Measuring Hotel Ad Performance: Attribution That Reflects Direct Bookings

Why measuring hotel ads is harder than it looks

A hotel booking is rarely a single click. The same guest might read an AI answer, find your organic listing, follow you on Instagram, see your rate on Google Hotel Ads a week later, then book after a brand search. Several channels touched that booking, but your booking engine usually credits only the last one. That single habit is where most hotel measurement goes wrong.

The hotel booking window is long and indirect. Leisure guests in particular research for days or weeks, across devices, comparing you against OTAs and against other properties. By the time they book direct, the channel that first put you on their list is long out of view.

This matters because the channel that gets the credit gets the budget. If your reports systematically over-credit the last click and ignore everything that led to it, you end up defunding the very channels that fill your pipeline. Measuring hotel marketing ROI properly starts with accepting that no single number tells the whole story.

What last-click attribution gets wrong

Last-click attribution hands 100 percent of a booking's value to the final touch before the guest booked. For hotels that final touch is very often a brand search or a remarketing ad, so paid search looks brilliant and everything earlier in the journey looks like a waste. The model is not lying, it is just answering a narrower question than you think you asked.

Picture a guest who discovers you through an Instagram reel, comes back via an organic search two days later, then books after clicking a brand ad on your hotel name. Under last-click, the brand ad takes the entire booking. Social gets nothing. SEO gets nothing. If you trimmed budget on that basis, you would cut the two channels that actually introduced the guest and keep paying to close bookings that were already half-won.

There is a second trap specific to hotels. Brand search and remarketing tend to sit at the end of the journey by their nature, so last-click structurally rewards them. That is useful to know (brand defence really is valuable, as our Hotel PPC Guide explains), but it is not the same as those channels creating demand. They mostly capture and protect it.

• ›It over-credits brand search, which usually catches guests who already intended to book you.
• ›It over-credits remarketing, which by design only reaches people who came from somewhere else first.
• ›It under-credits SEO, social and email, which open journeys but rarely close them.
• ›It encourages budget decisions that slowly starve your top of funnel.

How attribution models actually work

An attribution model is simply a rule for splitting the credit for a booking across the touches that led to it. Some rules are fixed and arbitrary; one, data driven attribution, learns the split from your own data. Understanding the handful that matter is enough to read any hotel report critically and spot when a channel is being flattered or robbed.

You do not need to memorise every model. You need to know which question each one answers, and which ones quietly distort a hotel's long booking window. The practical choice for most independents comes down to two: the last-click default that ships with most tools, and data driven attribution, which we cover in its own section below.

The fixed-rule models

• ›Last click. All credit to the final touch. Simple, and biased towards closing channels.
• ›First click. All credit to the first touch. The mirror-image bias: it over-rewards discovery and ignores what actually sealed the booking.
• ›Linear. Equal credit to every touch. Fairer than the extremes, but it pretends a passing display view mattered as much as the brand search that booked.
• ›Time decay. More credit to touches nearer the booking. Sensible for hotels in principle, though it still leans towards the end of a long window.
• ›Position based. Most credit to the first and last touch, the rest shared. A reasonable compromise when data driven is not available.

Why the model you pick changes the story

Run the same quarter through last click and then through a model that shares credit, and your channel league table can reorder completely. SEO and social often look two or three times more valuable once they get credit for the journeys they started. Nothing about the bookings changed; only the lens did. This is exactly why you should never compare two reports built on different models and treat the difference as performance.

Data driven attribution, and why it suits hotels

Data driven attribution shares credit across touches based on your own conversion data rather than a fixed rule. It compares the paths that led to bookings against the paths that did not, and assigns each channel credit in proportion to its measured contribution. For a business with a long, multi-touch journey like a hotel, that is usually the fairest default available.

In Google Ads and GA4 this is now the standard model, and it is the one we lean on for most hotel accounts. Instead of guessing that the last click deserves everything, it works out, from your real data, how much each step nudged the guest towards booking. A brand ad still gets real credit, but so does the organic visit and the metasearch click that came before it.

It is not magic and it is not perfect. It needs a reasonable volume of conversions to learn from, so a very small property with a handful of monthly direct bookings will get a noisier picture than a busy one. It also only sees what you let it see, which brings us to the part most hotels skip: connecting the ad platforms to the actual booking.

Connect the ad platforms to your booking engine

No attribution model can measure a booking it never sees. The single biggest measurement gap in hotels is ad platforms that track clicks or contact-form fills but never receive the real booking and its value from the booking engine. Fix that connection first, because every other number depends on it.

Most independent hotels run a third-party booking engine that sits on a different domain from the main site. If conversion tracking is not configured to follow the guest across that hand-off, the booking happens in a place your ad platform cannot see, and the click that earned it gets recorded as a failure. You then optimise towards the wrong things and wonder why performance drifts.

Done properly, the booking engine passes the confirmed booking and its revenue back to Google Ads, GA4 and Meta, so each platform can tie that revenue to the click and the campaign that drove it. That is what turns a dashboard of clicks into a report of bookings and pounds. It is fiddly, engine-specific work, and it is the part we sort out as standard in our hotel paid search service before we judge any campaign.

• ›Track the full path from ad click to confirmed booking, including the jump to the booking engine domain.
• ›Pass real booking value back, not a flat placeholder, so revenue and cost per booking are honest.
• ›Use server-side or enhanced conversion measurement where you can, since browser tracking is increasingly blocked.
• ›Reconcile what the platforms report against your PMS or booking-engine totals, and treat large gaps as a tracking fault, not a performance one.

Assisted conversions: the channels that start bookings

An assisted conversion is a booking a channel helped along the way without delivering the final click. The assisted-conversions report is where the hard-working but invisible channels finally show up. For hotels it is often the single most useful view, because it reveals what your last-click report has been hiding.

If a channel shows few last-click bookings but a high assist count, it is doing real work earlier in the journey. Social and content frequently look like this: rarely the closer, often the opener. Cut them on last-click logic and you will watch your 'efficient' closing channels slowly weaken, because the guests they used to close stopped being introduced in the first place.

The practical move is to look at last-click and assisted side by side for every channel. A channel that both opens and closes is a keeper. A channel that only ever assists is a discovery engine you fund deliberately, not an underperformer you trim. This is the same logic behind running SEO and paid as one system rather than judging each in isolation.

Report on bookings and commission saved, not clicks

The honest scoreboard for a hotel is cost per direct booking set against the OTA commission you would otherwise have paid, plus your direct-versus-OTA mix over time. Clicks, impressions and even raw ROAS are inputs. Bookings and commission saved are the outcomes that map onto profit and onto the dependency you are trying to reduce.

ROAS (return on ad spend) earns a place, but read it knowing it flatters channels that capture demand rather than create it. The measure we trust more is simple: what did this booking cost to win, and what would the same booking have cost in OTA commission at 15 to 25 percent? If a direct booking comes in for less than the commission you have avoided, the channel is doing its job even when the ROAS headline looks ordinary.

This framing also keeps everyone honest about brand search. The value it generates is largely commission saved on guests who were already yours, not brand-new revenue. That value is real and worth paying for, but labelling it correctly stops you double-counting it as growth. We go deeper on the ROAS trap in the Hotel PPC Guide.

• ›Lead with cost per direct booking, by channel, over a full booking window.
• ›Show commission saved: direct bookings won multiplied by the OTA commission avoided.
• ›Track the direct-versus-OTA revenue mix quarter on quarter as the headline trend.
• ›Keep ROAS and clicks as diagnostics, not as the verdict.

Build a measurement view that survives a long booking window

Hotel bookings ripen slowly, so any honest report has to allow for delay and for the multi-touch reality. Judge campaigns over weeks, separate brand from non-brand, and resist the urge to react to a single soft fortnight. A measurement habit built for a long window beats a prettier dashboard built for a short one.

In practice this means a few disciplines held steadily. Look at performance over at least a full month, ideally a quarter, so the booking lag has time to resolve. Keep brand and non-brand search reported separately, because they behave nothing alike and averaging them hides the truth. And anchor decisions to the booking and commission numbers, letting clicks and impressions explain the why rather than set the direction.

When measurement is set up this way, the wins become legible. We have seen a multi-property chain add around £125,000 in direct booking revenue over 90 days once tracking and channel mix were put right, and a Tuscan spa hotel lift direct bookings by about 20 percent in its first month after a new website fed clean conversion data back to the ad platforms. Neither result is visible on a last-click click report. Both are obvious on a bookings-and-commission one.

Where measurement fits in the wider hotel marketing picture

Attribution is not a reporting afterthought; it is the steering wheel for the whole programme. Get it right and every other decision (what to fund, what to cut, where to push) rests on bookings rather than guesswork. Get it wrong and you optimise confidently towards the wrong channels.

If you are building out paid search, this is the layer that tells you whether it is working, so plan it alongside the campaign work in the Hotel PPC Guide rather than bolting it on afterwards. If you are mapping the bigger system of SEO, website, social and paid, the hotel marketing guide shows how the channels connect, and clean attribution is what lets you see those connections instead of assuming them.