While I generally don't enjoy 2011 prediction posts, I really loved James Hoskins's article (20)11 predictions from the CMS coal face. I found it thoughtful, pragmatic, and meaningful. One of my favorite predictions is #8 Personalisation falters again. I have had similar challenges with personalization. Most content organizations do not have the maturity, discipline, and energy to fully leverage this kind of technology. Interest and investment inevitably subsides after irrational expectations are not immediately gratified. Investment is higher than expected: you have to manage more complex technology, you need to develop more content, and you need to test a lot more. Return is lower than expected: most companies don't even know how to measure the value of the return.
There are, of course, exceptions. The importance of display and placement are ingrained in retail culture (online and offline) and small tweaks can lead to big returns. Customer extranets are inherently personalized so I am not including that here.
Traditional media companies (magazines, newspapers, and television/radio stations) would like to personalize but they don't have quite the upside eCommerce sites do; on a media site, an extra click means a few more ad impressions rather than a potential sale. Most media clients find that the volume and turnover in content makes the cost of tuning personalization greater than the return. Showing articles that are related to the current article brings the biggest bang for the buck. Marketing sites are more campaign driven than personalized; you put up different pages with different URLs rather than creating a personalized user experience.
As problematic as it has been in the past, I think that succeeding with personalization is only going to get harder. First of all, multiple device support is going to eat up your development resources so you will have even less time to test and tune complex personalized views. Second, unless you are lucky enough to be one of the premier social networking sites, most of your audience will only be on your site for a page or two — if at all. Most visitors will follow a deep link to your site, scan the content, then go back to the conversation about it. Some visitors will take snapshot of the page using a service like Read It Later or Instaper. Some visitors will just read the conversation and never click through. The upshot is that most visitors will not hang around long enough to implicitly or explicitly build a profile that can drive personalization logic.
I think that the greatest potential for personalization will be to use a service like Facebook Connect like Levis is doing. When a visitor comes with Facebook Connect, they bring some additional context that can be used to drive personalization logic. Most web savvy people find Facebook Connect creepy (and will try to avoid it) but the vast majority of web surfers are either unaware or unconcerned with privacy issues. I would look for a major uptake of Facebook Connect. In particular, I expect to see recommended content display components that can be used in different presentation channels. In the near term, however, I the greatest returns will come from the "make every page a home page strategy" where each page promotes content that is related to the current page. That's not personalization. That's just good content re-use.