TripAdvisor promotion incentivizes users to submit hotel reviews in return for frequent flyer miles. A promotion was run in India last year and another is currently running in Germany. In theory this should not bias reviews because, unlike some hotel loyalty program schemes, reviewers receive the rewards regardless of whether the review is positive or negative. Properly controlled, this could improve the quality of reviews. This is because reviews submitted to sites like TripAdvisor can be prone to response bias—the people most motivated to submit reviews are those who have had experiences at the negative or positive extremes. By providing an incentive that is independent of users’ experiences, more people with typical experiences are likely to submit reviews.
Unfortunately, there’s nothing in the article to indicate that the TripAdvisor program incorporates the most basic of controls—verifying that submitters have actually used the product or service they’re reviewing.
TripAdvisor has a team that is charged with policing fraudulent reviews. It’s led by a new director of content integrity.
Yesterday I noted that Google was positioning ‘Sponsored Results’ in organic search results. Tnooz reports more details today with the news that Google has positioned Hotel Finder atop organic search results. Flight Search has been there for a while but may now at least get the ‘sponsored’ label.
Planning a Disney vacation? Tactics for finding the cheapest airfares to Orlando or Anaheim. (Actually the same general tactics will work for most destinations.)
Kayak wins Webby Award for best travel website.
Delta airlines buys refinery to control fuel costs. Airlines have commonly used hedging to control fuel costs, but owning an actual refinery is an interesting airline first.
Online marketing analytics: Last-click attribution models have a lot of drawbacks. The article talks about some other methods of attribution, the best of which is probably a linear model where all interactions are assumed to share evenly. I have a more scientific suggestion: systematically experiment on the key interactions along the conversion path to find out which ones are significant and what their weights are. This requires that you have sufficient statistical power, statistically educated personnel and managers who are willing to invest in proper analytics rather than voodoo, but the results will actually reflect reality.
After writing the click attribution through experimentation blurb above, I just happened to check Avinash Kaushik’s blog and he said the much the same thing, only in much more detail and with much more rigor. As I’ve said before, Avinash wrote the book on Web Analytics. If you need to measure attribution and do it right, stop what you’re doing and go read Avinash’s recent post on Multi-Channel attribution. If you do any sort of web analytics in your work, you should be regularly following his blog.
Occam’s Razor by Avinash Kaushik
Google offering big-data analytics in conjunction with its cloud storage.
Internet advertising: The law of [scatological] clickthroughs. Andrew Chen notes that advertising response rates decline over time, not just by promotion but by ad type and placement. As a result, marketers need to continually change their promotions and channels if they want to maintain engagement. The idea is hardly new; traditional advertisers have known this since before the days of Don Draper. The article has some nice charts and heat maps illustrating the phenomenon.
Will apps kill websites? Good overview of the pros and cons of apps as compared to websites with lessons learned that be applied to both.
Samsung is now the top mobile phone company, beating Apple in smartphone numbers and Nokia in overall mobile numbers.
Just One More: Ultimate geek-out: 3D-Printed Electromechanical Computer.