I benchmarked it against what I used before in my own time, and it was worth the effort. The short answer is that how blocking actually behaves is the first thing I would look at now. It was noticeable in a way I could measure once I had a few weeks of data.