A “bottom-up” technique for discovering connections between pieces of data. An individual or group starts with one piece of data (say, a customer need). They then look through the rest of the data they have (say, statements of other customer needs) to find other data (needs) similar to the first, and place it in the same group. As they come across pieces of data that differ from those in the first group, they create a new category. The end result is a set of groups where the data contained within a category is similar, and the groups all differ in some way. See also Qualitative Cluster Analysis.