I was thinking about this ever since writing about Feedburner last night. No one will argue that you need to know who your customers are. Figuring this out can be surprisingly difficult, especially if you have to confront "popular wisdom" held by the senior managers/founders/money guys who already think they know who your customers are.
An easy exercise for exploring this is to pretend that you're a freshly-minted new associated at McKinsey. Sit up straight in your cube and say "Me See". Ignore your cube-neighbors when they accuse you of imitating a bad 50s era asian movie stereotype, and say "Me See" again. Doesn't that feel smart?
Me See, or MECE, stands for "Mutually Exclusive, Collectively Exhaustive", and it's a principle you need to keep in mind when the time comes to describe the taxonomy of your target market.
Start at the top with the largest possible group of potential customers. If you were selling the canonical bag of chips, you might start with "people who eat". You'd have to conclude that some people who eat aren't going to eat chips - children under 2, for example. So below "people who eat" you'd create two groups: "Children under 2" and "People older than 2". Congratulations, you've segmented your market.
Keep segmenting that target audience until you've arrived a sub-grouping of your market below which it doesn't make a lot of sense to go. Think demographics, psychographics, and other inclusive/exclusive qualities, then see how your target market breaks up.
For a complete rundown on how McKinsey does this, go buy Ethan Raisel's book The McKinsey Way.
I used this technique earlier in 2005 when I looked at the target market for installation tools. No matter how I sliced it, the opportunity wasn't getting bigger. So I looked at it another way - what would happen if our target market wasn't just "people who write installations" but "people who write software"?
(For those of you scoring at home, this is working MECE from the opposite direction - instead of top -> down, I looked at it bottom -> up. Same concept, though.)
The result was a concept for a new product designed for application developers - they may not write the installation, but they sure do have a hand in it, since they are the "keepers" of installation requirements. Give them a way to capture these requirements, and make it easy for the installation guy to incorporate the requirements into the install, and viola, you should have a no-brainer of a line extension that could actually make life easier for anyone who produces software.
Decide for yourself if it makes sense. Then go back and look at how you've defined your marketplace and see if your "traditional" way of defining your customer base merits a fresh "me see" look.
For the record, if you're not constantly challenging how you define your market - and how your market defines you - then you'd better find another line of work.