When is the best time to send cold emails?

The default worth using

Choose two plausible windows within the recipient's working hours, such as Tuesday morning and Thursday afternoon. Account for local time and holidays. The starting window is an operating hypothesis; industry charts cannot determine the best time for a particular list and message.

Plan the cadence before the test. If comparing sequences, keep follow-up count and spacing consistent between groups so a timing comparison does not also become a frequency comparison.

Where the hour-by-hour charts go wrong

Hour-by-hour charts can help generate hypotheses, but their sample, time zones, outcome definition and campaign mix affect interpretation. A difference between adjacent windows may be real or sampling variation; it needs a comparable test before being treated as a rule.

When prospects actually reply: 4,375 human replies, by weekday and hour

We read the arrival time of every human reply to campaigns in six ReplyLead-operated sending workspaces between April and September 2026 - 4,375 replies across 115 campaigns, automatic replies excluded - and grouped them by weekday and hour in US Eastern time. The busiest day was Monday, with 18.8%; Tuesday to Thursday held 61.6% of weekday replies, only slightly above the 60% an even Monday-to-Friday spread would give, so these arrivals show at most a small mid-week lean; Friday fell to 13.0%; 17.3% arrived on a Saturday or Sunday; and 78.2% arrived between 08:00 and 17:59 Eastern, with the busiest hour at 14:00. Of the 554 replies labelled interested, 20.0% arrived at weekends, against 17.3% of all human replies - a small sample, noted rather than relied on.

Read this for what it is. Arrival times partly mirror when our campaigns were scheduled to send, recipient time zones are not known per reply, and there is no random allocation here - so it is consistent with the working-hours part of the hypothesis above without testing it, gives the mid-week part little support, and proves no best send time. Automatic replies, which fire when a message is delivered, show when the campaigns sent: 99.9% of them arrived between 06:00 and 19:59 Eastern and 24.3% at weekends. So the working-hours share largely reflects the sending schedule, and part of Monday's lead may be replies to weekend sends. The test at the end of this page is how you find yours.

Volume settings and send windows answer different questions

ReplyLead's stated settings are 10-12 cold emails per mailbox per day on Google Workspace and 1-2 on Microsoft 365. These are operating choices, not provider-approved inbox thresholds. Hold volume consistent during a timing test and review delivery responses if it changes. See campaign capacity planning.

Check readiness before comparing schedules

Our stated preparation period is 2-2.5 weeks; elapsed time alone does not prove readiness. Verify authentication and review recipient expectations, provider responses and reputation before increasing volume. The warmup guide describes the process; it cannot guarantee placement.

The thresholds a good send time cannot buy back

Check the following independently of the send-time experiment:

• Complaint monitoring. Use the receiving provider's guidance and dashboard definition. Do not substitute bounces or unsubscribes for spam complaints. Provider scope and thresholds.
• Google counts roughly 5,000 messages in 24 hours to personal Gmail accounts, aggregated by primary domain; assigned bulk status persists. Yahoo does not publish a fixed volume threshold. Both have requirements for smaller senders as well. Google scope and classification; Yahoo classification.
• Bounce investigation. Our 2% internal review threshold is not a universal safe-sending rule. Investigate a new spike and distinguish invalid addresses from policy refusals.
• SPF evaluation. More than 10 evaluated DNS-querying terms returns SPF permerror; DKIM may still pass. The DNS checker helps inspect the configuration. SPF lookup rules.

A dead address ignores your schedule

An old validation flag does not establish that an address is currently valid. Review data age and scope, recheck where appropriate and honour suppression requests. No schedule can deliver to a nonexistent mailbox; a valid address can still reject mail for other reasons.

How to find your own window

Randomly allocate comparable recipients to two predefined local-time windows. Keep the message, audience criteria, volume, sender mix and sequence comparable; do not send both variants to the same person. Define positive replies per contacted lead and, separately, booked and held meetings. Use an equal observation window that accommodates reply delay. Record counts as well as rates; small samples may be inconclusive. Repeat promising results before changing the default. Outcome definitions.

When this page does not apply

• Your buyers are in other time zones. The reply times are read in US Eastern time, and recipients' own time zones are not recorded per reply; shift them to your recipients' zones before using them.
• You need a proven best send time. No controlled test sits behind these figures; they are arrival times that partly mirror send schedules. Run the two-window test on this page.
• You send consumer or newsletter email. The data is B2B cold outreach replies; opted-in marketing mail has different reading patterns.
• You are reading this months from now. The replies cover April to September 2026; the date range is stated in the JSON download.

How this page was built and checked

The reply-timing figures count campaign replies (replies tied to a ReplyLead campaign) received between 1 April and 24 September 2026 in six ReplyLead-operated sending workspaces (not every workspace ReplyLead operates), read from the sending platform's reply records (metadata only: no message content was read or stored); automatic replies are excluded from the human-reply counts. Arrival times are converted to US Eastern (UTC-4 throughout the window). The rest of the page is ReplyLead's operating guidance; the volume settings and thresholds are cited to their providers below. No random allocation of send times was run for this page. Last checked 24 September 2026.

Sources and check dates

ReplyLead's own data and pages behind this page:

1. Cold email reply timing 2026 (open data, CC BY 4.0): the weekday and hour counts behind the figures (JSON version alongside).
2. Cold email benchmarks 2026: the reply-rate evidence from the same campaign book.
3. Cold email statistics: 36 cold email numbers with their denominators.
4. Cold emails per day: the per-mailbox volume settings.

Outside primary sources, each read on the date shown:

1. Google: Email sender guidelines: sender requirements, spam-rate thresholds measured in Postmaster Tools. checked 24 September 2026.
2. Google: Email sender guidelines FAQ: how bulk-sender status is assigned: about 5,000 messages in 24 hours to personal Gmail accounts. checked 24 September 2026.
3. Yahoo Sender Hub: best practices: Yahoo's sender requirements. checked 24 September 2026.
4. Yahoo Sender Hub: FAQs: Yahoo's bulk-sender classification, with no published volume threshold. checked 24 September 2026.
5. RFC 7208, Sender Policy Framework, section 4.6.4 (IETF): the 10-lookup limit that returns SPF permerror. checked 24 September 2026.
6. RFC 5322, Internet Message Format, section 3.3 (IETF): the date-time format of a message's Date field. checked 24 September 2026.