There is no universally best time to send an email. The right window depends on your audience, the campaign objective, the time zone used for scheduling and analysis, and the evidence in your own account. A weekday morning can be a reasonable starting hypothesis for a newsletter with little history. An event reminder, deadline-driven offer, or time-sensitive message should be scheduled around its business constraint instead.
HubSpot’s editorial article, The Best Time to Send an Email [2023 Research], updated in July 2026, reports that respondents most often selected 9 AM to 12 PM and 12 PM to 3 PM as high-engagement periods. Tuesday led the reported day preferences, followed by Monday and Thursday. These are survey responses from the article, not platform-measured results for your list. The article does not publish enough methodology to reproduce the findings, so use them to define candidate test windows rather than as a scheduling rule.
A practical decision order is simple: meet deadlines and consent requirements first, use reliable audience history next, choose a broad candidate window when history is sparse, and judge the result against the campaign’s primary business outcome.
Choose the campaign outcome first. Use general benchmarks to create test windows, then use your audience’s evidence to choose a schedule.
The best time to send an email depends on the outcome you want
Business type alone is not enough to select a send time. Recipient role and seniority, geography, buying cycle, campaign purpose, and local time can all change the useful window. A B2B newsletter, retail promotion, event reminder, and customer follow-up may each need a different rule.
HubSpot reports that 27% of surveyed U.S. marketers selected Tuesday as the day with the most engagement, followed by Monday at 19% and Thursday at 17%. It also reports that 47.9% of surveyed B2B marketers and 30.9% of surveyed B2C marketers selected 9 AM to 12 PM. The article does not provide subgroup sizes or enough methodology to establish these figures as representative industry benchmarks. Treat them as reported preferences that can help you choose a small set of plausible slots.
The article also reports lower preference for weekend sends, particularly Sundays. That is a survey-based suggestion, not a deliverability rule. A weekend send may be appropriate for a retail promotion, event, or audience that is more active outside business hours. Test it when the campaign makes it plausible, and account for any fixed deadline.
Choose the metric before choosing the send time
Define what “better” means before comparing results. An open-rate winner may not produce more clicks, replies, qualified leads, or revenue. Keep the metric and denominator fixed across every time slot or test arm.
| Campaign objective | Primary measure | Denominator or scope |
|---|---|---|
| Drive site visits | Click rate or click-through rate | Recipients who received the email, or recipients who opened it |
| Generate replies | Replies per delivered recipient | Recipients who received the email |
| Generate leads or sales | Conversions or revenue | Defined eligible recipient base and attribution window |
| Assess subject-line attention | Open rate | Recipients who received the email |
Current HubSpot documentation distinguishes click rate from click-through rate. Click rate uses recipients who clicked divided by recipients who received the email. Click-through rate uses recipients who clicked divided by recipients who opened the email. Do not compare one measure from one report with the other from a second report and call the difference a send-time effect.
HubSpot defines open rate as recipients who opened divided by recipients who received the email. An open may be recorded when images load or when a recipient clicks a link, and tracking settings and email-client behavior can affect measurement. Treat opens as one signal. For revenue or conversion campaigns, review delivered counts, bounces, unsubscribes, spam reports, and the business result together.
Write a short metric contract before analysis:
primary_metricdenominator_definitionobservation_window_hoursattribution_window_daysexcluded_eventstime_zone_rule
The time-zone rule should state whether results use recipient-local time when available, the account time zone, or another campaign time zone. HubSpot’s email performance documentation describes current metric definitions and analysis options. Its export documentation covers overall performance and recipient-level event exports, with timestamps for individual events available in advanced exports on eligible plans. These are documented reporting and export paths, not a promise of a public API, fixed export schema, or real-time event feed.
Use benchmarks to choose test windows, not to declare a winner
Use the HubSpot survey as a starting point, not as proof that Tuesday or a particular hour will outperform for your audience. The article says its respondents most often selected morning and early-afternoon windows and reports Tuesday as the most frequently selected day. It also gives Eastern Time examples for some day-and-time preferences. Those examples should not be treated as recipient-local recommendations for a geographically mixed list.
If you have little audience history, choose a small number of plausible windows, such as weekday morning and early afternoon, and compare them under similar campaign conditions. Testing every day-and-hour combination at once spreads the audience too thin and makes campaign differences harder to control.
If reliable history exists, segment it by audience and campaign type before selecting a time. Compare similar newsletters with similar newsletters, or similar promotions with similar promotions. Check whether the audience mix, offer, subject line, seasonality, send volume, tracking settings, or deliverability changed between the periods being compared.
Analyze historical performance by hour of day
HubSpot documents a custom marketing email report example that groups delivered date by hour of day and includes delivered rate, open rate, click-through rate, and reply rate. This is an aggregated reporting configuration. It can identify candidate hours, but it cannot by itself show that the hour caused the result or that every recipient has the same ideal time.
A practical analysis has separate layers:
- Raw event layer: individual sends and recipient events, including repeated opens or clicks when the source provides them.
- Normalized layer: validated identifiers, event types, timestamps, campaign context, and the declared time-zone rule.
- Aggregate layer: results grouped by campaign, segment, reporting date and hour, time zone, metric definition, and experiment arm.
- Decision layer: candidate windows, sample status, rationale, owner, and review date. This layer must not overwrite the observations.
The following is an illustrative analytical record, not a HubSpot export schema. It represents one recipient click occurrence, so the event-level identity includes a source event identifier and occurrence sequence:
{
"account_id": "illustrative_account_3",
"email_id": "illustrative_email_42",
"send_instance_id": "illustrative_send_2026_10_01_a",
"recipient_id": "illustrative_contact_17",
"event_type": "click",
"event_timestamp": "2026-10-01T14:23:00Z",
"event_sequence": 1,
"source_event_id": "illustrative_source_event_771",
"campaign_id": "illustrative_campaign_8",
"segment_id": "illustrative_segment_b2b",
"time_zone_rule": "recipient_local_if_available",
"experiment_arm": "morning",
"source_record_id": "illustrative_export_row_108",
"import_batch_id": "illustrative_batch_12"
}Be precise about row grain:
- Send or campaign row: one email send or campaign version, identified by an appropriate account, email, and send-instance key.
- Recipient-event row: one event occurrence for a recipient and email. A recipient-plus-email key alone can incorrectly collapse repeated opens or clicks.
- Link-click row: one click occurrence or defined click aggregate that retains link identity as well as recipient and email context.
- Aggregate row: one result for a defined campaign, segment, reporting date and hour, time zone, metric definition, and experiment arm.
For concurrent imports, enforce uniqueness at the actual source-event grain with a database-enforced unique constraint. Use an atomic insert-or-update or upsert. A check-then-insert sequence is not safe when two workers process the same event at the same time. This is a recommended implementation design, not a HubSpot-published synchronization method.
HubSpot does not prescribe a universal minimum sample size for selecting a send-time winner. Establish a fit-for-purpose rule before reviewing the results. A small count, changed tracking settings, or a major campaign-mix difference may require more data or a qualified recommendation rather than a schedule change.
Run a controlled test when history is inconclusive
When historical patterns do not settle the question, test a small number of candidate windows. HubSpot documents A/B testing that can vary send time, subject line, or content for Marketing Hub Professional and Enterprise. That confirms a product capability, not a universal statistical protocol or guaranteed lift.
- Choose one primary metric, denominator, observation window, and attribution rule before launch.
- Allocate comparable eligible recipients to each time slot.
- When testing time alone, keep the offer, content, subject line, audience criteria, and delivery conditions as consistent as possible.
- Apply the same exclusions and time-zone rule to every arm, and wait for the fixed observation window to end.
- Review the primary result alongside delivered counts, bounces, unsubscribes, and spam reports.
Keep each arm identifiable. Do not merge its results into the same hourly aggregate as another arm. If creative, offer, audience, or delivery conditions changed as well as the send time, report the result as an association rather than a causal timing effect.
| Trigger | AI or analysis job | Validation | Action and fallback |
|---|---|---|---|
| Comparable history is available | No AI required. Calculate fixed metrics by declared hour and segment. | Check event grain, time zone, sample status, tracking changes, and campaign comparability. | Use the strongest candidate as a test input or documented rule. Escalate mixed campaign types for segmentation. |
| Candidate windows are unclear | No AI required. Assign comparable recipients to a small controlled test. | Keep creative and offer consistent; freeze metric definitions and observation window. | Choose against the primary business metric and guardrails. Continue testing if allocation or delivery was compromised. |
| Individual engagement history is sufficient | Use HubSpot’s documented send-time optimization, not an invented custom model. | Confirm plan, 90-day engagement basis, selected range, and deadline tolerance. | Use the optimized schedule. Fall back to a deterministic window when history or timing conditions are unsuitable. |
A morning slot may lead on opens, an afternoon slot on clicks, and another slot on purchases while also producing more unsubscribes. Select the slot using the declared business objective and agreed guardrails, not the most flattering rate in isolation.
Use individual send-time optimization only when its conditions fit
HubSpot can suggest a send time based on selected recipients’ average best sending times. Marketing Hub Enterprise can optimize the send time for individual contacts. HubSpot says the calculation uses recent marketing-email opens and clicks from the prior 90 days. One-to-one emails and opens classified as bot activity are excluded. Suggested times display in the user’s time zone. Contacts without sufficient engagement data are sent at the beginning of the selected range, and individual sending can span up to 168 hours, or seven days.
The cited documentation does not disclose the complete algorithm, a universal minimum activity threshold, or a guaranteed performance lift. Check current plan availability before implementation and confirm that the campaign can tolerate delivery across the selected range.
A fixed schedule is usually more appropriate for urgent or deadline-based messages, small lists with sparse history, quiet-hour or legal constraints, and one-off events with no repeat data. In these cases, use a clear scheduling rule and a deterministic fallback rather than implying that sparse history supports individual personalization.
Use post-send workflows for follow-up, not initial timing decisions
HubSpot contact-based workflows can use marketing-email activities such as opens, clicks, bounces, and unsubscribes as enrollment triggers after the relevant email has been sent. That supports follow-up, not the initial send-time decision. Default enrollment is generally one time unless eligible re-enrollment is configured, and suppression or unenrollment rules can prevent or stop workflow participation.
Before writing a follow-up event to a CRM record, verify the contact, email or campaign identifier, event type, subscription status, suppression status, and event grain. A contact-level follow-up action is not the same as a raw recipient-event record. Keep the source event and processing batch available so a repeated import cannot create a second action accidentally.
See HubSpot’s documentation on workflow enrollment triggers and enrollment settings for the documented trigger and enrollment behavior.
Turn the analysis into a repeatable operating decision
For each recommendation, save the campaign and segment scope, reporting time zone, metric and denominator, analysis period, event exclusions, sample status, selected slot, rationale, owner, and review date. Keep raw events, aggregate reports, and scheduling recommendations as separate records. Do not overwrite the observations when the team adopts a new rule.
Marketing owns the campaign objective and comparability of the sends. Analytics or marketing operations owns metric definitions, event grain, and report validation. The platform owner checks plan availability, suppression, and scheduling behavior. For teams formalizing HubSpot reporting and scheduling processes, HubSpot systems consulting is an optional service resource.
- The campaign objective and primary metric are explicit.
- The denominator, observation window, attribution window, exclusions, and time-zone rule are fixed.
- The compared campaigns or test arms are sufficiently comparable.
- Raw events are preserved at the correct grain and duplicate processing is prevented with enforced uniqueness or an atomic upsert.
- Sample status, delivered counts, and negative signals have been reviewed.
- A named owner has approved the rule and set a review date.
Once those conditions are met, change the schedule for the defined audience and campaign type, then review the next comparable send. A useful send-time recommendation is a documented operating decision that can be checked and revised, not a permanent rule inferred from a general benchmark.