Whose time is it anyway?
Why timing matters in email marketing
We’ve all been there. You’ve done the hard work of designing and executing a brilliant campaign. The copy is snappy (thanks GPT!), the creative beautiful, the offers irresistible, and incentives will surely maximize margins. You’re about to hit go. Wait – when are we going to send the emails? Sending times can be the component of a campaign that’s invisible, but it’s time we brought it into the light. At OfferFit, we’ve found that time of send is a powerful lever for personalization.
Picking the winning time
Much research has been done to address the question of when to send, and a quick google will give some quick answers. Here are a few:
|
|
|
|
9am - 12 pm | 12pm - 3pm | 6am - 9am | |
10am | N/A | N/A | |
10am | 9am | 8am | |
3pm | 6pm | 12pm |
The result of this research is perhaps all too inevitable – we’re all getting marketing emails at 10am local time, on the dot. Of course, most marketers – including anyone who read past the headlines of the studies cited above – know it’s not so simple. These results represent averages – the global “winners” across large populations.
Marketers want to find the “winners” for their own customers, and the typical tool is A/B testing. But this approach has the same fundamental problem as choosing times based on industry trends. When an A/B test comes back with a clear result, it feels like a win. “10am sends got more clicks than 4pm sends!” The problem is that some customers clicked on the 4pm email. If marketers adopt the “winning” option from the A/B test, they are simply letting the majority rule. In an election, getting 51% of the vote means you win. In a marketing campaign, engaging 51% of your customers is probably a losing strategy – 49% of your customers aren’t engaging with your email.
What about segments? Dividing customers into segments is better than nothing, but it doesn’t solve the problem. You are still deciding how to engage each customer based on a single data point – their segment. Marketers’ rich first-party data describes hundreds of customer characteristics, and it's simply not feasible to take all this data into account when building segments. With a customer population of thousands or millions, dividing customers into 100 or 1000 segments merely repeats the “majority rules” problem 100 or 1000 times. Marketers need the best time for each person, not each segment.
AI testing platforms like OfferFit experiment and learn the best options 1:1 for each individual, including the most effective time of send. While personalizing 1:1 is of course the most effective option, there are general lessons to be learned. OfferFit’s data scientists have dived deep into the data to understand the impact of send time on lifecycle marketing campaigns. Let’s take a look at the results.
Time of send matters
The most important takeaway from our research is simply this: time of send matters! OfferFit’s AI testing can simultaneously optimize across many dimensions – such as product offer, incentive, channel, message, creative, time of day, day of week, and frequency of communication.
We’ve seen marketers get 20-30% of the overall lift
in campaign performance from finding the best time
to contact each individual. Furthermore, choosing the right day of the week
can have a similar impact. Of course, time of day and day of week are interrelated – most of us check our email at different times on workdays versus weekends, for example. That means that 40-60%
of the value of 1:1 personalization can come from when you contact the customer – the right time on the right day
.
Intuitively, that makes sense. All of us get marketing emails we never read, and most of us can remember converting from a message or ad that came at the right time. And of course, marketers should continue to personalize the “what” – messages, subject lines, product offers, incentives. But the “when” is equally important.
Try unusual send times
A marketer committed to experimenting with different times of day – whether through old-fashioned AB testing by segment, or modern AI testing – needs to pick times of day to test. All too often, marketers default to picking “top of the hour” times – 6:00am, 10:00am, 4:00pm, etc. Some marketing automation platforms only offer marketers the option of testing times that are exactly on the hour.
OfferFit has found that “off hour” can outperform their “on the hour” siblings. A major beauty brand used AI testing to optimize an email marketing campaign. OfferFit’s AI experimented to find the most effective email sending time for each customer in the campaign. (Here “most effective” means the choice that maximized conversions.) The chart below shows how often OfferFit chose each sending time. Overall, the “on the hour” sending times were chosen much less frequently, because the AI found that most customers were likely to ignore those emails. For example, 6:41am, 7:19am, and 8:47am all significantly outperformed 6am and 8am.
![unusual times](/_next/image?url=https%3A%2F%2Fimages.ctfassets.net%2Fvwvqbifr6a0g%2F2HVDLEICBSpVRBlf4J2L8j%2F5b79146c05dbd0f88f5fd27b40aca925%2Ftimes_graphpng.png&w=1080&q=100)
Why might this be the case? It may simply be that most marketing emails come in at the top of the hour. It feels strange to select 8:47am as a sending time, and most marketers don’t. Many of us start meetings or other appointments at the top of the hour – I might not be able to look at my phone at 1pm or 2pm, but might sneak a glance at 1:24pm. Notice also that the pattern is not universal – 8pm performed well. Maybe dinner is over and the kids are finally in bed?
Still, absent further information about the specifics of your customers or campaigns, try “off hour” sending times – you might be surprised at the lift over sending on the hour.
The right time to send depends on the individual
Of course, emailing every customer at 9:41am is not really any more personal than emailing them all at 10:00am. (Though 9:41am will likely outperform 10:00am even so for the reasons given above.) The right time to email a particular customer could depend on the topic of the email, the nature of the campaign, the day of the week, or hundreds of other characteristics unique to that particular individual.
OfferFit’s AI finds patterns in these relationships. The table below shows the patterns the beauty brand found between customer characteristics and the times at which those customers were more likely to engage with an email.
![Arrows](/_next/image?url=https%3A%2F%2Fimages.ctfassets.net%2Fvwvqbifr6a0g%2Fwd5yni6JfQNn8H0TOWq0V%2F17ce0875289fcdfb53d44d436b5f2291%2Farrows.png&w=1080&q=100)
Some of these trends are expected – if a customer has clicked evening emails in the past, it’s common sense to email them in the evenings. Other patterns are less obvious, and likely wouldn’t have been discovered through A/B testing. For example, customers with higher historical spend are more likely to engage at off-hour send times. Perhaps customers with higher spend are also higher earners with busier jobs, who are always starting meetings at the top of the hour. Perhaps these big spenders simply get more marketing emails overall, and so it’s more critical to stand out from the crowd.
The only way to truly find out the best time to engage with each individual is to discover it empirically, and the only way to experiment 1:1 at scale is with AI testing. And as you test and learn, remember: what you send is no more important than when you send it.
William Palmer develops AI testing software for marketers. Before joining OfferFit, he worked as a data scientist for an IT consultancy. He holds an MS in data science from Harvard University.
Nathaniel Rounds writes about AI and machine learning for nontechnical audiences. Before joining OfferFit, he spent 10 years designing and building SaaS products, with an emphasis on educational content and user research. He holds a PhD in mathematics from Stony Brook University.
Experimentation unleashed
Ready to make the leap from A/B to AI?