Jupiter Personalization Story Only Gets It Part Right
Posted by Bill Rini @ 12:10 pmIf you're new here, you may want to subscribe to my RSS feed. Thanks for visiting!
Hi [FILL IN YOUR NAME HERE]:
This post has been specifically written just for you!
Ok, so I’m trying (probably unsuccessfully) to be funny. Jupiter’s report on web personalization seems like the typical hack job done by someone who only partially understands what they’re talking about. The study says that most companies would enjoy a greater benefit from improved navigation and search. Well, duh! From the C|Net summary of the report it seems as if Jupiter failed to draw a dividing line between sites that were poor to begin with and tried to patch that up with personalization and those that were using it for strategic purposes. That seems to be a fundamental flaw which can poison the rest of the research.
Over the years as a consultant as well as the head of various technology departments I’ve heard the “personalization” request many times. Sometimes it’s as simple as putting “Hell Bill” on the top of a page or as the greeting line in a mass emailing. Other times it means providing a different experience based on user preferences. Seldom though has it been about offering a specific product or service based on the user profile which is where the true value of personalization comes in.
I’ve probably used this example a hundred times but I used to do quite a bit of travel to and from London on business. I specifically flew a certain airline (because my company was paying for the trip and required it) and I specifically flew the same flight coming and going because the flight to London put me there at a time where I could still get from the airport to the office with just enough time to catch one meeting (which was a requirement the company imposed for bumping employees up to business class) and the flight home was timed so I had the maximum amount of time to sleep before returning to work. I don’t remember exactly how many times I flew that route but I felt it was enough that when I went to the Expedia website they should recommend it to me when I told them my itinerary was LAX to LHR. Instead I had to hunt through the flights and find my preferred itinerary each and every time. Now, that’s an example of where personalization would be very valuable and would probably add a certain degree of customer loyalty based on the learned preferences of the user.
Amazon has the right idea as well and so does NetFlix. Both keep track of items you like and dislike and then make recommendations based on some sort of collaborative filtering. It’s not difficult to see the value in that to both the user and the merchant.
The problem is they are but the few shining examples in a mess of personalization attempts. But let’s not just limit ourselves to the online world. I get a snail mail letter every other week from the same mortgage lender offering me a great rate on a second mortgage on my property. The offer reads something like this:
Congratulations Bill Rini. I have been authorized to offer you a rate of X% on your property located at . . .
Which would be cool if I owned that property but I don’t. I live in an apartment which should have been the first clue that they might not want to offer me a second mortgage without some further research but more importantly, I have banking accounts with these people and I’ve listed on numerous forms that I rent. On top of that, I’m fairly confident that my landlord might have some issues if I tried taking out a second on his building. They’re just pouring money down the drain every few weeks when they send me those offers. How many other people are they sending the same offer to who don’t own property? I live in a neighborhood that is almost all apartments so even some very basic demographic research would have proven that the offer should be limited to the very small number of people who own the apartment buildings instead of to every resident.
My other “favorite” form of offline personalization is when you get a form letter with your information mechanically inserted into it. The fonts are usually different and since it was done with two different methods the ink, the alignment is just a tad off, and the whole look and feel are different. I always get a good chuckle before throwing them in the trash.
The thing about personalization is that it’s not easy. It takes time to develop a good system. You have to have good metrics to measure. You have to know why you’re doing it. Amazon and NetFlix understand that the more they can recommend that is on target with your tastes the more you’re likely to purchase. But hard doesn’t mean expensive. Travel sites like Expedia or my airline should start to clue into the fact that my most recent trips have been to places like Phuket (Thailand), Cozumel (Mexico) [twice], Maui [twice], The Bahamas, and Grand Cayman and quit sending me information about special fares to Chicago. Theoretically I could have a business opportunity pop up in Chicago and I might take them up on one of their specials but I’m far more likely to purchase a great deal on a trip to a similar location as to the ones I’ve been to in the past. They have the data on my past travels so it shouldn’t be that expensive to develop a system that primarily offers you specials on trips to those areas. Heck, even if they got it down to the region (Caribbean, South East Asia, etc.) I would be much more happy than getting another “Super Special Fares” email offering trips to places I would normally not go to unless I had some compelling reason to go.
The report also mentions that people feel uneasy about handing over information to websites. I do but not because of privacy issues but because it doesn’t lead to a better experience. Wow, they put a little weather ticker with Los Angeles weather because I gave them my zip code. Big whoop. I want real personalization in exchange for my data. I think a cool offering from someone who had the traffic (Yahoo, MSN, AOL, Blodgex, Popdex) to make it really feasible would be a personalized news service. No, not the pick these 10 news sources and we’ll send you the headlines but how about a link at the bottom of every article asking you to rate it? Again, it’s collaborative filtering but it could recommend stories that may be of interest based on my previous rankings of other stories. I might have an interest in politics but with a particular view. Isn’t that what many blogs are anyway? Many (but certainly not all) are nothing more than links with a sentence or two of commentary. In effect, they’re just a human replacing software in the collaborative filtering process.
The trick with personalization is that it can’t be purchased in a product or by hiring some consultants to come in and do it for you. You have to know your business and know your customers. In fact, personalization shouldn’t even be conceptualized by you as personalization but rather as “if there was only a way to do X it would really drive some additional revenue per customer.” If X happens to fall under the umbrella of personalization then fine but personalization for the sake of personalization leads to exactly the kinds of failures that has prompted Jupiter to conclude that it doesn’t work. If you’re thinking “if I could only do X” way, you’ve already milked better navigation and squeezed all you can out of better search capabilities. You don’t see personalization as a silver bullet but rather as a continuously evolving method of better understanding and meeting the needs of your customers.
License
This work is published under a Creative Commons Attribution-NonCommercial-NoDerivs 2.5 License.
If you enjoyed this post, please consider getting our free RSS feed so you can be notified of other posts like this.
- BROWSE / IN TIMELINE
- « More on the Email -> RSS Gateway
- » Leaders or Rock Stars?
SPEAK / ADD YOUR COMMENT
Comments are moderated.
