monty python spamalot

ECRM Database Segmentation Rocks

By Paul Morris
Let’s face it discussing database segmentation with the Mrs does not get her in the mood. It doesn’t really get me in the mood either however ECRM segmentation truly does rock. When applied to your ECRM (email customer relationship management; a fancy name for email marketing) database it will lead to improved engagement, better brand perception, increased immediate sales and incremental LTV (Life Time Value).

As I detailed in a previous post on the topic of succesful email marketing it is far too easy to destroy your database hence here goes with a few tit bits to help you on your ECRM way.

Let’s start with the basics – Segmentation allows you to break your database down into smaller groups of people who share similar characteristics. To begin with make sure your segmentation approach links back to your KPI’s (generally the most basic is increasing engagement/ relevance > more sales/ leads).

Then when you start to segment make note of the amendments, segments mailed (and when), measure the impact and then how you will evolve your segmentation strategy as otherwise what’s the point?!

What data do you have available? What bits of data do you think will prove key to exceeding your KPI’s? When was the data collected (data relevance clearly degrades with time and this degradation is exacerbated in certain sectors more than others)? Can I improve my ECRM data by combining it with any other data that is ‘laying around’ the place (customer loyalty information, Offline Point Of Sale, online sale, Click through, etc)?

Basically if you only have limited customer data intelligence you are going to struggle with meaningful segmentation. However if that is the case then it’s your job to do something about it. How can you improve your data/ customer knowledge? It might be, as I highlight above, by tying that data in with other pots of data to enrich it. It will almost certainly be linked with a combination of the following 3 tactics:

1/ Ask your customers what they want

Lifestyle based questionnaires as an example will prise more data from your email recipients. If you utilise reciprocity (essentially quid pro quo) then there will be value in your customer willingly surrendering data to you. If they really like you/ see the benefits then the person might even update their own profile online e.g. as is the case with Amazon.

2/ Find out what your customer is interested in NOW

Behavioural data, “you are what you click”, is extremely powerful and Amazon do this amazingly well by tying in your logged in user profile with their ECRM back end. I clicked around the watch section of Amazon yesterday and funnily enough today I received an email with an update on special offers and different types of watches I might not have otherwise considered. Brilliant!

We must however remember that not all companies have people who are logged in and who have willingly handing over credit card data for one click check out thus other targeting methodologies need to be employed. You could start by simply segmenting on when the customer last clicked a link (and what they clicked on). You would send differently worded/ strength reengagement emails asking them to come back to you and if they do they receive a shiny X% discount of their next order (a reciprocity reengagement approach).

Also when you do start to segment make sure you have checks and procedures in place to stop the following case study taking place – A past client (they will remain nameless and were not a client on the ECRM front) has been sending a steady stream of female related fashion emails to a colleague for the past two years. He is male and thinks he only engaged with their emails once-twice hence why on earth has he been getting incorrectly targeted super basic gender emails for 2 years?!

3/ Profile your customer

Essentially you will improve your profile data by asking more about them (linking in with point 1 above) or more likely tying in your ECRM database with your other database data (sale or loyalty card for example) thus understanding age, gender, geographic location, spending habits, etc.

Walk before you Run

Do not start to segment into hundreds of pots at the start of your segmentation journey e.g. men, 30-45, cross dressing, one legged smokers. These data segments will likely be too small to make economic sense in targeting with unique messaging and it will overwhelm your tiny mind and your team.

Start with only a handful of segments/ ideas and take things from there. As an example the amazing Tesco club card ECRM program only started with 4 segments and now look at them! Please do understand though that ‘1’ segment can still be highly targeted e.g. split your database out into UK geographic locations, however this 1 principle should be manageable before you try anything more complex with those 1 legged cross dressing mid-life smokers.


Once you figure out your basic segmentation strategy you might eventually go down a smart arse route utilising Recency, Frequency, Value modelling. You could assign a code to each customer relating to the recency of the last purchase, the frequency of the purchase and the value. You then group people together dependent on the RFV codes. Simples.

An easier RFV route would be to scrap the code route (personally I have never tried the code grouping route as it hurts my tiny mind just thinking about it in depth) and pick mailing campaigns based on RFV insight. e.g. target high value normally regular purchasers who have not made a purchase for 2 months.

Do something with the segments for God’s sake

Then when you have built the segments do something with them and tie in with personalisation to further improve relevancy (whether that is in the subject line or body of the email). Understand the segment characteristics, what they might want/ need and then understand how best to convey those messages. These differences might emanate themselves in different: colours, email structure, call to action messages, tone/ structure of wording, products, etc and relate to customer age, sex, geographic location, expiry of membership, birthday, RFV (e.g. VIP exclusive offers), etc.

Go forth and spam-alot (sorry I mean strategically engage with your customer base)!