When our clients ask this question, they are typically looking for a "time of day" and "day of week" that they can apply to every campaign they send. But unfortunately it's never that easy...
- Is timing the most important factor in eliciting an interaction from your subscribers?
- Can you keep your hypotheses open to uncover unlikely insights?
- And is there a more effective way to test timing by using historical analysis?
From our experience, of all the things that you can test about your regular newsletter or offer-based email campaigns, time of day and day of week testing provide the least valuable insights.
A far greater number of people on your list are likely to buy because they have just been paid than the number of people who buy because you sent the email before 1pm on a Thursday. See more about the effects of payday in this past article.
It's worth bearing in mind that we're talking about regular 'one-time' or 'push' mailings here, those that are scheduled in your mailing calendar. When you create transactional or lifecycle emails, the time lapse before or after the event or lifecycle point can be critical, but we'll talk about timing tests for those in a future issue.
Why do Subscribers Interact
Research on consumer motivations has shown that consumers' personal motivations exert more influence than the time of day a message is sent. When Forrester asked "How much did each of the following influence you to open the emails you receive from companies? Only 9% of consumers cited the time and/or day they received the message as having any influence.
By comparison the top four were:
Your interests 59%
Your needs at the time 51%
The sender 48%
I like the company 46%
Exact Target have also published a Channel preference survey which looked at usage patterns in email by time of day, it identifies an 'email prime time' of maximum engagement between 8am and 11am.
Testing & the Problem with Averages
One of the most prevalent mantras in email marketing is TEST, TEST, TEST & TEST again; so there must be hundreds of marketers out there planning to roll out day of the week and time of day tests at this very moment. But is this the best approach?
Alchemy Worx has found if you really want to optimise send frequency; it is vital to target all behaviour not just the average, because not everyone on your list wants the same thing. This is also true of send time. If you conduct a standard day of the week test, all you will find out is the best day for your list and not what is best for any given individual or groups of individuals. For any given day of the week or time of the day that you could choose to send your campaign out, including 3 am Sunday night, there will be at least one person on your list who thinks that is the perfect time or day and one person who thinks that it is the worst time or day.
Your Goldmine of Historical Data
An effective shortcut for any timing analysis will be to look back at what has happened over the past 12 months. Over that time period most companies will have sent messages on every day of the week most times of the day - often without prior intent to test. You will also have data on the full impact of your campaigns, by waiting for sales and revenue rather than relying only on immediate opens and clicks.
This historical data is a goldmine! Using this data, you can analyse the effect the time and day a message was received has had on the performance. At the very least you can confidently make your test plan more manageable by focusing on the most likely looking days or days where you do not have any data; best of all you will also have a benchmark to work from.
Our golden rule is this; before you test anything ask yourself if the answer (or part of the answer) is right under your nose in your historical data.
One of the other benefits of conducting a forensic audit of your historical data is that it can yield unanticipated and valuable additional benefits. In setting out to discover the best day of the week or time of day for our clients to send their campaigns out, we discovered that which week of the month a message was sent had a far greater impact on revenue than other temporal considerations!
Expanding your Analysis
Our in-depth analysis over the past 5 years has consistently shown that the best week of the calendar month for generating revenue (in the UK) is week 5!
Does your list also fit this pattern? Why not test this for yourself?
The reason for it is simple, most people feel richest just after payday, and in the UK, the majority of salaried people get paid between the 25th and 30th of each month.
Implementing your Findings
So if your campaign is monthly, aim to get your message out as soon as possible after payday - weeks 5 or 1 depending on the month. If at all possible try to avoid week 2 and to a lesser extent week 3.
If your campaign is weekly you may want to consider lower discounting or adding less value in weeks 5 and 1. Conversely if you need revenue in weeks 2 and 3 you will need a much stronger offer.
So back to the first point; if you also think about send time/day or week in terms of size of segment i.e. how many of the people on your list are likely to feel the same way about something or be in the same situation; a far greater number of people on your list are likely to buy because they have just been paid than the number of people who buy because you sent the email before 1pm on a Thursday.