This one is a pretty common tradeoff that we sometimes fail to even acknowledge. So you heard that you have to talk to your users to learn about their problems and needs. But you also read somewhere that you have to be focused and say no. So you’re watching users and talk to them and then you realize that that user you’ve been completely focused on — the one you decided to focus on so that your product will be focused — does not really exist. You realize that you have several types of users, each type with different needs and problems.

Or maybe you don’t have a product yet. Suppose you were a new car manufacturer and you‘re watching users drive your prototype of an SUV (electric vehicle, of course). You notice that six out of the twelve you tested think it’s too big or heavy and would want a smaller, lighter car. The other six are actually quite happy with the prototype. What do you do?

Let’s say, for now, that those six interviewees who tell you they want a smaller car are really representing half of your overall market (we will get to the problems with this assumption in later stories). Do you stay true to your current design? Do you pivot to make a smaller car? Maybe you can decide to manufacture two models: a compact car and an SUV? Or maybe one intermediate-size car as a compromise (that classic horse designed by a committee)?

In software, unlike in car manufacturing, it is possible and very tempting to ship a product that is both big and small. While it’s hard to imagine a car that can change size with a click of a button, letting users customize the product is possible in software. This may make your bosses or investors expect you to address all those issues in your product use cases in your product. A car that is both small and big! Magic.

But extending your product to make it customizable does not come without a price. Sometimes use cases contradict each other: for example if advertisers, publishers, and readers differ on how many ads they want to see. Even when separate solutions can coexist in one product, the mere multiplicity of them are distracting: they add more items to your checklists, add to technical complexity and therefore make future development harder. If you had five different templates and you wanted to add a new feature you had to add it into all five templates. The time you invested in the three least-used templates could have been put into getting the two top ones upgraded sooner. Each variation of your product is like another weight you’re putting on your team. You may be OK carrying those weights now, and you’re showing off to everyone, but once you try to run to beat your competitors you will probably regret that. Or worse, you won’t even have the time to see your competitors sneak behind you.

“Yes!” You say. “I need to stay focused. I’m dropping the customization and going with one product.” Which will it be then? Small or big? Maybe something in between to satisfy everyone?

Answering this question depends on where your users really are: what is their sweet spot (optimum) and how much do they care about your product getting away from that sweet spot.

Economists use location models to try to describe where firms choose to place their products. These models usually focus on the behavior of firms competing for a location and how how good or bad their choice was from a social standpoint. But we don’t care about industrial organization. Competition or monopoly, the basic premise stays the same: firms will choose the location that will maximize their revenue. Revenue will be affected by the product’s location because customers will find higher value in products located closer to their optimum. Those customers who end up too far from you will vote with their feet.

So if your customers’ sweet spots have a bimodal distribution—having two separate peaks—then you’ve got distinct segments and trying to find a middle ground is going to end up providing lower value and lower revenue. You’re better off staying with your prototype, or dropping it and going all the way to the small screen size (yes, we switched from cars to screens because everyone is on mobile now). Whatever you do, don’t try to find a middle ground. This is shown in the figure above, plotting customers’ sweet spots on a continuum of possible screen sizes. There are two types and customers are split evenly between them.

But the same interviews could have actually represented a different distribution of user types. A more continuous one where customers can have several intermediary sweet spots. This distribution is shown below. When that is the case you definitely need to go for a middle ground because it will be closer for a larger number of customers relative to your prototype. It may still alienate some of the extreme customers: the ones on the edges, who may get very vocal. But if you decided you’re only going to ship one product then that’s the best place to be

“Wait, so what do I do with this? How can I know where my users really are?” Well, you can’t really, but you need to aim there. Users never tell you their sweet spots. In many cases they won’t really know until they tried. Still, it is important to do as much as possible to see how distinct your segments are, and how much the value really drops for customers when it gets farther from their sweet spots. Both will help you determine whether you really want to stick to one product and where to place your product(s).

In the mean time: keep your focus and keep listening to your customers. It is a tradeoff you need to acknowledge and constantly manage.