My original comment was really focused on my narrow read of the question in the thread title. My comment is basically "because there is clear A/B test data to support the decision".
I take your comment as a valid answer to a broader question -- something like "thinking about the bigger picture, and envisioning all possibilities, how could and should SFUSD work?" I have a lot of thoughts on that too.
But if I take your comment as a direct rebuttal, on my same topic, I think it's an impractical way to manage the system. It's like you're arguing that middle school alg is a cool feature that shouldn't cause the app to fall over, and we should redeploy the cool feature and reengineer everything. I don't agree that your conclusions flow from your premises and I think the pragmatic approach (given your premises) is to disable the feature while we do the reengineering.
I take your comment as a valid answer to a broader question -- something like "thinking about the bigger picture, and envisioning all possibilities, how could and should SFUSD work?" I have a lot of thoughts on that too.
But if I take your comment as a direct rebuttal, on my same topic, I think it's an impractical way to manage the system. It's like you're arguing that middle school alg is a cool feature that shouldn't cause the app to fall over, and we should redeploy the cool feature and reengineer everything. I don't agree that your conclusions flow from your premises and I think the pragmatic approach (given your premises) is to disable the feature while we do the reengineering.