Attribution for Meta and Google campaigns in hotels

Attribution assigns credit to marketing touchpoints before a booking. It does not by itself prove how many bookings would not have occurred without advertising. Hotel budget decisions require separately verified bookings, costs and an explanation of how each system attributed the result.

Why booking counts differ

Let's imagine an educational example: Google Ads shows 45 conversions, Meta 30, and a booking engine 52 confirmed online bookings. The sum of 75 is not the number of guests nor proof of error. Systems can assign the same booking to different contacts, use different periods, or count different actions.

Read the settings from a specific account. Don't assume that each account uses the same attribution window or that each platform always exaggerates the result.

How to reconcile your monthly report

1. Select the same group of bookings. - For acquisition, use the date the online booking was made. Revenue earned during the month is a separate business measure.
2. Separate confirmed, cancelled and completed stays. - The booked amount is not yet collected or realised revenue.
3. Align currency and what the value includes. - Document taxes, fees, additional services and changes to your stay.
4. Check the measurement. - Examine transaction identifiers, duplicates, cross-domain linking and consent. Note what in GA4 cannot be linked to the booking record.
5. Show channels separately. - Do not add platform attribution to total hotel bookings.

There is no universal threshold according to which a 15% difference is acceptable, and 20% is proof of failure. A small difference can hide two opposite problems; a larger one can stem from a different scope of data. Explanation and change over a comparable period are more important.

Which model to use in GA4

Google documents the data-driven attribution and last-click models. First-click, linear, time-decay and position-based models are no longer available since November 2023. Source: Google's attribution overview.

For the same key event, compare models in the attribution models report. If the channel estimate changes significantly by changing the model, this is a reason to be cautious when allocating the budget, and not a reason to choose a more favorable figure.

Separate the session traffic source from the report that assigns credit to key events. In the monthly report, write the name of the used dimension and model so that next month's comparison makes sense.

What GA4 cannot explain on its own

A guest may see an Instagram post without clicking, hear a friend's recommendation and later search for the hotel name. We cannot assume GA4 sees that entire journey. Path review shows recorded interactions; the absence of a channel from the path is no proof that it did not play a role.

Similarly, a simultaneous decline in Meta spending and Google brand search is not sufficient proof of causality. Both can be affected by season, availability, pricing, and other communication. If you want to assess the additional effect of advertising, agree a comparable control, test duration and decision criterion in advance. With a small volume, the conclusion may remain indeterminate.

How to make a budget decision

Start from the dates the hotel wants to sell and check whether the offer matches availability. Then review the cost, quality of bookings, and data limits.

You can track the ratio of total direct channel revenue and media spending as an aggregate indicator. Do not present it as evidence of revenue caused solely by advertising. For a business decision, include the full cost per booking.

What to prepare for verification

Prepare conversion definitions, attribution settings, and a reconciled, aggregated booking report. First check what to track in GA4 and the booking engine, then arrange a measurement review. The goal is to know on what data we can make a decision and what else we need to check.