Thumbs-Up to Travel Sites with User Data

There's more to travel sites than travel.yahoo.com and orbitz.com. Sites like Virtual Tourist and Trip Advisor are sources of user reviews for hotels and advice for getting the best from every destination. Where to now, though? There's more to travel sites than travel.yahoo.com and orbitz.com. Sites like Virtual Tourist and Trip Advisor are sources of user reviews for hotels and advice for getting the best from every destination. Where to now, though?

They're data silos. There might be value in saying to these groups "you're collecting all this great data--how's it accessible beyond your wiki or your comments fields?" I look at the rich collections of reviews, collections to rival Amazon's vaunted user data, and wonder: Where are their APIs so that Google Maps/random startup can overlay ratings and search reviews when mapping and finding places?

When I met with Moto Labs on Friday, the VP raised the interesting question of fraud. When you collect user reviews, how do you avoid ballot-stuffing and other poisoning of your data? I'd like to know how many people Amazon has working on keeping the user reviews clean. It's interesting that there are no published best practices in this field.

Amazon's able to mix questionable user review data with the absolute data of what people have purchased, which is something these review sites can't do. The credit rating companies are the ones who know what people actually buy, the analog of Amazon's ability to mine your purchase history. If I were going to build a "People Who Bought Also Bought ..." for the rest of the world, I'd have to go through the credit card companies. That'd let me say "people who spent a lot of money at book stores and bought coffee at this Starbucks went on to buy books from the second-hand book store around the corner ...", as well as use the rate of product return and the number of second purchases to determine quality of the store.

The only other alternative is to collect satisfaction information from people at the time of purchase. At the moment, nobody's going to dig out their phone and thumbs-up thumbs-down a business. But if you used your mobile phone to pay (as they do in Japan) then it becomes a possibility. I still like the idea of purchase histories as implicit thumbs-up behavior for consumers. How does location play into that, though? Does it provide implicit thumbs-down by letting us deduce that you drove past two other stores to get to the one that you purchased from?

Still lost in location ...

--Nat