In all organisations it’s really important to be delivering the most valuable thing, and in many organisations it’s obvious what that is—and people just get on with it. But in many organisations—particularly large ones with many competing demands—it’s important to estimate the value of our work to determine what gets delivered next with the limited time and resources available.
This can be more difficult when we break work down. The value of a project may be obvious, but when we have to choose which of the constituent parts to deliver in what order it’s sometimes tricker. If every feature or deliverable has an independent monetary value, say, or some other real-world value, we can estimate fairly. But when the individual features are more complex, and they’re not directly comparable, it’s more difficult.
When I was speaking to my friend David Stoughton about this he suggested a different approach. Rather than trying to claim that this feature delivers £100,000 of value and that feature delivers £80,000 of value, consider them as part of a whole. Think about the project delivering, say, 10,000 points of value, and then ask how many of those points are contributed by this feature or that feature. The shift from money (or whatever our real-world measure of value was) to points makes it easier to compare unlike features, and thus allows us to take a more top-down approach. By contrast when we only think of real-world metrics we fall into the pattern of estimating value bottom-up—i.e. estimating the value of each feature in its own right.
So we have two approaches to estimating value, and when we want to compare unlike features then estimating top-down is helped by abstracting a little bit away from the real world.