For all my time at Yahoo I've had the good fortune of working under Tapan Bhat. I worked for him directly for the first year as we tried to sort out My Yahoo!.
"My" was and is a pretty geeky product, with a lot of power user features. Our job was to figure out how to turn it into a mass market product. Having it for several years before abandoning it I felt that I had a great sense of what users wanted in a "personalized home page," and what we needed to do to make the product grow again. I spent endless hours working with our engineering team to spec out what I thought were killer features, only to see them blow up in the lab. (Fortunately, Tapan and the designers kept most of the really dumb ideas from ever leaving paper). It turns out that I had a great idea of what I wanted on a personal home page, but no clue as to what our users actually needed.
At the time our strategy was to upsell front page (yahoo.com) users to My Yahoo! I realized that I was nothing like a typical front page user, and so everything I came up with fell flat. To make matters worse, I had a hard time identifying and empathizing with those users, and reading the research didn't help.
What finally worked was to use people I knew, who used the home page, as personas. I had classmates, parents, and siblings who all used the front page, and like to personalize things. So what worked for me was to put myself in their shoes and reconceive each potential feature in terms of how - or if - they might understand it.
Eventually, putting myself in their shoes became second nature. But from time to time I still get worked up about a product I own not working exactly as I might want it to, or a feature I really need getting deprioritized by someone on my team. If they're wrong, they're wrong. But often times I just take a deep breath and then slowly say "No, Joseph, you are not the user."
Marty has his own, similar rule in his "Top 12 Product Management Mistakes."