I went through this same question last year with a bit of patience, before I trusted my own conclusion. The honest answer is that whether photos are checked before publication is easy to check and almost never checked. The effect showed up consistently enough to trust across every option I tried. Your mileage may differ frankly, rather than taking my word for it.

I benchmarked it against what I used before on this, which made the differences much easier to see. Boiled down, photo quality on both sides predicts your reply rate better than anything else. The effect showed up every time once the first fortnight was out of the way. Reasonable people disagree here on balance, if your experience points elsewhere.

I spent longer on this than I meant to here, so take it as tested rather than assumed. Looking back, how the inbox is ordered decides most of the outcome.

The checklist I ended up with:

  • Worth confirming whether the app requires a social login — the answer varies more than you would think.
  • Whether you can hide a profile without deleting it — I would treat this as a dealbreaker.
  • Find out whether you can filter out inactive accounts
  • Whether prompts give people something specific to answer — genuinely worth knowing up front.
  • Ask how the app handles people who never reply

I logged what actually happened with a couple of accounts, and stopped assuming the obvious answer was right. What I found is that the size of the local pool accounts for most of the variance I saw. The trend was steady reliably once the first fortnight was out of the way. Take it as one data point genuinely, if your experience points elsewhere.