The median measured cold DM campaign on X gets a reply from about 1 person in 175, or 0.57%. That comes from the largest dataset, 43.7 million DMs sent through a rival tool. Our own users averaged 1.2% from 23 August to 22 September 2026. Tool blogs quote 3% to 40%.
We sell a cold DM tool, so a high number would suit us better. The measured figures are lower, and most of the gap comes down to what each source counts as a reply.
What the measured data shows
The only large public dataset comes from Xreacher, a cold DM tool that competes with ours. It counted 43,672,794 first DMs sent through its product in the 180 days to 3 July 2026, across 15,045 sending accounts, and any answer at all counts as a reply. Read the rows as different cuts of the same data, sorted from the highest rate down:
| Cut of the data | Reply rate |
|---|---|
| Top tenth of campaigns. | Above 3.7%. |
| Small teams on 2 or 3 accounts. | 2.51%. |
| One sender on one account. | 2.24%. |
| Median campaign, of 366 with 1,000 DMs or more. | 0.57%. |
| All DMs pooled. | 0.43%. |
| Operations on 11 or more accounts. | 0.39%. |
Big operations sent 97.5% of the DMs and got the lowest rate, 0.39%, so volume did not buy replies.
Our own figure sits between the median and the solo senders. All campaigns on our software averaged 1.2% from 23 August to 22 September 2026, with single scripts ranging from about 1% to 8%. Xreacher's window ended around the time X added its cap table, some time between 24 June and 4 August 2026, so we cannot say how much the caps change its numbers.
What tool blogs claim
Pages that rank for this question quote far higher numbers, and none of them shows a dataset behind its figure. We read them all on 5 October 2026. Read down the middle column, sorted from the highest claim, and compare each with the measured median:
| Who says it | Their number | How they label it |
|---|---|---|
| Autoreach. | 25 to 40% for personal DMs. | Benchmarks, with no dataset. |
| TweekSocial. | 8 to 20% "with real research". | "From campaigns we see". |
| XAutoDM. | 8 to 15% for a well run campaign. | A table set beside cold email. |
| Cold DM Calculator. | 7 to 11% as the middle range. | "Illustrative planning ranges". |
| Twittrz. | 3 to 8% is healthy. | "Planning figures, not benchmarks". |
Two of the five say plainly that their numbers are for planning, which is fair of them. Twittrz goes further and says a rate under 1% means "something is broken". By that line, the median measured campaign is broken.
Why the numbers disagree
Each source is answering a slightly different question, and five differences explain most of the spread:
- What counts as a reply. Xreacher counts any answer, "no thanks" included. A positive rate counts only people who want to talk, and it is always lower.
- How warm the list is. Autoreach's 25 to 40% describes personal DMs, which usually go to people who already know the sender.
- How big the list is. Twittrz says rates above 15% usually mean "a small hand-qualified list".
- Which average. Pooled, median and top tenth come from the same Xreacher data and differ about ninefold.
- Numbers borrowed from email. Cold email has large public studies and X does not, so some pages reason from email instead.
What is a good reply rate on X?
Above 0.57% beats the median measured campaign, and above 3.7% puts you in the top tenth. Look up your own rate here:
| Your rate | Where it sits |
|---|---|
| Under 0.57%. | Below the median. Check the list first. |
| 0.57% to 2.2%. | Above the median, below solo senders. |
| 2.2% to 3.7%. | Level with solo senders and small teams. |
| Above 3.7%. | The top tenth. |
If you sit near the bottom, the list is the first suspect. In our own outreach in September 2026, people who had commented on a "drop your startup link" post replied with interest 3.0% of the time, against 0.3% for followers of big marketing accounts. Which kinds of X lead lists answer DMs covers that side.
How to measure your own reply rate
Count people, not messages, and wait a week before you judge. In Xreacher's data, 22.6% of replies arrived more than 7 days after the first message. Your reply rate is the people who replied divided by the people who got a first message, and your positive rate counts only the people who want to talk.
Sample size matters more than most people expect. At 1%, 500 sends bring about 5 replies, so one extra answer moves your rate by 0.2 points, and a fair test needs 1,000 sends or more. The opening line is the next thing to test after the list, and our 12 cold DM scripts give you a starting point.
Every other X DM number, with its source and date, is on our X cold DM statistics page.
Xsender shows sent DMs, failed sends and reply rate for each account, so you do not have to work these out in a spreadsheet.
FAQ
Why is my cold DM reply rate so low?
Check the list before the message. In our own outreach in September 2026, changing the list source moved positive replies from 0.3% to 3.0%.
Do follow-ups raise the cold DM reply rate on X?
In Xreacher's data, a follow-up after 3 or more silent days raised the reply rate. But X lets only one message into a pending request, so a follow-up cannot reach a stranger until they accept.
Does the time of day change cold DM replies?
In Xreacher's data, DMs sent between 19:00 and 02:00 UTC got about 1.4% replies, against 0.86% between 09:00 and 12:00 UTC. That is a pattern in one vendor's data, and it does not prove the hour caused it.
Try it on your own X accounts. See plans and the 7-day trial
Sources and method
Sources and dates
- Measured rates, account counts, timing and follow-ups: Xreacher's Cold DM Benchmarks, published 3 July 2026, read 5 October 2026. Vendor data from its own product.
- Our 1.2%: all our users' campaigns, 23 August to 22 September 2026. The 1% to 8% range: our users' campaigns by script.
- The list comparison: our own outreach, September 2026, 26 positive replies of 854 against 5 of 1,804.
- One message per pending request: observed by the founder; our sending has followed it since 1 October 2026.
- Autoreach: its Twitter cold DM templates post, updated 3 June 2026.
- TweekSocial: its Twitter cold DM page. XAutoDM: its cold DM benchmarks post. Cold DM Calculator: its response rate benchmarks post. Twittrz: its good reply rate post.
- All blog pages were read on 5 October 2026.