When we measure our progress there are—by some perspective—three kinds of things we might be measuring.
The most immediate things to measure is our activity: Are we doing what we said we would do, and at the rate at which we said we’d do it? This is about measuring ourselves against our plan and is the most basic kind of measurement. Most project managers would measure this as their stakeholders ask “How are we doing against our plan?” It doesn’t usually tell us anything about the success of what we’re doing though.
The ultimate, most distant, thing to measure is the goal we’re trying to achieve. For example, if our ultimate aim is to increase sales then measuring the sales before and after our work is important to demonstrate success. Taking checkpoints in the middle is valuable, too, to show progress towards a goal.
But often we also need to measure something in between these two. For example, in the case of wanting to increase sales there may be many things which influence the ultimate figure which are largely outside our control—competitors’ offerings, the weather, and so on. Because of these influences measuring success of the ultimate goal doesn’t give us sufficient clarity of our own project’s success.
Therefore much of the time it’s also important to measure a third thing, that lies between the two: the immediate impact of our activity. This is more than just measuring our activity, but not as grand as measuring the overall goal. It allows us to see how well our plan measures up to reality. It’s also valuable to construct our plan so that we can get early sight of our impact, as that gives us a feedback loop to adjust the plan as we go along.
There is a line from our activity, to the impact of our activity, to our overall goal.
For a more detailed example of this see the article the Matthew Leitch and I wrote on “How does a risk expert behave?” (also on this blog) and look for the section on deciding what to measure.
niksilver.com 
Proxy metrics are often perilous, though – c.f. Goodhart’s Law.
@Brunns, Good point, and I entirely agree. My aim here was to make the “impact of our activity” metric something we can be directly responsible for, but always with the caveat of “…because we want to improve [distant end goal]”. And if it looks like it’s not having the right impact on the end goal then we should be able to change that more immediate aim. We should never forget the end result we’re trying to achieve.