How much contact will your recipients tolerate before hitting that spam button or ignoring you? Is twice a week too much? But what if once a week is too little? If our recipients get 300 emails a day in their inbox, that’s 2,100 emails a week. From address (brand reputation), subject line and preview pane design have to be amazing to get attention among the other 2,099 emails from the boss, partner, customers. Would it really hurt to send 2 or 3 a week?

We at Lyris have been doing some testing on the effect of different message types on recipient frequency tolerance. The results are surprising. Initial findings show that the right combination of service messages and marketing messages can allow much more frequent communication and even increases your response rates and conversion.

Two experiments, in particular, with Gwan Yip and Amanda Metcalfe of fashion retailer www.koodos.com have given thought provoking results.

Experiment one

Sending the MENS range to the Female segment on the same day as the normal WOMENS range email. The Female segment is calculated as those who have told us they are Female through our welcome programme, or those who have a Female title (customers).

Friday Morning, Clearance sale day. The WOMENS range email is sent to the Female  segment throughout the morning UK time,

• Subject line:  80% off CLEARANCE - going...going...gone
• Deliverability – 99.8%
• Open rate 2% above average for gender targeted email
• Clickthrough rate 4% above average
• Conversion rate – average for targeted clearance email
• Unsubscribe rate - average

Friday Afternoon, same Clearance sale day, the MENS range email is sent to the Female segment:

• Subject line: 80% off our CLEARANCE - going...going...gone
• Deliverability 99.8%
• Open rate 2% above average for gender targeted email, higher total opens than WOMENS mail
• Clickthrough rate 3% above average with lower total clicks than WOMENS mail
• Conversion rate – average for targeted clearance email with  lower total revenue than WOMENS mail.
• 10% More transactions than WOMENS email
• 14% more web activity than WOMENS email
• Wider range purchased
• 39% who opened MENS email had not opened WOMENS email
• Unsubscribe rate - average

The first key point to note here is the fact that the unsubscribe rate on the MENS email is average. If our Female recipients had considered this email to be inappropriate or irrelevant, we would have seen an increase in the unsubscribe rate and the complaint rates from Hotmail and Yahoo. We saw average responses on these metrics and normal deliverability across both mailings. This shows that for this particular segment, sending two emails in one day does not provoke a negative reaction.

The second key point is that if we look at the two mailings as part of a ‘clearance campaign’ we increased our unique open rate overall to 8% above average. We more than doubled predicted transaction and web activity numbers and nearly doubled predicted revenue. This is clearly a winning campaign.

It is interesting to compare the profile of the open rate over time for the two mailings:

WOMENS mailing to female segment - this is a normal pattern for WOMENS content to the Female segment

MENS mailing to female segment - this mailing had repeated spikes of interest over the weekend and particularly on the days of the following week.

The Female segment is clearly interested in the MENS range. You will remember that the Female segment is defined by recipients identifying themselves as Female through the welcome programme or their title. We have made an assumption that they would be interested in the WOMENS range, rather than the MENS range. The response rates for targeting range to gender have been good for the past 12 months, and the Female segment had not had much email exposure to the MENS range. However, the results of this experiment show that our Female segment also has interests in the MENS range. This is borne out by the increased web activity and wider breadth of range purchased from the MENS email.

These results may have been different if the Female segment had been defined by declared interest ‘I am interested in your WOMENS’ range or previous behaviour (those who had clicked through on WOMENS range items in an email or spent significant time on WOMENS pages on the site.

It is worth bearing in mind that the time of day may have (morning and afternoon) will have had an impact on these results. In order to eliminate this we should set up a control group that receives either a second WOMENS range email instead of the MENS range email at the same time in the afternoon along with a second control group that received nothing.

Is it possible that mailing twice in one day actually had a beneficial effect too? 61% of those who had opened the MENS email had also opened the WOMENS email. The reaction to receiving two in one day, seems to have inclined strongly towards “Wow there’s lots going on at Koodos!” rather than “That’s too much contact from Koodos”.

The most important result of this experiment was that 42% of the openers who engaged with the MENS range email who had not opened the WOMENS range email that morning had not opened an email in the previous 60 days. This means that we were able to re-engage a significant number of recipients with the fresh content of the MENS range.

These results tie in with the results of a previous experiment with sending two messages of a different type to the same database in the same day at Koodos.

Experiment two

The morning email was sent in the morning of 17 July.

• Subject line:  FREE DELIVERY plus designer festival chic on the cheap
• Delivery, Opens and Clicks were in line with average and revenue was above average. 
• Unsubscribe rate: average
  

The afternoon email was then sent at Midday on 17 July to the same recipients.

• Subject line: Your early access to the Lulu Guiness sale
• Results were average for deliverability opens and clicks.
• Revenue was above predictions and excellent.
• Unsubscribe rate: average
• 65% of those who had opened the ‘Festival Clearance email’ also opened the Lulu Guinness email.

Here again, we see no negative impact on unsubscribe rate or delivery and we see very positive open repetition and engagement. In this case the email is not a different range, but a different type of email. We have a single designer and early access to a sale of a particular range. This type of email is perceived more as an alert and therefore as a ‘service message’ rather than the ‘sale message’ of the ‘Festival Clearance’ email. We have seen with other fashion retailers such as www.net-a-porter.com that service messages such as ‘What’s New in Stock for you today’ emails can be sent in close proximity to more editorial style emails.

Andrew Robinson
Director of International Professional Services
Lyris Inc.