Online Dating Advice: Optimum Message Length

September 3rd, 2009 by Christian Rudder

Picture this online dating scenario:

  1. You see someone you like.
  2. You read their profile, and wow.
  3. You send them a long message.
  4. You hang tight and…
  5. …you never get a reply.

Sadly, this is a typical story. Even on a lively site like OkCupid, only about a third (32%) of first messages get any response.

Some people, however, get much better results.

In the next several posts here on OkTrends, we’ll be taking a closer look at messaging and finding some ways to improve your own message response rate. We will not be dispensing generic advice. No. We’ve done research, and we have actual numbers.

. . .

As we began to dig into OkCupid’s messaging data, the first thing we noticed was that most people’s contact attempts are way too lonnnng. Almost 16% of first messages are over 2000 characters (roughly 400 words), and the average is 743! At least in terms of using your time efficiently, your messages should be much shorter. Let’s start with this chart:

The y-axis is reply percentage; the x-axis is message length, in characters; and the two lines are as follows. Red is the ratio of messages that get any reply. Green is the ratio of messages that get a reply that in turn gets replied to by the original sender. The idea is that this is the ultimate goal of the first message: to start a conversation with someone.

Messages sent by guys are, overall, only about half as likely to get replies as similar messages from women. But when you consider we’re including dudes who send out messages such as:

Your hot

DAm I got inch for you


Your people need to get out of Israel

a baseline reply rate of 22% is looking pretty darn great. (All those were actual first messages, by the way.)

Now, our graph clearly shows that in raw terms, it helps guys to write longer messages. But when we factor in the actual time it takes to compose a given message, it becomes clear that in terms of time put in vs. likelihood of starting actually having a conversation, shorter is actually better. If we imagine that people type messages at about 200 characters per minute, we get the following table:

Of course, we shouldn’t forget that there’s a certain amount of overhead involved with contacting someone (scanning her profile for common interests, thinking of jokes to make, taking a deep breath, clicking around, and so on). If we include 5 minutes of forethought, we find that the actual ideal first message length is 200 characters, or 1 minute’s worth of typing for the average writer.

Yes, brevity is key. Something we learned building SparkNotes, in our pre-OkCupid days. If you’re the kind of person who spends a little more time reading a profile and thinking about your message, say, 10 minutes, then the optimal length goes up a few words (to 270 characters), but, still, short is better. Savor this advice, men, for there are not many things in your world that fit this paradigm.

. . .

For women, the most efficient message is even shorter.

The shortest messages get almost the best absolute response rate, and the reply rate actually goes down as messages approach extreme length. Apparently, after about 360 words (1800 characters), you start scaring people off. A message like that is the online equivalent of a face tattoo. Of your life story. Let’s generate our efficiency table for women:

Incredibly enough, the optimal first outreach from a woman to a man is just 50 characters long! I’m willing to speculate that this graph is telling us that a guy decides whether or not to reply to a woman’s message regardless of what the message actually says, and that the first message’s true function is simply to bring her profile to his attention.

My guess is that he looks at her picture and if she’s his type, he writes back. On the one hand, such a superficial reality is depressing. On the other, over 40% of female-to-male first messages do get replied to, so, as a woman, if you’re writing to a few people who fit your basic demographic the odds are very good you’ll get a response. Anyhow, all this implies that the average woman’s time is better spent looking for the right people to write to, rather than composing detailed messages.

To our bi and gay readers: we also ran the numbers on same-sex messaging, and though we don’t have nearly as much data for them, those charts are here, along with some discussion.

A quick note about privacy on OkCupid

Though this post talks in detail about the content of people’s messages on OkCupid, all messages have been thoroughly anonymized, with sender and recipient data stripped out. In addition, our sifting program looks at the content of messages only two or three words at a time, to track the success of certain phrases (like “what’s up” vs. “wats up”), then aggregates results by phrase. No human has read any actual user messages. The longer messages I give as examples in this post were actually forwarded to us by their annoyed/amused recipients.

This post’s line charts are smoothed with a bezier spline. Message length is rounded up to the nearest 50 (or we wouldn’t have significant data on some points.) So the first data point on the graph is messages of length 0–50, next is 51–100, and so on.