Hotel teams use channel reports to guide marketing priorities. Those reports are useful, but they are not complete records of every interaction. Attribution rules, missing source information, device changes, and technical handoffs can leave parts of the journey unresolved. The right response is not to invent the missing answer, but to understand what the available evidence can and cannot support.

Last click hands the credit to whoever touched the guest last

Analytics has to answer one question with an imperfect method: which channel gets credit for a booking? Google's attribution documentation describes attribution as assigning credit for a conversion across the ads and clicks along a user's path, and it is plain about how its last click models work: they ignore direct traffic and hand the full value of the booking to the last channel the guest clicked through before converting.

Consider what that does to a hotel report. Picture a guest who discovers your property through your website in March, checks reviews in April, and books in May after clicking a metasearch listing. Last click credits the listing. Every earlier touch your website supplied is invisible, and the report reads as if the website did little. That is not a harmless quirk. It argues, month after month, for cutting the exact work that starts guest journeys, and for feeding budget to whoever happens to stand nearest the checkout.

The direct bucket often means the source is unknown

The "(direct) / (none)" row on your report looks precise. It is a junk drawer. Google's own documentation lists what lands there: links without campaign tags, redirects that strip those tags away, shortened URLs, links opened from PDFs and other documents, and visits where ad blockers interfered with the cookies used to identify the source. Few of those guests typed your address into a browser out of nowhere. Something sent them. The dashboard just cannot say what.

For an owner the cost is specific: you cannot see which effort produced those room nights, so you cannot defend the spend behind it, and work that is quietly producing bookings looks optional at budget time.

The most expensive number on your dashboard is the one that quietly undercounts the channel you own.

AI answers send guests who arrive unannounced

Guests are arriving from AI assistants, and those visits are unusually hard to count. Semrush notes that many AI clicks show up as direct traffic in GA4 because "AI platforms don't always pass referrer information." Ahrefs describes a second layer: assistants often mention a property without any clickable link, so a guest who was effectively referred by an AI answer arrives later through a search or a typed address, and the report files the visit under search or direct.

An analytics dashboard usually cannot show whether an untracked visit followed an AI answer, another untagged source, or a later search. The absence of a visible referral does not prove either AI visibility or its absence.

One trip, three devices, and a booking engine on another domain

A guest researches on a phone during a commute, shortlists on a tablet, and books on a laptop. Analytics can only stitch those visits into one journey when it has a reliable way to recognize the same person. Google's documentation on reporting identity calls User-ID the most accurate identity space precisely because it depends on data a site collects about signed-in users. Most hotel websites have no sign-in, so Analytics falls back on a device identifier, and one guest's trip is liable to read as several unrelated visitors, with the booking credited only to the last device's path.

Then comes the handoff that hurts hotels most. Many booking engines run on a different domain from the hotel website. Google's cross-domain measurement documentation is explicit: without cross-domain setup, a single user moving between two domains is identified as two users and two sessions, because each domain sets its own cookies, and self-referrals can appear. In hotel terms, the click from your website into your booking engine can start a fresh session, and the campaign or page that actually began the trip loses the booking. The purchase still happens. The credit evaporates in the handoff.

How HTB supports clearer measurement

HTB does not sell a way around any of this, because there is none. What we do is make the numbers honest and keep them that way.

We identify how your website, booking engine, and analytics are wired together: whether cross-domain measurement is in place, whether campaign links are tagged consistently, and how large the direct and unassigned buckets have grown. We monitor those buckets over time, because a rising unattributed share is an early sign that measurement is drifting. We report what the tracking can genuinely see, in owner terms: room nights, booking value through standard Google ecommerce tracking, and the sources the data can actually support. We recommend the fixes, some of which sit with your booking engine provider to implement, and where the work is website content, we do it directly.

Just as important is what we refuse to do. HTB has no access to your accounting, your banking, or your booking engine's internals beyond standard Google ecommerce tracking, and we say so. Where a number is an estimate, we label it an estimate. Where the tracking is blind, we tell you it is blind rather than filling the gap with a confident guess. You price real decisions on these numbers: rate positioning against your comp set, marketing spend, how much commission to keep paying. The least we owe you is a report that admits what it cannot see.