A/B test email send times for higher engagement

Timing can determine whether an email is read, clicked or ignored, yet the best send time is rarely the same for every subscriber. Australian audiences add another layer: Sydney and Melbourne share a time zone, Brisbane sits 30 minutes ahead, and Perth is two hours behind. A morning campaign can look strong nationally while missing customers elsewhere who are still commuting or opening messages over breakfast.

A useful A/B test isolates delivery time while keeping the audience, subject line, offer, creative and call to action consistent. It compares a control time with treatment windows using a predefined engagement measure. This produces evidence that is safer to act on than choosing a “best” hour from industry folklore, giving Australian marketers a repeatable way to improve email engagement without disturbing deliverability.

Define one timing hypothesis

Treat the first campaign as a hypothesis, not a search for a universal sweet spot. A retailer might predict that delivery at 12:30 pm will generate more clicks than 8:00 am because shoppers browse during lunch. In an A/B test of email send times, change only that variable. Identical content makes any difference attributable to timing rather than a stronger subject line, larger discount or redesigned email.

Australian lists complicate simple clock-time rules. A 9:00 am send reaches Sydney and Melbourne at 9:00 am, Brisbane at 9:30 am and Perth at 7:00 am, so define windows in recipient-local time where possible. Daylight saving also matters: New South Wales, Victoria and South Australia shift their clocks, while Queensland and Western Australia do not. Record testing dates and use UTC scheduling so clock changes do not distort the comparison.

Set the hypothesis before seeing results. Choose a contrast, such as a weekday morning against a weekday lunch window, and state the expected effect. For an electronics campaign, current electronics discount codes can supply offer material while send time remains the only variable. The landing page, product mix and email content must stay stable between groups.

Build fair audience groups

Build mutually exclusive treatment and control groups from eligible subscribers, using random assignment within meaningful segments. Segment by factors that can change behaviour, such as customer type, purchase recency and state, then randomise people inside each segment. This prevents an imbalance between engaged repeat buyers and dormant contacts from distorting the result. Avoid testing a subject line or creative simultaneously, because that would confound the delivery-window comparison.

Keep the creative journey consistent. A contact-lens merchant can use contact lens cleaning comparisons as a fixed destination for a replenishment reminder, then compare two send windows with the same destination and offer. The control group sees the original time; the treatment group sees the challenger. If one time produces more clicks but fewer completed purchases, the landing-page journey may be revealing a qualification problem that timing alone cannot solve.

Decide the sample split, primary metric and guardrails before launch. Click-through rate, conversion rate or revenue per delivered email may suit commercial campaigns, while open rate can be a supporting signal because privacy features and blocked images make opens less reliable. Set limits for unsubscribe rate, bounce rate, spam complaints and inbox placement. A higher open rate has little value if recipients mark the message as junk or disengage permanently.

Run a clean, measurable experiment

Run the test long enough to capture meaningful behaviour across several weekdays and, where useful, a weekend cycle. Two to four weeks is a common starting point, but volume and purchase frequency determine the duration. Do not stop when a variant briefly leads. A clean experiment needs stable sending rules, no overlapping campaigns and consistent treatment of bounced addresses or unsubscribed contacts.

Control the destination as carefully as the message. Beginners comparing email landing page builders should select one page for every variant, since a faster mobile page could imitate the effect of a better send time. Check that links work, tracking parameters are preserved and the call to action is identical. Use the same conversion event in both groups, excluding bot clicks, internal traffic and duplicate orders.

Evaluate results with a confidence interval or suitable significance test, then consider business size as well as statistical confidence. A 4% relative lift in clicks may matter to a high-margin retailer but not a campaign with thin margins. Examine results by state, device and engagement segment to see whether the winning window works for everyone or one cohort. Report absolute lift and expected incremental revenue, not just an impressive percentage.

Protect deliverability and customer trust

Timing can amplify deliverability problems. A message sent at 1:00 am may arrive in the inbox but remain buried by morning, while repeated late-night campaigns can appear intrusive. Australian consumers commonly check email on phones during commutes and lunch breaks, particularly in cities such as Sydney, Melbourne and Brisbane. Choose windows when recipients can reasonably engage, avoiding times that create notifications outside their routine.

Commercial email remains subject to Australian marketing and privacy rules. Under the Spam Act 2003, senders generally need consent, clear identification and a functional unsubscribe route for marketing messages. Organisations covered by the Privacy Act 1988 and Australian Privacy Principles must also handle personal information lawfully and transparently. Timing tests create no exception: a low complaint rate cannot replace valid consent or a clear sender identity.

Build frequency caps and a suppression process into the platform. Immediately exclude anyone who unsubscribes, repeatedly bounces or reports the message as spam. Compare inbox placement and complaint signals between groups. If a challenger raises unsubscribes or spam reports, reject it even when clicks rise, because sender reputation and audience trust matter more than a short-lived engagement spike.

Turn the result into a repeatable schedule

Choose the winner using the primary metric, guardrails and expected commercial impact. The result should show that the timing difference is unlikely to be random and large enough to matter. If evidence is weak, keep the current schedule or run a larger test. Avoid declaring a universal winner from one campaign: a retailer’s 7:00 pm audience may differ from a membership organisation’s active members, even within the same city.

Scale gradually and retain a control where feasible. Apply the preferred time to one segment first, watch deliverability and conversions, then expand while continuing to measure the baseline. Local-time scheduling suits national Australian campaigns because it treats Perth’s morning as Perth’s morning rather than Melbourne’s afternoon. Revisit the schedule after major changes to the email, audience, product range or tracking.

Seasonality can alter the answer. Australia’s retail calendar includes June EOFY promotions, Christmas gifting campaigns and Boxing Day sales, each creating different inbox competition and shopping habits. Compare like-for-like windows during comparable periods rather than blending a quiet week with a sale event. The strongest outcome is a documented send-time strategy tied to engagement data, deliverability safeguards and customer context, rather than a rigid hour copied from an overseas campaign.