September 2009 - As we enter the annual hotel request for proposals season, I'm struck by how little sourcing innovation we've seen in this important category. Yes, we have a standard RFP format, and several good RFP technology platforms, but sourcing hotels is still like a long walk in the hot sun: You'll eventually get there, but not without cracking a good sweat and wishing for a quicker and cooler way. The good news is today there is a better way, and the story involves some lazy thinking and a bottle of wine.
First, let's agree on some of the fundamental problems involved in sourcing hotels:
•
Dirty data Agency and card data must be merged, de-duplicated and scrubbed.
•
Data overload Which hotel data can you safely ignore?
•
Stale bid lists It's easiest to just use last year's bid list with a bit of tweaking.
•
Clumsy market buckets Using city names or zip codes to organize hotel markets has severe limitations.
•
Small-market knowledge It's really hard to know which chains best fit a program's smaller markets. Is Marriott or Hilton the better fit in Tucson? What about Toronto and Tampa?
•
Defining meaningful markets It's nice to know how many room nights are produced in Chicago, but how many of those nights are really in play for hotel X in the Loop?
•
Credible competitors Buyers want to play one hotel off another, but it doesn't work if the hotels are 10 miles and two stars apart.
•
Predicting savings Management always wants early predictions.
•
Policy options Try calculating, credibly, what a tier-down strategy would save.
•
Compliance analysis Was a hotel's low volume due to poor compliance or a slow market?
Clearly, there's room for improving the hotel sourcing process. The solution is based on reframing a key ingredient in any negotiation--the relevant competitive market. In the airline category, the relevant competitive market is the city pair. City pairs are geographically specific and have well-defined suppliers. The buyer and the airlines know how much volume was produced on the city pair, and they can negotiate for one or any combination of relevant city pairs. I thought the hotel category needed the equivalent of city pairs. That was the extent of my lazy thinking for quite a few years. Now, about that bottle of wine.
One night a couple of years ago I was admiring the artwork on a wine label. It showed a beautiful cluster of grapes and made me ponder how a vintner chooses grapes for a batch of wine. He doesn't pick the grapes one by one, nor does he take all the grapes from a section of his vineyard. Instead he picks clusters of grapes--aha!
I was struck by the concept of clustering hotels into neighborhood-level competitive markets. Throw away the artificial boundaries of city names and zip codes, and just group the hotels that are near each other into logical and local clusters. Make the clusters small enough to be relevant competitive markets, say one to three miles in diameter, more or less. Seed each cluster with hotels getting a significant number of room nights ("high-stay hotels"), and then grab all the nearby hotels as part of the cluster. Ignore any low-stay ("orphan") hotel that doesn't fall within any of the cluster diameters.
Of course, you need to be able to do this automatically for hotels in any city, any country. As long as there is good latitudinal/longitudinal data and a reliable quality indicator (e.g., star rating) for the hotels, no problem. I helped TRX Travel Analytics develop their hotel clustering tool last year, and they are applying it to hotel sourcing projects this season. It's a powerful innovation that will make hotel sourcing projects deliver more value in much less time.
Grouping nearby hotels into clusters opens the door to much more meaningful analysis of any corporate hotel program. This analysis arms the buyer with fact-based insights, quickly quantifies savings options and instantly creates relevant competitive markets for bidding and negotiating.
|
Clustering allows buyers to quantify the negotiating leverage they have over any property, brand or chain, and which competitors are the best alternatives.
— Scott Gillespie
|
|
|
Buyers know with confidence which chains need to be at the negotiating table based on each chain's capacity to serve a corporate program. (These capacity numbers are based on each program's footprint and stay patterns.)
Measurements show how much travelers are supporting or avoiding any single property, brand or chain, so buyers can easily track savings achieved by the preferred hotel program in any market and/or quickly spot preferred hotels that are not getting their fair share of room nights. As such, clustering allows buyers to quantify the negotiating leverage they have over any property, brand or chain, and which competitors are the best alternatives--in every market, no matter how small.
This strategy also enables buyers to quickly prepare relevant bid lists by automatically weeding out hundreds or thousands of hotels that have room nights but are not relevant to any cluster. Buyers can know with certainty each hotel's nearby competitors, regardless of zip code or city name, making rate negotiations faster and more credible.
Clustering provides senior management with quick, yet credible, estimates of savings for: trading down, improving compliance with the existing preferred hotel program and taking advantage of the best rate in any cluster for a given level of hotel quality--roughly equivalent to a spot-buy strategy.
Additionally, the data input for clustering is equivalent to traditional hotel sourcing projects. The clustering process takes traditional hotel data (agency bookings and card payments) and scrubs it as part of the pre-clustering logic--but faster and more thoroughly than any buyer can do today. This eliminates all data scrubbing work on the buyer's side!
Once buyers grasp that clusters are simply neighborhood-level markets, the rest is straightforward. I see no barriers to buyers adopting this innovation and predict clustering will soon be a hotel sourcing best practice.
Suppliers will have some challenges, as well as some significant benefits, from taking a cluster-based view of corporate hotel programs. It remains to be seen if hoteliers will appreciate this view of their world.
Scott Gillespie--founder and former CEO of Travel Analytics, a travel procurement consultancy he sold to TRX--is the author of a U.S. patent covering airline bid analysis and co-author of a patent application covering cluster-based hotel data analysis. He is writing a book on travel procurement. 
