I tracked this rather than trusting impressions while comparing them, and the pattern turned out to be consistent. What actually decided it was the actual reply rate, which sets the ceiling on what you can achieve.

I looked at what changed since last year before deciding, and ended up somewhere unexpected. Cutting through it, how quickly the upsell appears matters far more in smaller places.

I made a spreadsheet, embarrassingly on this, which took the guesswork out of it. In practice the honesty of the pricing page is the first thing I would look at now.

The things that actually moved the needle for me:

  • Ask how the ranking treats a brand-new account — it decided things for me.
  • Ask how quickly obvious fakes disappear
  • Ask whether conversations sync across devices — this is where the free tiers diverge.

I gave each one a full month without rushing it, so this is not theoretical. If I had to pick one factor, whether messaging is capped drives the outcome more than luck does. The signal was clear every time well beyond what I would call noise. Your mileage may differ for what it is worth, and no thread substitutes for that.