Many of the software people that use Agile recommend that teams track the velocity of the team to know how big of a scope they can realistically take on each iteration. Others warned of how management tries to use the velocity metric to compare teams, which it is not designed to do.
As it turned out for us, we had to use earned value management reporting anyway and had to meet the estimates made up front to be profitable. So for these situations, we did not need a velocity metric, so we ignored it. We had to impose the pacing on the teams. We had to explain why we were not hitting the goals any time we missed one. This is not ideal, but it was our situation.
This is why from our perspective, the cadence of the two-week iteration and the iteration goal of X number of LOs per day was what helped the entire team reduce the costs of tracking where everything was. The structure of the visual board helped minimize the transaction waste of the daily standup meeting. The shared accountability helped the team manage itself to a higher degree than we’ve been able to accomplish on other projects.
It just turns out that a velocity metric did not really help us. So we ignored it. You will have to determine that for you.
One of the really strong ideas behind Lean-Agile is that if something is not working, change it. It is not a dogmatic approach of specific practices. If velocity works for your team or sets of teams, then by all means use it. For our circumstances it did not provide value, so we dropped it.