The first message problem.

The opening message in a dating app conversation is functionally a cold call to a stranger who has indicated minimal interest — the swipe — but not warmth, not curiosity, not intention to respond. The match is the minimum signal: it means only that this person was willing to indicate they would accept a conversation with you if you initiated one. Whether they respond when you do is a separate question with a separate answer, and the answer is not primarily about the quality of your writing.

Understanding what produces responses requires treating the first message as what it is: an asymmetric communication problem operating inside a platform architecture that was designed to maximize engagement, not to facilitate connection.

The baseline response rate

OKCupid’s internally published blog data — released publicly between 2010 and 2014 under the editorial project OKTrends — documented messaging behavior across millions of interactions. Some of the findings were counterintuitive enough to attract significant media attention at the time, but they were quickly buried under the subsequent news cycle and are rarely cited now.

The finding that received the most attention: the physical attractiveness of the recipient is the primary determinant of response rate, not the content of the message. Messages sent to recipients rated in the top attractiveness tier receive responses at dramatically higher rates than messages sent to recipients in lower tiers, regardless of message quality. The complementary finding: messages sent to recipients whose attractiveness rating significantly exceeds the sender’s receive very low response rates regardless of effort.

The content finding: message length does not reliably predict response rate in the way that most people assume. Longer, more effortful messages do not consistently produce higher response rates than shorter messages. In some analyses, very long messages to recipients in high attractiveness tiers produce lower response rates than shorter ones — a finding that requires a structural explanation, not a stylistic one.

What Bruch and Newman found

Elizabeth Bruch and M.E.J. Newman’s 2018 Science Advances paper is the most methodologically rigorous academic study of messaging behavior on a dating platform to date. They analyzed a dataset of 186,000 users on a major dating platform across four U.S. cities: New York, Boston, Chicago, and Seattle.

The core finding on messaging behavior: most users message significantly above their own desirability rank. Desirability in the study was defined recursively — a measure derived from how many messages a user received and from the desirability of the people who sent those messages, analogous to PageRank. The result: the average user’s opening messages go to recipients substantially higher in the desirability ranking than the sender.

This is aspirational messaging, and it is nearly universal. It is not pathological — the researchers describe it as a rational strategy in an uncertain environment — but it has a predictable consequence: the response rate is low. Most messages to users significantly above the sender’s desirability rank go unanswered.

Bruch and Newman also documented escalation behavior: users who receive no response often send increasingly effortful follow-up messages to the same recipient. The analysis showed no evidence that increased effort improves response rates in these conditions. The mechanism is not message quality.

The question framework

Where message content does appear to matter, the evidence points toward a specific factor: evidence of genuine attention to the recipient’s profile. Messages that reference something specific in the profile — a photograph, a stated interest, a piece of profile text — produce higher response rates than generic opening lines. The analysis on this is consistent across multiple data sources, including OKCupid’s published data and subsequent academic work.

The mechanism is legible: a message that demonstrates the sender actually looked at the profile signals selectivity. It implies that the sender chose this specific person rather than broadcasting the same message to many people. In a platform context where mass-messaging is common, selectivity is itself a signal that carries value.

The type of reference matters. A functional question about something in the profile — specific, open-ended, not creepy — outperforms generic conversation openers. “How was your weekend?” is generic. A question about the location in a photograph, or about a book mentioned in the bio, or about an activity shown in an image — these are specific. The specificity does the work, not the cleverness.

The question format outperforms statements as openers, generally. A statement requires no response. A question creates a mild, low-pressure invitation. The combination of specificity and question format appears in multiple analyses as the single most reliable predictor of response among message-content variables. This is not a large effect — the primary variance is still explained by desirability gap — but it is the content variable most consistently supported by evidence.

What this means structurally

The first message problem is not primarily a writing problem. It is a targeting and attention-signaling problem. The content of the message accounts for a small share of response rate variance. The primary predictors are:

One: the relative desirability positioning of sender and recipient. Messages to recipients significantly above the sender in the platform’s implicit desirability hierarchy are unlikely to receive responses regardless of content.

Two: evidence of genuine attention to the recipient’s profile — the specificity signal. This is within the sender’s control and reliably matters.

Three: the absence of disqualifying content. Opening with sexual content, with generic copy-paste text, or with content that signals the sender has not looked at the profile are all reliable suppressors of response rate.

What is not consistently supported: the idea that wit, humor, or sophistication of writing reliably drives response rates above the baseline set by desirability positioning. This is a finding that most people find uncomfortable because it reduces the significance of effort, and because it implies that the primary variable is something largely outside the sender’s control.

On effort and reciprocity

The counterintuitive finding about message length deserves a structural explanation. In a cold-call context, a long, high-effort message creates an obligation asymmetry before any connection has been established. The recipient has provided a minimal signal of openness — the match — not an indication of interest sufficient to warrant an extended personal communication. A long message implicitly asks for a substantial response. A short message asks for a brief reply. The asymmetric obligation is a deterrent.

The reciprocity principle in social psychology (Cialdini 2001) applies here in an inverted way: when an overture creates an obligation disproportionate to the stage of the relationship, it produces avoidance rather than reciprocal engagement. The instinct to demonstrate effort and investment in an opening message is understandable, but it is not well-matched to the structure of the interaction.

Short, specific, low-pressure messages that demonstrate genuine attention to the profile — and that ask an open-ended question — are what the data supports. This is not a formula; the specificity requirement means it requires actual engagement with each profile rather than template deployment. That engagement is the actual effort the data rewards.