Q. What is the best email frequency to maximise response?
Is there an optimal email frequency? Is it one email a quarter, week, month or day even? Is less more or or is more more?! Here we present 2010 data on emails per month for the UK and 2011 emails per week data for the US where the frequency seems to be higher.
This is a basic question every digital marketer has to try to answer to maximise profit or response of email activity. I thought I’d share some some testing suggestions and case studies which could help you decide.
Average UK Email Frequency
The 2010 UK DMA National Client Email Marketing Report which I wrote the conclusion for this year provides useful insight on this question. It shows a wide variation in email frequency, but with a pattern evident of lower frequencies in 2010.
We are looking to get the right balance between email overexposure and underexposure. With overexposure, the recipient receives e-mail from the same company so frequently that they don"€™t have the time to read it or feel they are being spammed. They become "€œemotionally unsubscribed"€.
On the other hand with underexposure, opportunities and sales are lost since the customer does not receive e-mails sufficiently frequently.
Average US Email Frequency (2011 update)
Information on average frequency in the US is available from a compilation on Retail Email Blog which has a regular email frequency curve based on popularity of different days amongst US retailers.
This example shows how there is quite a variation in frequency – as high as nearly 5 emails a week near in peak-selling before Christmas but dropping to an average of around 3 in other trading periods. The two curves certainly show there is a higher email frequency too.
Evaluating current email frequency and customer response behaviour
The first step to answer this question is to assess the impact of your email marketing frequency on customer activity and perceptions. If frequency is too high, subscribers will tune out. The obvious thing to measure is aggregate open and click rates and most email broadcast systems are good at this. Mark Brownlow has a good roundup of research on customer perceptions of email frequency. In one study 73% said that frequency was the main reason for opting out. Ouch!
Another measure is to look at the average of number of emails you and your competitors send to subscribers per week, month or year.
Econsultancy reported that the average frequency in 2009 for US eretailers was 2.5 emails per week and 11 emails per month giving an average of 132 emails per year.
But you need to go beyond this and use these measures that most systems can’t measure readily, so you need to do some more analysis to identify:
- Average frequency of email received and plot profile by frequency for different list members – to see the proportion of the list who are receiving too many or two few emails – see chart.
- List activity – The % of your list that open, click and buy within a period, e.g. quarterly or annual.
- Recency of response – what is the average for the last open, click or purchase – a good tip is to store recency in your email database as a field for analysis. Alternatively score list members by activity and store this in the database also.
- Break down list activity and recency measures by different type of list members – it may the frequency is working for some segments but not others.
- Break down list activity by time on list – commonsense suggests, that the longer they are on your list, the less responsive your emails will become
Testing options to decide on the best email frequency
It"€™s not an easy question to answer by gut instinct, so you have to test. So how do you decide on frequency? Here are some ideas and examples showing how you can approach frequency testing:
Defining a random control group to test frequency changes against. Here you continue with current mailing frequency for the control group and then vary the frequency for other groups and review changes in response and in particular revenue per 1000 subscribers. In one case a bank tried frequencies of 1,2,3,4 times per month and found the right frequency this way.
Example 1 Toptable
Sean Duffy reported how Toptable measured the long term impact of increased frequency by creating a control group with half the new customers that joined in a month held back from the second send.
After three months this control group was measured against those who had joined the site at the same time yet received the default setting of two emails a week. Open rates were 86% higher, unsubscribe rates 57% lower.
But the main figured that proved why sending too many emails leads to long term damage "€“ those receiving only one email a week had made 14% more bookings than those receiving two emails over that three month test period!
Example 2 Net-a-porter
In this case Fashion e-retailer Net-a-Porter.com reduced the number of emails it sendt to customers from up to 10 per week to two according to Brand Republic.
It had been emailing some customers up to 10 times a week with information including generic updates, highlights from specific designers and details of new products.
After the experiment Net-a-Porter.com now sends each user two automatically generated emails a week that take into account their specific interests and preferences. Conversion rate has increased. Product update emails get a conversion rate of more than 10% and newsletter emails are opened by nearly half of recipients.
The report also shows the importance of getting email marketing frequency right. The company sends out around 300,000 emails a week. Email drives 32% of Net-a-Porter’s sales and generates more than £1m in revenue each month.
If you have a single email newsletter as in the Toptable example, testing is more complex than these examples suggest since there are a range of different types of emails such as enewsletters, promotional offer emails and also individually tailored event-triggered emails. Different offers or creative to each segment will also be overlaid upon this.
Other options to solve the frequency dilemma include:
A. Reduce Email frequencies automatically for lower responding customers? Set a database field for activity or engagement level for each customer to help implement this. Amazon is good at implementing this and increases frequency through event-triggered emails sent in response to someone browsing, searching or buying "€“ that"€™s the smartest approach.
B. Change frequency for different segments. One frequency size is never going to fit all. So if you find that open or click response is lower for certain segments, then decrease the frequency when they are inactive.
C. Give customers a choice on frequency. You do this through their profile or "€œcommunications preference centre"€. Give options to change content and frequency preferences through profile or survey (E-mail, DM)?
D. Increase Direct Mail or SM for customers with a lower Email response. This is sometimes called "€œright channeling"€. To test the value of this use a holdout group. This small group, perhaps 5% of your list or a specific segment who doesn"€™t receive the catalogue (or email if you"€™re testing this) at all.
E. Re-engagement campaigns. Re-activation campaigns use content or discounts to encourage email subscribers to become active again.
I’d be interested to hear what you think where you’ve tested this, or what you think is too much.