Hotel Tonight’s Owning the Night

Why tapping into users’ pasts keep them tomorrow night.

You can only be in one bed per night.

Usually.

Hotel/berthing concierges like AirBnB and Hotel Tonight compete for the same thing—deciding where I want to rest for the night. AirBnB professionalized couchsurfing—Hotel Tonight deals surplus capacity from hotels at the absolute last minute. I’ve had incredible experiences for both—but there’s one area where AirBnB has the edge: Prediction.

I have no idea if the HotelTonight hotel I want will be available the night I will be in town.

That’s based on when HotelTonight gets its rooms—they can’t get them in advance because hotels don’t want to downgrade rates until the absolutely last possible moment. But what if I could base my choice on history? What if I knew what was available before and gamble on that?

Worst case—I could instant book through AirBnB, couchsurf, or find some other discounted fare at the hotel I really want to stay at.

Part of the data stream HotelTonight users don’t have is history—we don’t know how often in the past a hotel has been available. We don’t know how many rooms either (probably because hotels want to keep that secret from their competition)—but at least knowing if a hotel has used the service could be a start.

AirBnB markets is unique and different rentals—palaces, lofts, bungalows. HotelTonight can do the same thing—showing users the amazing hotels that have used the service before. They get hotels across the board—1 to 4-stars. But we only know for the current day—we don’t know the last month or last year.

I’ve tried to use prediction before to guess when a hotel could be available in advance—but it takes too much time for a user. I’d rather just go mainstream, Hotwire, or Kayak instead of gather that data. Especially if there isn’t a payoff.

Hotels could also benefit from broadcasting the percentage of rooms filled through HotelTonight. We don’t need to know exactly how many rooms—but wouldn’t it be cool to know 100% of a hotel’s rooms on the service have always been used? Or 50%? Yes—hotels could start gaming that metric based on how many rooms they want to make available to the service. But would they want fewer rooms constantly booked, or more rooms not booked as frequently?

Makes for an interesting marketing behavior question.

It’s all a gamble—but Vegas succeeds the same way.

Past data keeps people on the site, enhances credibility for the service, and gives hotels another venue to market their availability and desirability. It’s a win-win for users, hotels, and HotelTonight.

Whether it’s future data or past data—longitudinal data helps users make better decisions and be more comfortable in the decisions they make. Data becomes reinforcing—yes, sometimes self-serving (confirmation bias anyone?). But the market of people who book hotels based on a single day of information is small. The market hasn’t been used to that since Manifest Destiny. If HotelTonight can show us their past—we’ll be more comfortable gambling on them for our future.