Your bid engine spends most on the rooms you would have sold anyway

A metasearch bidding vendor shipped a product this week on the premise that platform bid automation pushes your budget toward the properties already winning — and said so in its own marketing

by Markus Busch

The launch. Koddi, which runs hotel campaigns across Google, Booking, Expedia, Trivago and Tripadvisor, released a feature called Occupancy insights on July 15. It lets a hotel group feed its own occupancy, MPI, RGI and direct-share figures into the bidding layer, so a property dropping into a low-occupancy tier automatically triggers a bid boost on specific check-in dates. A routine vendor launch. The interesting part is the problem it names to justify itself.

The sentence. "Traditional bid automation optimizes for clicks and return. But this often forces your budget toward already-popular assets, higher daily rates, and historic click-through patterns." And a line above it, describing the publishers' own approach: it "traditionally rewards already-popular listings."

Read who is saying it. That is a bidding vendor telling hoteliers that the automation the publishers offer spends their money on the hotels that were already selling. Koddi is not a disinterested witness. It sells the alternative, and the post ends at "contact your account team." Take the pitch at a discount.

The mechanism isn't in dispute, though. An algorithm optimizing on historic click-through and conversion will favor the properties with historic click-through and conversion. That isn't a flaw in the system; it is the system working as designed. The room that sold last summer looks like the safest bet for this one, so it takes the budget, and it sells again, and the pattern hardens. Efficiency and self-fulfilment look identical in the report.

You have seen this shape before. It is the AI-discovery problem, one layer down. Up top, the model names the hotel it knows best. Down here, the auction funds the hotel that already won. Both reward established presence and call the result performance. The difference is that this one arrives as a line item, and you pay it monthly.

The limits. None of this is measured. Koddi publishes no data showing that platform bid automation skews toward incumbent properties, and no independent party has tested whether it does. What exists is a mechanically plausible claim from a company with a product to sell. That is a reason to look at your own numbers, not a reason to take theirs.

What to do Monday. Pull your metasearch spend by property and by date, then ask which of those bookings would have happened without the bid — the high-occupancy dates, the properties that fill regardless. Koddi's own dashboard flags exactly that as "critical overspend," which tells you the pattern is common enough to build a feature around. If your budget concentrates where you were already winning, the algorithm is not finding you demand. It is billing you for it.

Enjoying this analysis? Hospitality.today delivers daily insights on hotel distribution, AI trends, and travel commerce — straight to your inbox. Subscribe for free at Hospitality.today →