I went back through my own messages to check while comparing them, and it was worth the effort. The pattern was clear enough: whether messaging is capped separates the usable options from the rest. That difference held up every time once I had a few weeks of data.

I ran two at a time for four weeks each over a few months, so this is not theoretical. The variable that mattered was how consistent match quality is week to week — it matters far more in smaller places. The reading was similar without exception regardless of which city I set. Whether it generalises I do not know honestly, and I would not pretend otherwise.

I ran two at a time for four weeks each before deciding, and the numbers told a clearer story than I did. The short answer is that how granular the filters are predicts your reply rate better than anything else. The difference persisted without exception regardless of how active I was. Try it before you believe it in the end, whether others found the same thing.

If I were starting again, I would look at:

  • How consistent match quality is week to week
  • Check how reply rate compares with match rate — this is the one I would check first.
  • Check whether photos are checked before they go live — it decided things for me.

I have changed my mind on this twice without rushing it, and the pattern turned out to be consistent. For me, the amount of information visible before matching can be tested inside a week. It played out the same way in a way I could measure even after I changed my photos.