freshest episode on how the booking funnel has changed is right here:
Where GA4 Stands in 2026
One year into GA4, we graded it on the podcast: a C+ and a B-. A few years on, we’d stand by that grade — and we’d still tell every vacation rental manager to use it. Both things are true.
The complaints are real. Data processing is slower than Universal Analytics ever was: reports commonly lag 24 to 48 hours, and month-end numbers keep shifting for days into the new month. The multi-touch funnel reports many of us relied on are gone without a true replacement. And the interface still throws 85 numbers at an owner who needs about ten of them.
Here’s the other side: GA4 is free, it’s good, and it’s wired directly into Google Ads bidding. It is really hard to compete with free and good. Every alternative we’ve evaluated — Fathom, Clicky, Amplitude — earns a place next to GA4 at best, never instead of it.
One vocabulary reset before we go further, because Google renamed things after most guides were written: everything in GA4 is an event, and since March 2024 the actions you mark as important are called key events. “Conversions” in GA4 now refers specifically to actions shared with Google Ads. If your saved reports and your agency’s slides use different words for the same thing, that’s why.
The Setup Checklist We Run on Every New Client
When we take over a vacation rental website, this is the pass we make on GA4 before we trust a single number in it:
- Install through Google Tag Manager. Hard-coded snippets die in redesigns; a GTM container moves with you, and it’s where your event customization will live anyway. On WordPress, a performance plugin that injects gtag.js is fine for base tracking — but the moment you want custom events, you want GTM.
- Match timezone and currency to your PMS. GA4 converts transactions using daily exchange rates; a mismatch guarantees your revenue never reconciles.
- Configure site search parameters. Date searches are the top of your booking funnel. GA4 captures q, s, search, query, and keyword by default — add whatever parameters your booking engine actually uses.
- Set up cross-domain tracking and unwanted referrals. If your checkout hops to a booking engine domain, this is the most important — and most skipped — configuration in vacation rental analytics. It gets its own section below.
- Filter internal traffic. Define IP addresses for your office and your reservation team. Your own staff should not be your best-converting audience.
- Extend data retention from 2 to 14 months. GA4 still ships with a 2-month default, and Explorations can’t analyze data you didn’t keep.
- Set reporting identity to Observed. The Blended setting fills gaps with modeled data — sessions that didn’t observably happen. On small and mid-size sites, we’d rather report less and trust it more.
- Decide on Google Signals with your eyes open. It’s how you get age, gender, and interest data, but it no longer contributes to cross-device reporting identity (that changed in early 2024), and it can trigger heavy thresholding — rows of data hidden for privacy — on smaller sites. Demographics or complete reports: pick which you value more.
- Add multiple admins, including the owner. There is no reliable recovery path when the one person who owned Analytics leaves the company. And the account should belong to you, never to your agency. When you lose access to your analytics account, you are toast.
- Link Google Ads and Search Console. Then go to Reports > Library and actually publish the Search Console collection — creating the link alone doesn’t surface the reports.
- Verify with Tag Assistant and DebugView. And watch for double-tracking: several PMS website templates auto-inject GA4, so adding your own tag on top quietly double-counts everything.
The Only Funnel That Matters: Search, View, Checkout, Purchase
Strip everything else away and vacation rental analytics is four events, in order — and they should use GA4’s default names, not homemade ones, so the built-in ecommerce reporting works and your setup stays portable:
- search — a visitor entered travel dates. They’ve told you when they want to come.
- view_item — a property detail page view. On a vacation rental site, this is a shopping event: dates in hand, looking at a specific home.
- begin_checkout — they clicked Book Now and started the reservation process.
- purchase — a completed booking, with revenue attached.
Here’s a real month from one of our clients: 27,000 sessions produced roughly 12,000 date searches, 14,000 property page views, 872 begin-checkouts, and 125 completed bookings. Every layer of that funnel is a question you can act on. Plenty of searches but few property views? Your search results page is failing. Plenty of checkouts but few purchases? Something in the booking flow — fees, forms, trust — is scaring people off.
Why “Book Now” Clicks Aren’t Bookings
On most booking engines, starting checkout is the only way to see the true total with taxes and fees. So a large share of begin-checkouts are really quote requests: someone generating the all-in number to take back to their mother, brother, cousin, husband, or wife so the group can figure out how to split it. That’s not abandonment — that’s how families book vacation homes. Read begin_checkout as serious shopping, not as failed payment.
The One Number to Watch
Across our client base, a healthy vacation rental site converts roughly 10% of begin-checkouts into purchases. The precise figure matters less than your baseline. If you normally run 10% and this month prints 5%, go find out what broke — a payment gateway, a rate rule, a new fee. If it prints 20%, go find out what worked. GA4 now has a benchmarking feature with travel peer groups if you want outside context, but our position hasn’t changed: your own trend line beats an industry average. (When you do need industry numbers, we keep a sourced library of vacation rental statistics.)
Segment the funnel by device, operating system, and traffic source. Mobile and desktop behave like different websites, and two channels sending identical session counts can differ several-fold in how many quotes they generate. One workflow note: GA4’s funnel explorations still can’t compare time periods side by side, so once a month we export the funnel numbers into a simple scorecard. Ten minutes, and now you have trend lines.
Revenue Tracking Your PMS Will Agree With
The purchase event is where most vacation rental setups fall apart, because GA4’s ecommerce model assumes you’re shipping sneakers. We’re not shipping a vacation rental. The payload that works for lodging:
- value — the gross booking amount, fees included
- tax — separated out, never lumped into value
- items — the property (ID and name), with price set to the commissionable rent and quantity set to nights
- every fee as its own line item — cleaning, damage waiver, pet fee, hot tub. It may be a $500 booking with $500 in fees, and some of those fees are pure margin. You want to see them, not bury them.
- transaction_id matching the PMS reservation ID — so every booking in GA4 reconciles one-to-one against a reservation.
Then reconcile monthly. Expect GA4 to under-report direct bookings by around 10% even when the setup is healthy — ad blockers, declined consent, cross-device stragglers. When it’s not healthy, the gap gets absurd: one account we audited took 88 bookings in a month while GA showed 54. A careful configuration can get within a percent or two. Either way, the PMS is the place of record. GA4 is the directional layer that tells you where the bookings came from.
One non-negotiable to hand your web developer: a unique confirmation-page URL after a completed booking. It’s the fastest tracking health check that exists, and it’s the difference between diagnosing “tracking broke” versus “bookings stopped” in five minutes instead of a week.
Booking Engines: Where Vacation Rental Tracking Actually Breaks
The moment a guest hops from yourdomain.com to a hosted checkout on another domain, GA4 sees them leave and a “new” visitor arrive. Session split, campaign credit gone — and the booking lands in your reports as a referral from your own booking engine. Your ads and your SEO get credit for nothing.
Two settings fix the common case, both under Admin > Data streams > Configure tag settings: Configure your domains (add your domain and the booking domain) and List unwanted referrals (add both there too, so the hop never counts as a traffic source). Then run a test booking start-to-finish yourself and watch it in DebugView.
And re-test after every PMS update. The cautionary tale we tell everyone: a PMS template shipped a cookie-consent banner on its hosted booking pages, almost nobody clicked accept, and GA, Google Ads, and Meta recorded essentially nothing for weeks — while bookings kept flowing. Marketing looked dead. The business was fine. Nobody caught it until someone compared the numbers to the PMS.
Platform Notes From Our Internal Setup Docs
- OwnerRez — the best out-of-the-box GA4 ecommerce we’ve seen: line items, fees, and taxes broken out properly. Still needs the cross-domain and referral-exclusion settings above for its secure booking domain.
- Streamline — the GA4 measurement ID lives in the website plugin’s integration settings; verify it after any template change.
- Uplisting — native connect pages for GA4, Google Ads, and Meta; paste your IDs and you’re tracking.
- Guesty — GA4, GTM, and pixels configured through the booking engine’s tracking settings.
- Hostaway — supports GTM and pixels; the purchase event gets built in Tag Manager.
- Homerunner — tracked via a GTM container import; double-check the ecommerce value variable mapping or revenue won’t pass.
Different platform? The pattern is always the same: get GTM or gtag onto the booking pages, pass the ecommerce data layer, then fix cross-domain.
Web analytics also can’t see phone bookings or emailed quotes, and for some managers that’s still a meaningful slice of revenue. Patch it with call tracking numbers, a “how did you hear about us” field with checkboxes instead of radio buttons (guests have more than one answer, and forcing one is how attribution myths start), and promo codes as fallback attribution for offline channels.
Metrics We Ignore on Purpose
- Sessions. Bottom of the scorecard, on purpose. If your metric is sessions, you’re on the wrong track — and it got worse in late 2025 when bot and AI-agent traffic surged. One site’s server logs showed 175,000 hits in a window where GA reported a few thousand sessions; nobody can tell you precisely what’s human anymore. Validate with events only humans perform: a bot’s not going to enter dates and add a credit card.
- Bounce rate and time on site. A one-page visit that runs a date search and generates a quote might be your best visitor of the day. Thirty seconds is plenty of time to read a description and check availability. Both metrics punish exactly the behavior you want.
- Default event clutter. Seventeen thousand scroll and click events on 1,300 sessions tells you nothing. The events you configured on purpose are the ones that earn a place in a report.
- Real-time reports and 7-day windows. Real-time undercounts and tricks you. And GA4 data takes 2 to 4 days to finalize, so a last-7-days view compares finished days against incomplete ones. We report on a lagged window, always.
Before You Declare a Traffic Emergency
The triage we run before anyone panics: never judge a 7-day window. Zoom out. Is the drop in every channel or just one? Is it macro — your market’s seasonal booking curve — or is it your site? Check rankings before blaming your SEO. The real alarm is organic traffic down 20% or more and staying down for weeks. Everything else is usually noise, seasonality, or a reporting lag.
Attribution Without Lying to Yourself
GA4 offers exactly three attribution models now — data-driven (the default), paid-and-organic last click, and Google-paid-channels last click. Data-driven attribution is better than nothing, but in practice it still hands most of the credit to the last click, which is usually branded search. Our standing line on brand campaigns: branded search stops up all the credit, but it didn’t do the work — the guest was 99% of the way there, and that last 1% got toppled over. Switch attribution models on the same date range and watch 30-50% of conversions move between channels. That’s not precision; treat every attribution number as directional.
The most underrated screen in GA4 is the attribution paths report (Advertising > Attribution paths). That’s where the real story lives: 47-day journeys with 11 touchpoints, an email sent in 2023 assisting a booking years later, an AI referral showing up at step two of five. Booking a vacation home is a committee decision made over weeks — the path report is the only place GA4 admits it.
It changes how you judge channels. Two hundred newsletter clicks that produce 110 date searches, 47 property views, and 2 completed checkouts in the same session is not a failing channel — that’s email moving cold subscribers deep into the funnel in one visit. Judge email and social by funnel movement, not last-click bookings, or you’ll quietly starve your best top-of-funnel channels while overfeeding branded search.
Three habits make attribution meaningfully better: manual UTM tags on everything you control (every email, every QR code on a postcard — untagged links all pile into direct); in Google Ads, keep the Google-tag conversion as primary and the GA4 import as secondary so you can compare both; and push phone and offline bookings back into the platforms via enhanced conversions from your PMS or CRM.
AI Traffic in GA4: What’s Real in 2026
As of May 2026, GA4 finally has a native “AI Assistant” channel: referrals from ChatGPT, Gemini, and Claude get classified automatically. Two gaps remain: Perplexity isn’t included, and visits from AI apps or copied links carry no referrer at all, so they land in Direct. Keep a custom channel group with a hostname rule covering chatgpt.com, chat.openai.com, perplexity.ai, claude.ai, gemini.google.com, and copilot.microsoft.com — and note that ChatGPT tags its web links with utm_source=chatgpt.com, which makes segmentation easy.
Now the honest numbers, from our own client base. When we pulled LLM referrals across the roster in early 2025, they were 429 sessions out of 734,376 — about 0.06%. One client site in 2026 logged roughly 800,000 Google referrals against 17 from ChatGPT. Referral clicks from AI assistants are still a rounding error for most vacation rental managers.
And yet Similarweb’s research keeps ranking travel and hospitality as the most-cited industry in AI answers — 22.6% of citations. Both are true, and squaring them is the whole game: AI influence rarely arrives as an “AI visit.” People research in ChatGPT, get your brand recommended, then Google your name and show up as branded search or direct. The citation click often carries nothing. So watch branded query growth in Search Console as your AI-influence proxy, and filter GSC with a regex for prompt-shaped long-tail queries — full-sentence questions are already showing up in the data.
One more AI effect that shows up in GA4 whether you track it or not: AI Overviews answer informational queries on Google itself, so blog clicks fall even at stable rankings. The article gets summarized, the reader gets their answer, and there goes your click. Impressions become the leading indicator for informational content. Getting recommended inside those AI answers is its own discipline — that’s our GEO and LLM optimization work, and it pairs with everything in this post.
What We Run Alongside GA4
- Google Search Console — just as important as Analytics for a vacation rental site. Non-brand clicks year over year is the honest SEO health metric; the count of branded queries is your brand-demand trend.
- Microsoft Clarity — free session recordings and heatmaps. Watch recordings at 2x-4x speed, look for rage clicks and dead ends on property pages. The best free complement to GA4 there is.
- Your PMS — the system of record for revenue, always. Every GA4 number is a lens on it, not a replacement for it.
- A reporting layer — nobody should hand an owner raw GA4. Seven to ten numbers on one page, month over month, with the funnel front and center.
- An AI assistant, genuinely — have Claude or ChatGPT build your GA4 custom reports and explorations, or analyze a CSV export for anomalies. It’s faster than clicking through the settings maze, and it doesn’t get bored.
Frequently Asked Questions
Which vacation rental booking platforms support Google Analytics tracking?
Nearly every major platform supports at least a GA4 measurement ID, and most now support full ecommerce tracking: OwnerRez does it best out of the box, Uplisting has native connect pages, Guesty and Hostaway support GTM and pixels, and Streamline and Homerunner are configured through plugin settings or a GTM import. The real question isn’t whether your platform supports GA4 — it’s whether cross-domain tracking and the revenue payload are configured correctly.
Why doesn’t my GA4 revenue match my PMS?
It never will match exactly. Around 10% under-reporting is normal — ad blockers, declined cookie consent, phone bookings, cross-device journeys. Gaps much bigger than that usually mean broken cross-domain tracking, a missing purchase payload, or a consent banner blocking tags on the booking domain. Set the GA4 transaction ID to your PMS reservation ID and you can identify exactly which bookings go missing.
How do I see ChatGPT traffic in Google Analytics?
GA4’s native AI Assistant channel (added May 2026) classifies ChatGPT, Gemini, and Claude referrals automatically. Add a custom channel group to catch Perplexity and others, and look for utm_source=chatgpt.com in your session data. Expect small numbers — most AI influence shows up later as branded search and direct traffic, so watch branded query growth in Search Console too.
Is GA4 still free?
Yes. The standard version is free, including the BigQuery export (up to one million events per day). The paid GA4 360 tier exists for enterprise-scale properties, and virtually no vacation rental manager needs it.
Does a vacation rental website need a cookie banner?
If you take European traffic, yes — consent plus Google Consent Mode v2, or you lose remarketing and ad measurement for those visitors. US-only sites face a patchwork instead: twenty states now have comprehensive privacy laws on an opt-out model, and as of January 2026 a dozen require honoring Global Privacy Control signals. Practically, that means a real privacy page and an opt-out mechanism rather than an EU-style consent wall. We’re marketers, not lawyers — confirm your setup with a privacy attorney.
What replaced bounce rate in GA4?
Engagement rate — bounce rate still exists as its inverse. We don’t watch either on vacation rental sites: a guest who lands, runs a date search, and generates a quote in one visit is a win that both metrics can score as a failure. Watch the funnel events instead.
The Setup Is the Easy Part
GA4 in 2026 is what it’s been since launch: free, powerful, and completely indifferent to whether you configure it correctly. The managers getting real value from it track four events with a clean revenue payload, reconcile against the PMS monthly, and ignore the other 85 numbers on the screen.
We run this exact setup — and the monthly reading of it — as part of every vacation rental marketing engagement. If your funnel doesn’t look like the numbers in this post, or GA4 and your PMS are telling two different stories, get in touch. The analytics audit is where we start.
