Learning Trumps Results
For obvious reasons, we care a lot about "results" in the business context. Like all good things, however, results has a bit of a shadow side. By focusing so much on the results, we often will seek to get (or at least try to prove) results at the expense of actually learning from what we're doing.
In The Lean Startup, Eric Ries makes reference to "vanity metrics" in the world of technology startups. If you're building a new app, for instance, tracking a gloriously upward trend of the number of users may make you feel better (look at the "results" we're getting!!), but unless that helped you learn how to make the product better or more relevant to your market, then you're not actually making progress. What are those users doing, and can we say (with confidence) why they are choosing to act that way? That kind of learning is valuable, so let's get metrics that track those results.
Want to improve employee engagement? Then be careful about simply tracking an engagement score from year to year. Unless you can craft some powerful hypotheses that you can test, it may amount to wasted effort. Results (and their metrics) still matter, of course, but if you don't have real learning at the foundation, you'll end up falling farther behind DESPITE the results you're getting.