As per Erik's comments on my last entry, I agree that with metrics there comes a danger of "teaching to the test."
That is part of the challenge of developing appropriate metrics (that take into account cultural aspects of the community served, for example). However, metrics are important to determine the impact made on the community and, although at times they may be difficult to qualify (or, even more so, quantify), they can help to make the most significant impact (and make sure that the venture continues to maintain its standard of service/support over time).
Despite the danger of "teaching to the test" with metrics, I assume that social ventures aim to make a positive impact. I think that being able to measure some aspects of the impact attests to the societal value of the organization (whether the metrics are quantitative or qualitative). I guess these metrics have to be very specific to the organization, however, in order to avoid the problem of broadly-defined outcomes that might encourage "teaching to the test." However, if teaching to the test means that there are some positive societal outcomes, it might not be a negative approach.