I find the Pew reports really interesting. Here’s a snippet from a recent one:
Pew Internet: Home Broadband Adoption 2007
47% of adults have high-speed internet connections at home as of early March 2007, up five percentage points from a year earlier.
For those of us living in Silicon Valley, the idea that only 47% have broadband is pretty unthinkable. Here we have broadband galore! Heck, there’s free WiFi in Mountain View, Sunnyvale, and a few other cities. Meanwhile, in places like South Korea, broadband penetration passed 70% in 2003.
Yet, we as a nation only made a 5 percentage point gain year over year. In fact, this was one of the slowest years of broadband uptake according to the report:
The fact is that the United States of America is a pretty huge country, and we have people living scattered throughout. It’s going to be really hard… no… really expensive to get everyone on broadband as we don’t live in densities like they do in Korea, Japan, etc. How will people in South Dakota and much of New Hampshire ever get broadband? I suspect that there’s probably a good correlation between Starbucks availability and broadband penetration. 
Another point is that the definition of broadband is kind of fuzzy – my parents’ broadband is 992kbps where as mine is 15,132 kbps. Yep, mine is almost 15x faster. And it’s not like my parents live in the middle of nowhere either!
For those of us working Internet applications and services this is something to keep in mind: broadband isn’t everywhere, and it probably won’t be everywhere for a long time. And even then, the speed won’t be all that consistent – some people will have super fast access, and some people won’t. We must build products and services that are usable by all of these people, if we want to have the largest possible customer set.