ResearchAnalysis

Should a Cold LinkedIn Message End With a Question? Mostly, and Not With a Meeting Ask

Our analysis of 355 distinct first messages sent to 58,207 accepted LinkedIn connections found that messages ending in a question other than a meeting ask drew an 18.54% reply rate, against 15.53% for messages ending in a meeting ask and 14.42% for messages ending in a statement.

Dataset. Reply to the first message after acceptance by how its closing sentence ends, across 58,207 accepted LinkedIn connections and 355 distinct first-message texts, from requests sent 8 January 2025 to 16 October 2026, pulled 16 November 2026.

By Jordan Kwan · Published Nov 16, 2026

TL;DR: First LinkedIn messages that ended in a question drew more replies than ones that ended in a statement, but a question asking for a meeting did only slightly better than a statement. The exception: at 75 to 99 words, statements beat questions.

Our analysis of 355 distinct first messages sent to 58,207 accepted LinkedIn connections found that messages ending in a question other than a meeting ask drew an 18.54% reply rate, against 15.53% for messages ending in a meeting ask and 14.42% for messages ending in a statement. Our analysis

Key Takeaways

  • First messages ending in a non-meeting question drew 18.54% replies (19.17% campaign-weighted), the highest of the three endings.
  • Ending with a meeting ask drew 15.53% (15.59% weighted), only 1.11 points above ending with a statement (derived).
  • Ending with a statement drew 14.42% (13.84% weighted), the lowest.
  • A question beat a statement in every length band under 75 words, with the widest gap in messages under 25 words.
  • At 75-99 words the pattern reversed: questions drew 8.32% and statements 13.68%.

What did we measure?

"Always end with a question" and "always end with a clear call to action" are both standard cold outreach advice, and they often point to the same closing line: a request for a call. This study separates the two and measures them on the first direct message a person receives after accepting a LinkedIn connection request.

Reachium's outreach sequences store each message's text, so the closing line of every first message can be classified and compared against reply rates, across 355 distinct first-message texts.

How was it measured?

Dataset window. Connection requests sent between 8 January 2025 and 16 October 2026, pulled 16 November 2026, matured at least 30 days. InMail sends are excluded.

The first DM, denominator and outcome. The first message node inside the "connection accepted" branch of a sequence, read as stored. The denominator is matured, accepted requests that were sent it. "Reply to the first DM" means the person replied while that message was still the last one sent; sequences stop on a reply, so the reply is attributed to the first message rather than a follow-up. Our analysis

Exclusions. Accepted requests whose first DM was never sent (10,402); requests whose copy had been edited after sending (55); campaigns whose first DM carried more than one variant text; and 63 requests whose first DM would have rendered a blank company or title token. Clean sample: 58,207 people, 355 distinct messages, overall reply to the first DM 17.05%. Our analysis

The closing line. The last sentence before any sign-off. Trailing short lines without a question mark, such as "Best," or a name, are dropped. The classifier reads that closing sentence only.

Three endings. A statement (no question mark). A meeting ask: a closing question that mentions a call, chat, meeting, Zoom, calendar, booking, a number of minutes (15, 20 or 30), this or next week, hopping on, catching up, a demo or a walkthrough. Any other question. Both question types end in "?". The other-question group includes qualifying questions and opinion questions. As illustrations written for this page (not taken from the data): a closing like "Worth a quick chat next week?" counts as a meeting ask; a closing like "How are you handling this today?" counts as another question; a closing like "Happy to share more if useful." counts as a statement.

The effective sample is messages, not people. Every row reports distinct messages and a campaign-weighted reply rate (campaigns with 50 or more recipients, each weighted equally). A cell with fewer than 10 distinct messages is labelled too few messages to read.

What did we find?

How the first DM endsPeople who got the first DMReply to first DMReplied at any pointDistinct messagesCampaign-weighted replyLargest workspace share
Ends with a statement12,38814.42%23.84%8413.84%34.59%
Ends with another question33,94218.54%27.55%20319.17%14.62%
Ends with a meeting ask (question)11,87715.53%25.10%6815.59%19.52%
Reply rate to the first LinkedIn message, by how it ends
Another question18.54%
Meeting ask15.53%
Statement14.42%

Questions beat statements, and the meeting ask is most of the way back to a statement. Non-meeting questions led statements by 4.12 points pooled and 5.33 campaign-weighted (derived). The meeting ask led statements by 1.11 points pooled and 1.75 weighted (derived). Combining both question types, messages ending in any question drew 17.76% (derived), across 271 distinct messages. Our analysis

Within length bands. Length is a confounder: the length study on the same people found replies falling as messages got longer. So the comparison is repeated inside each length band. Here "ends with a question" includes both question types.

First DM length (words)Ends with a question: people / reply / messagesEnds with a statement: people / reply / messagesGap, question minus statement (derived)
Under 252,378 / 18.42% / 22717 / 11.85% / 11+6.57 pts
25-4924,668 / 20.56% / 1433,551 / 16.59% / 38+3.97 pts
50-7416,228 / 14.63% / 823,719 / 13.77% / 20+0.86 pts
75-992,188 / 8.32% / 193,984 / 13.68% / 13-5.36 pts
100-149357 / 19.89% / 559 / 3.39% / 1too few messages to read

A question beat a statement in the under-25, 25-49 and 50-74 bands, with the advantage shrinking as messages lengthened. At 75-99 words it reversed: statements drew 13.68% and questions 8.32%, on 13 and 19 distinct messages. The 100-149 band has too few messages to read. Our analysis

What does this mean for the best CTA in LinkedIn outreach?

On this data, the useful distinction is not question against statement. It is what the question asks. A closing question that did not ask for a meeting drew the most replies. A closing question that asks for time on a calendar did only a little better than ending on a statement.

A plausible reading is that a meeting ask is a bigger commitment to answer in a first message, while an opinion or qualifying question is cheap to answer. The data is consistent with that and does not prove it. Copy that ends in a meeting ask may also differ in what comes before it.

The 75-99 reversal is the part most likely to be misread. It is a small set of messages (19 questions, 13 statements), and the 75-99 band is also the most concentrated readable length band in the companion study (41.12% one workspace). It says the question rule did not hold for these longer messages; it does not establish that long messages should end in statements. The shrinking gap from under-25 to 50-74 is the steadier pattern: the question ending mattered most in short messages.

None of this changes what happens next. Follow-ups still add replies, replies arrive fast after acceptance, and when people reply is measured separately. The companion study on personalisation tokens covers the opening of the same messages. Teams reworking sequence copy can compare setups on Reachium's pricing page.

How do you use this in practice?

An SDR whose first message closes by asking for 15 minutes next week is in the meeting-ask group, which replied at 15.53% here. Messages that closed with a question that did not ask for a meeting drew 18.54%. The data does not say rewriting that close would move the SDR's rate by the difference, only that the two groups differed by 3.01 points (derived). Our analysis

A founder who writes short first messages and ends them with a statement is in the under-25 statement cell: 11.85%, against 18.42% for under-25 messages ending in a question, the widest gap in the table. Our analysis

A team lead with longer templates, 75 to 99 words, is in the one band where statements did better (13.68% against 8.32%). With 13 and 19 distinct messages behind those figures, it is a pattern to watch in their own data rather than a rule. Our analysis

The Savvy Capital case study shows one team's outreach numbers on the same platform.

What are the limits of this study?

Observational, not an experiment. Workspaces chose their own copy. Nobody was assigned an ending, so targeting, offer and everything above the closing line ride along. A/B variants are excluded, not analysed.

The classifier reads one sentence. Only the closing sentence is classified. A message with a meeting ask in the middle and an opinion question at the end counts as another question.

Keyword-based meeting detection. Meeting asks are identified by the terms listed in the methodology. A meeting ask phrased without any of them is counted as another question.

Small cells. The within-length comparison has cells as small as 11 and 13 distinct messages, and 100-149 has too few messages to read.

Concentration. The statement group's largest workspace sent 34.59% of it; the other-question group is the most spread at 14.62%.

Reply is not a booked meeting. A meeting ask could produce fewer replies but more meetings; this study measures replies only, not positive replies or bookings. Teams that want help turning replies into meetings can talk to the agency team.

FAQ

Should a cold LinkedIn message end with a question?

Usually, on this data. First messages ending in a question drew 17.76% replies (derived) against 14.42% for statements, though statements did better in 75-99 word messages.

What is the best CTA for LinkedIn outreach?

In the first message after acceptance, a question that is not a meeting ask drew the most replies: 18.54%, against 15.53% for a meeting ask.

Does asking for a meeting in the first LinkedIn message hurt replies?

It did not hurt against ending with a statement (15.53% against 14.42%), but it was 3.01 points lower than ending with another kind of question (derived).

Does message length change which ending works best?

Yes. The question advantage was largest under 25 words (+6.57 points), shrank through 50-74 words, and reversed at 75-99 words (-5.36 points, derived).

Sources

How to cite this study

Reachium (2026). Should a Cold LinkedIn Message End With a Question? Mostly, and Not With a Meeting Ask. Reply to the first message after acceptance by how its closing sentence ends, across 58,207 accepted LinkedIn connections and 355 distinct first-message texts, from requests sent 8 January 2025 to 16 October 2026, pulled 16 November 2026. https://www.reachium.io/research/question-cta-linkedin-messages

Related research

Your competitors are automating.
You're still copy pasting.

Try Reachium

Live in 10 minutes.