We Counted 3,060 Conversations. 77% Died at the First Message.
Most messaging advice assumes the hard part is writing an opener good enough to earn a reply. Our own numbers say the hard part is earlier and blunter than that, and the conversations that survive separate themselves almost immediately.
Of 3,060 conversations that got an opening message, 2,358 never got a second one. Only 148 ever reached eleven messages. The cliff is not somewhere in the middle of a conversation; it is at the very first reply.
Every conversation on one platform, seed accounts excluded. The shape is what matters here, not the decimal places.
There is a whole genre of advice about keeping a conversation alive. How to avoid the dreaded lull, how to ask better questions, how to move to a date before it goes stale. We went looking for where conversations actually die on our own platform, expecting to find the lull.
We did not find the lull. We found a cliff, and it is at message one.
The distribution
Every conversation, bucketed by how many messages it ever contained:
Exactly one message: 2,358 conversations. Somebody wrote, nobody wrote back, that was the whole thing. This is 77% of every conversation that was ever opened.
Two or three messages: 344. A reply arrived and then it stopped, usually after the obligatory exchange of greetings.
Four to ten: 210.
Eleven to thirty: 129.
More than thirty: 19.
Add the last two together and you get 148 conversations, out of 3,060, that turned into something you would recognise as talking to someone. That is 4.8%.
There is also a quieter number underneath all of this. A further 316 conversations exist on our servers with no message in them at all. Two people matched, the thread was created, and nobody ever typed anything into it. We had not expected that to be a category.
Why the cliff is where it is
If conversations mostly died of boredom around message six, the advice industry would be right and the fix would be conversational. That is not the shape. Once somebody answers once, the odds change enormously: of the 702 conversations that got a single reply, 148 reached eleven messages. Getting a reply at all is roughly a one in four proposition. Getting somewhere after a reply is closer to one in five.
So the first reply is not one step of many. It is the step. Everything downstream of it is comparatively easy, which is the opposite of how people allocate their effort.
This also fits what we found when we timed every reply on the platform: 82% arrive within a day and 41% within the hour. Answering is a same-session behaviour. Someone opens the app, reads a queue, and either engages or does not. A message that does not get answered in that session mostly never does.
Who opens, and who keeps going
Men opened 1,780 of those conversations and women opened 1,280. That is a 58 to 42 split, which is a far more even distribution of first moves than the category's reputation would suggest. The idea that women never message first is not true here; they are doing about 42% of the opening.
The volume, though, is wildly asymmetric once conversations are running. 65 men sent 3,771 messages between them, an average of 58 each. 276 women sent 3,526 messages, an average of 12.8 each.
Read those two sentences together, because the combination is the interesting part. Four times as many women are sending messages, and they are each sending about a fifth as many. A small number of men are doing an enormous amount of talking and a large number of women are each doing a little.
The practical consequence is that a woman's inbox and a man's inbox are completely different objects. Hers is a queue to be triaged. His is a set of conversations he is actively working. Advice written for one is frequently nonsense for the other, which is worth remembering the next time you read a universal rule about messaging.
What actually moves the first reply
We have gone at this from another angle before, reading 482 opening messages to see which ones got answered. The two findings that held up are both unglamorous.
Photos matter more than words. Members with three or more photos got replies at 11% against 7% for one or two, and nobody with no photos has ever had a reply. Before any question of what you wrote, there is the question of whether there was enough there to answer.
And a written sentence beat a greeting: 15% against 6% for a one-word hello. Not a clever sentence. Not an opener with a hook in it. A sentence, addressed to a specific person, that shows you read something.
Neither of those is a technique. They are both closer to showing up properly, and both of them operate entirely before the cliff we have been describing.
What this should change
Spend your effort at the start, not in the middle. The energy people put into keeping a conversation going is being spent on the part that mostly takes care of itself. The conversation you are trying to rescue at message four was decided at message one.
Treat a non-reply as weather. At a 23% reply rate, three quarters of your messages will go unanswered no matter what you write. That is not a referendum on you. Reading it as one is how people end up sending a wounded follow-up to a stranger.
Volume with care beats agonising. The numbers argue for messaging more people and writing each of them one real sentence, rather than writing one perfect paragraph to one person. The first is a strategy that works with the distribution. The second is a strategy that fights it.
If someone answers, take it seriously. One reply moves you from a 23% world into a 21% chance of a real conversation from a much smaller pool. Whatever you were saving for later, that is the moment for it.
What we cannot see
Conversations that move off the platform. When two people swap numbers early, our record of it is a short thread that looks identical to one that died. Some unknown share of the 2-3 message bucket is not failure at all, it is success leaving the building. We know it happens in roughly 10% of conversations because people mention it, and it is occasionally a warning sign rather than a good one.
We also cannot see intent. A thread that ran to forty messages and went nowhere counts here as a thriving conversation, and a four-message exchange that ended in a dinner counts as a near miss. Message count is a proxy for engagement, not for whether anybody was happy.
And this is one platform, at a size where a few hundred conversations can move a percentage point. The cliff at message one is far too large to be noise. The finer splits deserve more caution.
Frequently asked questions
What percentage of dating app messages get a reply?
On our platform, about 23% of opened conversations ever received a second message. Measured per opening message rather than per conversation, the rate we found in a separate sample was around 9%.
Why do conversations die after a few messages?
Mostly they do not, because most never start. Only 344 of 3,060 conversations ended in the two to three message range. The dominant failure is a first message that is never answered, not a conversation that fizzles.
Should I send a follow-up if nobody replies?
Our timing data does not support it. Since 82% of replies arrive within a day, a follow-up sent later is arriving to someone who has already decided rather than someone who forgot.
Do women message first on dating apps?
Here, yes, often. Women opened 1,280 of 3,060 conversations, about 42%. They send fewer messages each than men do, but they are not waiting to be approached.
How many messages is a good sign?
Eleven or more puts a conversation in the top 4.8%. In practice, one genuine reply is the meaningful threshold, because the odds improve sharply after it.
Does a longer opening message work better?
A single written sentence outperformed a one-word greeting by roughly two and a half times. Beyond a sentence we have no evidence that length helps, and a paragraph to a stranger has its own problems.
Ready to get started?
Are you a successful professional? See why Arranged is worth it
Join the community → Free dating advice from real members