Behavioural email – Instant or targeted on behaviour

By: Garry Lee | June 30th, 2010

When is the best time to send a behavioural email?

Behavioural email is a hot topic nowadays; plenty of people want to add it as a key part of an overall conversion strategy. Of late the thing that seems to be generating the most column inches is around the best time to send these emails. So as someone that’s been doing behavioural email for longer than most I thought I’d throw my opinion out there.

First up, let me address the one thing that annoys me more than anything else; the general assumption ‘instant is best’. You’ve all heard the argument’s. “They are in the buying mode hit them immediately” or “all responses to behavioural emails happen within an hour”. Now, don’t get me wrong, I’m not anti instant emails, however, before you start triggering emails to go off left, right and centre, here’s a few things to remember:

Never assume: With behavioural email we don’t assume when a customer wants to get an email, in the same way we don’t assume what content to send them.
Use the rich data available in your web analytics databases to understand customer behaviour (in the same way we did to create the first emails 10 years ago).
Each sector & emails within the sector have different response times. ‘Save quote’ on insurance sites might work best with an instant quote, but that does not mean it’s the ultimate for retail catalogue basket abandonment.

The other argument I hear is people stating that most responses to instant behavioural emails happen within the first couple of hours. Well of course they do. Email is an instant response tool. If I send the email 1 hour after they abandon, then the response is mainly within a couple of hours. If I send the email after one day, the response is mostly the first few hours. If I send the email after 3 days, most of the response is within the first few hours. See the pattern!

So how do we go about deciding the timing to send emails?

Look at the web analytics data and look at response curves for people in your behavioural cells.

TEST. Try sending at different times and look for the email with the best results.
Talk to the customer. If you are going to make an assumption not based on hard facts, please at least let it be your customers voice that shouts the loudest (let’s try a survey people).
NEVER assume one size fits all. I know from experience (years of analysis and testing!) every sector is different, products within sectors are different, even people using search will respond differently to people from affiliates. (I’ll save that for another blog.)

It’s not just about one off timing but deciding the best time to send a programme of emails. We have seen recently across retail clients the follow up email is getting a better response then the initial abandoned basket email. In the most recent tests the follow up conversion has been up to 12% higher!

To further highlight this, below is a response curve showing people who fit into classic behavioural cells of abandoning purchases. It looks at activity over a period of time (not from sending an email but their actual site behaviour).

As you can see, the instant response is clear and carries on for well over a day, then drops sharply. This is then followed by an increase around the middle of day 4, than another jump around the middle of day 7.

These figures will differ by sector and product but they show the absolute need to have a co-ordinated programme of emails, not a one off, one size fits all approach.

All of this is not to say that instant isn’t an option that should be used, because if the data says it works then you should use it. Just remember a few important things:

Instant can be very big brother. Consider the message and type of customer to ensure people are not put off by how much you know about their buying behaviour.
You don’t give them the chance to realise they have abandoned. Seems an odd statement but for certain products if you send an instant email the user might still be doing something towards the purchase and not understand why they got this email and be put off.
You give away unnecessary offers. Test if people responding to an instant behavioural email would have purchased anyway. When we started behavioural email 10 years ago, the first thing we did was have to prove that we were getting incremental sales that would otherwise have been lost. I’m not seeing these kinds of case studies around anymore.

Without giving too many state secrets away (if you want to know more give me a call) I will share a few of the observations we have, based on the actual results we have seen from testing in the last couple of years.

Save quotes work well with a quick response email (told you I wasn’t anti instant emails), but generally require a follow up to maximise conversion.
Retail abandonments require a series of follow ups, but the timing and message depends on the product and type of site (i.e. catalogue customers have a very different buying pattern to the average ecommerce site)
Travel sites should utilise a more rounded message and not be obsessed with exact details. People respond to a creative email around a subject of interest. Don’t scare customers with how much you know about them.

So three killer things you should remember about the timing of behavioural emails:

Use the data you have to see what is happening. Never assume.
It’s not just a one off, it’s about a rounded programme with the timing of each critical.
Test, test and test

In the meantime, if you want to understand more about behavioural email, please call 01270 848 490 or drop me an email at blog@redeye.com.