Personalization does not have to be “Creepy”

On Sunday, The New York Times ran an article arguing that retailers’ personalization attempts are backfiring because they are becoming too personalized. The piece quoted several retailers who had tested sophisticated personalization tools with poor results. They concluded that consumers just don’t like it if retailers know them “too” well.

I think that conclusion is off the mark.  Retailers are having problems not because they are too personalized but because they are not giving customers the four things needed to make personalization succeed:  Transparency, Control, Accuracy and Trust.

Let’s start with transparency. People feel something is “creepy” when they either don’t understand it or didn’t know it was happening.  Personalizing your site is meant to help the customer–so treat it like the awesome feature it is instead of a devious secret you don’t want the customer to know.  Announce on your Web site that “The site will get better with every click you make.” Tell customers from the beginning what you think you know about them (and how you figured it out) and update it as they click so they can see how your algorithm works.  Most people are secretly enamored with what technology can do and a site can go from “creepy” to “cool” very quickly if you just explain how you do your magic.

Second, to succeed with hyper-personalization, a site must give their users control.   People really like to be in control (just read Dan Gilbert’s popular book Stumbling Upon Happiness).  When personalizing a site, give your users a chance to change what you know about them, and/or turn it off completely.  Most people won’t make a change, but everyone will appreciate that they could.  On top of that, people have a strong psychological urge to change things that look incorrect. So if you make it easy to change the things that might be wrong, you’ll have users happily giving you the correct information.

Third, to win at personalization, a site has to be accurate.  The New York Times article cited an example about women with families who complained that a retailer showed them only women’s clothing (and not men’s) because of their gender. This is not a problem of hyper-personalization; this is a problem of inaccurate personalization.  Companies assume the best way to “understand” a user is to watch the few things they click and buy (and where else they travel on the web) and make judgments based on that, but that is the equivalent of watching a person shop for a few pairs of jeans and shorts and then assuming that they are not likely in the future to buy shirts.   Great personal shoppers interact with their customers and start a dialog–they talk to them and find out what they really need both now and in the future.  If you really want to be a retailer who differentiates on personalization, you are going to have to do the same.

Which leads me to the final  (and probably the most difficult) challenge:  Trust.  Consumers will be willing to give you their information and let you use it if they trust that you have their best interests at heart.  If your company is using personalization as a way to boost revenues, you’ll only get so far —  the customer will quickly become aware of this and feel exploited. If, on the other hand, your company truly is dedicated to building a great customer experience (even above revenue) your customers will be much more willing to share their information.  Most retailers would say they are customer-focused but few really are.  A good litmus test:  Would your company prioritize an investment in personalization if people loved it but it was guaranteed to produce no additional revenue?

So there you have it–sites can succeed in personalization, but they need all four of Transparency, Accuracy, Control and Trust.  Or to say it in a pithy way, sites will need a whole lot of TACT.