
Send-Time Testing: A Step-by-Step Protocol for Your List
Send-time optimisation is one of the most over-promised features in email marketing. Platforms present it as a reliable revenue lever, and the advice that circulates online often reduces to a single line such as "send on Tuesday at 10 a.m." Neither approach reflects how your subscribers actually behave. The averages in published studies describe many audiences at once, and your audience is not an average. It has its own time zones, devices, routines and relationship with your brand.
This guide sets out a testing protocol you can run on your own list. It explains when send time matters and when it does not, how to read your existing data, how to design a comparison that is fair, and what to do when the results are small or unclear. For the broader debate about whether send-time features earn their keep, see the send-time optimisation analysis.
Send-time optimisation works by matching sends to when your own subscribers engage. Test it with a clean comparison, decide on clicks or conversions rather than opens, and re-test each quarter because habits change with the seasons.
When send time matters, and when it does not
Send time matters most in three situations: when your subscribers are spread across time zones, when the content is time-sensitive, and when your list has a clear daily or weekly rhythm that you can see in the data. It matters least when the offer is strong, the content is relevant and the list is small. In those cases, a sensible send time is enough, and further tuning will produce little.
There is a helpful way to frame the question. Send time changes the moment an email arrives. It does not change whether the email is worth opening or acting on. If the content is weak, a perfectly timed send will still underperform. Fix relevance first, then work on timing, because timing is a multiplier on whatever is already working.
Read your own data before testing
Begin with at least ninety days of engagement history. Look at clicks and conversions by day and by hour rather than opens, and split the results by the factors that most often shape timing. Time zone is usually the most important, since a send at a fixed clock time can land in the middle of the night for a large share of subscribers. Device matters next, because mobile and desktop habits differ, and a morning send that works on a work desktop may be ignored on a phone in the evening. Segment matters too: new subscribers and long-standing ones often follow different patterns.
The aim of this step is not to pick a winner. It is to form a hypothesis. If the data suggests that your mobile readers click most on weekday evenings, that becomes the window you test. If the data shows no clear pattern, that is useful information too, and it tells you that timing is unlikely to be a large lever for your list.
Design a fair comparison
A send-time test is only informative when the comparison is fair. Send identical content to both groups, so that only the time changes. Split the audience randomly between similar groups, because giving the morning group your most loyal customers will beat the evening group for reasons unrelated to time. Where possible, send both versions on the same weekday, so that the day of the week does not confound the result. Choose the primary metric before launch, usually clicks or conversions, and decide how many recipients each group needs before the difference becomes distinguishable from noise.
Run the test long enough to see a pattern, which may mean several sends rather than one. A single send is a single observation, and a list with mixed habits can produce a misleading result on any one day.
Why opens are the wrong measure
Opens are unreliable for timing decisions. Privacy features such as Apple Mail Privacy Protection load images automatically, which means some recorded opens reflect when the email was delivered rather than when anyone read it. The result is that open times can cluster around delivery, regardless of when the reader actually engaged. For that reason, use clicks and conversions as the decision metric. They require a deliberate action, and they tell you what the reader did, not what their device loaded.
Interpreting the results
Interpreting a send-time test is largely a question of proportion. A clear difference in clicks or revenue that holds across several sends means timing matters for this list, and the winning window should be adopted. A small difference that is consistent across sends suggests a modest effect, worth keeping but not worth a large investment of effort. No difference is a real finding, and it means you should stop testing timing and move your attention to the content. A difference that reverses between sends usually indicates noise, or a segment effect hidden inside the averages, and calls for a larger and cleaner test rather than a rollout.
Common mistakes
The most common mistake is testing on uneven groups. A morning cohort full of loyal buyers will outperform an evening cohort of lapsed subscribers whatever the time of day, which makes the test meaningless. Trusting open times is the second. Image pre-loading makes them unreliable, and decisions built on them will drift. Declaring a winner after one send is the third; a few sends rarely settle a question for a list with mixed habits. Ignoring time zones is the last. A single best time in your home zone may land in the middle of the night for subscribers abroad, and the averages will hide that completely.
Built-in send-time features
Many platforms now offer per-subscriber send-time prediction. These tools can be useful when each subscriber has enough history for the model to learn from. They are less useful for new subscribers and small lists, where there is not enough data to predict from, and they can make results harder to explain when something goes wrong. The sensible order is to run your own time-zone test first. That gives you a baseline, so you can see whether a predictive feature adds anything beyond a simple local-time send, and it tells you how much the feature is worth before you pay for it.
A quarterly routine
Send-time patterns are not fixed, so the test should be repeated. Each quarter, repeat the data review and retest the two strongest windows against your current baseline. Around clock changes, check that time-zone sends still land at the local time you intended, since daylight saving can shift the outcome. After any major change to the list, such as a large acquisition or a new sign-up source, re-test, because the new audience may behave differently from the one you measured before.
Send-time optimisation earns its place when it is tested honestly. Measure your own list, decide on the metric that reflects what readers do, and let the results, not the averages, set the schedule.
Key Takeaways
- Generic best-time studies are averages; your own engagement data is more reliable.
- Segment by time zone first, since it is usually the biggest and simplest gain.
- Morning and evening sends can perform differently on desktop and mobile, so track both.
- Test one send-time variable at a time and track both opens and clicks.
- Re-test quarterly, because subscriber habits shift across the year.
Related Reading
Frequently Asked Questions
Aim for at least ninety days of engagement history, and enough recipients in each test group to see a meaningful difference. Small lists rarely produce reliable winners from timing alone.
It matters less than content and relevance. On small lists, consistent, relevant sends and time-zone sends usually do most of the work.
Only when your data shows different patterns. Test per segment before you assume that the patterns differ.
Privacy features that load images automatically record opens whenever the email is delivered, so open times are unreliable. Use clicks and conversions as the decision metric.
They can help once there is enough history per subscriber. Run your own time-zone test first, so you know whether the feature adds anything beyond a simple local-time send.
Was this useful?