Stan Pollitt one of the fathers of advertising planning introduced what at the time was a radical innovation. He simplified the tangled web of market segments down to two. Users and non users.  Or heavy users and light users. This was a powerful innovation because you don't just learn by considering the opinions of a group of people. Often it is much more useful to look at the contrast with another group. So this simple initiative made it possible to look at the impact of new advertising ideas on those who already had a strong impression of the brand from usage. And those who had little impression of the brand other than through advertising. Their response to advertising would be different. And the difference was what was useful.

The communications world is an altogether more complicated place now. But these polarities are still useful. One of my first jobs when i started working in ad agencies was to work on a campaign for Guinness - alas not one of the famous ones everyone remembers but during a time when Guinness were trying to lure the fringe drinker so the advertising was aimed at them primarily -  they changed the campaign (and agency) when they worked out that the ads were putting off the core drinkers who were accounting for most (but not enough) of the volume. The new agency (O&M) came up with a brilliant campaign that reminded core drinkers why they wanted to drink Guinness. But didn't scare off fringe users who wanted to drink it more often but were intimidated by the macho associations.

So to summarise don't try to talk to everyone (not even in their segments. Look to create a bi polar order where the contrast is useful by choosing just 2 groups of people described in the simplest way. But make sure you engage with both of them. The numbers usually require you to motivate both groups, you can't afford to target just the core or just the fringe. Or the yolk and the white as Guinness used to call it.