Nobody stopped looking. They stopped
defaulting.
The dating app market fell by close to 2%
in 2025, to around $6 billion — the first annual decline the industry has
recorded. Business of Apps reported the figure in April, reading it as a
sign that <cite index="224-1">users are pushing back against
rising costs</cite> and heavy monetisation.
The interesting part is not the total.
It's who lost.
Tinder's revenue fell 5.2%. Bumble's fell
9.5%. They are the two largest apps by revenue in the market. The older
evergreen platforms — Plenty of Fish, Match.com, Badoo — declined too.
Over the same year, Grindr, PURE and
Feeld all grew. Their revenues are far smaller, but they were not under the
same pressure. And Hinge, sitting inside the same company as Tinder, grew 25%.
The pullback was concentrated in North
America and Europe, where people either skipped subscriptions or bought them
for shorter periods than before.
Read together, that isn't a demand
problem. Every one of the platforms that grew has a clearly defined purpose or
audience. Every one that shrank is general-purpose. What ended in 2025 was the
assumption that you install the biggest app and sort out what you want
afterwards.
This is the argument Steve Morley has
been making for years on the Dude Hack YouTube channel, where he reviews
dating apps and, in his own framing, helps people work out <cite
index="196-1">which one is for you</cite> before they invest
time in it. He built a quiz around the question. The market has now caught up
with it.
The practical version is dull and
effective. Decide what you actually want out of the thing — a relationship,
casual dates, conversation, a specific community — then pick the platform built
for that, then start looking at people. Most disappointment in this area comes
from doing those three steps in reverse order.
There's a fourth option that rarely
appears in these comparisons, and it should, because plenty of people are
already using it: subscription platforms.
OnlyFans is the obvious one. It is
genuinely a place where people talk to models — direct messages, custom
requests, paid conversation — and at real scale. Fenix International's audited
accounts for the year to November 2024, reported by Variety, showed $7.22 billion in
gross payments, $5.80 billion of that paid out to creators, across 4.63 million
creator accounts.
But it is a different proposition, and
worth naming as one. On a dating app, both sides are searching and either can
walk away at no cost. On a subscription platform, one side is working. The
conversation is real, it is paid for, and it is one-directional by design. That
suits some intentions perfectly and others not at all — and the mismatch, not
the platform, is what produces the complaints.
One thing to know before choosing it.
Many creators use management agencies to run their inboxes. That isn't hidden:
when a group of subscribers sued over it, a US federal judge dismissed the
claims in December 2025 partly because OnlyFans' own terms already disclose it,
stating that creators may use third parties to manage their accounts and that
transactions run between fan and creator rather than through the platform. If
exclusive one-to-one contact is the point for you, that's a fact to weigh going
in, not to discover later.
The market spent a decade training people
to pick the app with the most users. The 2025 numbers are the first hard
evidence that this stopped working — and that the platforms doing well are the
ones that can finish the sentence "this is for people who want ___".
So finish it yourself first. Then compare
what's on offer against that answer: what the platform is designed to produce,
what the payment model is, what its rules say about third parties, and what the
free tier actually gives you. Review sites such as DudeHack
exist to lay those differences side by side, and the terms of service are
public in every case.
Match Group is already reworking Tinder —
cheaper tiers, in-person meet-ups. It may well recover. But the lesson of the
first down year is not that dating apps are finished. It's that the default one
no longer beats the right one.