How rare is a reply to a cold email, across the whole dataset

Before asking whether any particular cold-email tactic changes the reply rate, it is worth establishing what the reply rate looks like with no segmentation at all. This study tracked 287,790 cold emails sent between 2026-01-26 and 2026-08-14, sorted into 45 distinct sending groups, and classified each one into one of 2 outcomes: replied or no reply. No rows were excluded from this count (0 excluded) and no adjustments were made to the raw classification (0 mitigations applied).

The headline number is stark: 3,368 of 287,790 emails, 1.2%, received a reply. The remaining 284,422 emails, 98.8%, did not. This is the baseline against which every content-segment comparison in this study has to be read. When a later finding says one segment replied at a certain rate and another segment replied at a different rate, both numbers are small deviations around this same low floor, not swings between a high-performing tactic and a low-performing one.

This floor is not evenly distributed across sending groups. The 45 sending groups in this dataset averaged 6,395.33 emails each, and the largest single group accounted for 14.8% of all emails in the study, meaning no one group dominates the result but the groups are also not uniform in size. Looking at how the "replied" class varies group to group, across 33 groups the reply rate ranged from a low of 0% to a high of 4.5%, with a median of 0.8%. That is a wide spread. A sending group at the low end and a sending group at the high end are separated by several times the median rate, which means factors specific to the sending group, list quality, sender reputation, industry, time of year, likely explain more of the variation in outcomes than any single content choice does.

Why does this matter for reading segment comparisons? Because a segment's reply rate is a blend of whatever content effect exists plus whichever sending groups happened to contribute more emails to that segment. If, for instance, the percentage-stat segment happened to draw more heavily from sending groups that already sit toward the low end of the 0% to 4.5% range, some of its lower reply rate could be attributable to group composition rather than content. This study reports segment-level and group-level spread separately, but it does not report a group-by-segment cross-tabulation, so a reader cannot fully separate the two effects from what is published here. That is a real limit: someone who wants to isolate the content effect from the sender effect would need that cross-tabulation, or a controlled test within a single sending group.

What this baseline tells a team making a go or no-go call on cold email as a channel: the arithmetic of cold outreach means that even a highly effective message will produce a reply from a small minority of recipients, and the volume has to be large enough to make that minority meaningful. A team evaluating a new campaign against internal benchmarks should compare against this kind of low, single-digit-percent floor, not against reply rates typical of warm channels like referrals or inbound.