Don’t make the mistake of not using your historical transactional data

Why it’s important to make use of past purchase data in your email marketing

So, you’re a retailer, running either a click or mortar business, or maybe both, and you’ve been in business for a number of years. This will no doubt mean that you have a significant amount of transactional data about your customers. What they purchased, when they purchased it (recency), how many orders they placed (frequency), and how much they spent (monetary value). We can use all of this data in the RFM (Recency, Frequency, Monetary Value) model to analyse our customers and send out more focused communications. This is important in helping us understand their behaviour and foresee what they might do in the future.

The RFM model

The RFM model can be interpreted as follows:

  1. Recency: how fresh your relationship is with your customer; the fine line between a customer being active or inactive
  2. Frequency: a measure of demand; the number of orders in a certain period of time.
  3. Monetary value: how much your customer is worth; their average order value.

The advantages of applying the RFM model in analysing and segmenting your customers are manifold, allowing us to potentially;

  1. drive up-sells & cross-sells
  2. improve retention
  3. increase customer loyalty
  4. grow revenues
  5. increase profit margins

Segmented and targeted emails generate 58% of all revenue – DMA

How to make use of your customer’s transactional history

The first step in being able to import your customer’s past purchase data. Once you’ve done that, you’ll be able to segment your customers based on their purchase propensity, i.e. taking into account the recency, frequency and monetary value of their past transactions.

To import your transactional data, you could create a one-to-one mapping of your existing data model (be that from your e-commerce software, CRM or other data source) with ExpertSender’s data tables (a user friendly relational database model). Some of the data that might be of interest to you includes order details such as: the transaction total, transaction date, acquisition source and product details such as the product type, model and category. In a relational database model with normalised table relations that might consist of the following: order, order_item, product, product_category and category tables.

Example: table relationships in the ExpertSender system.
Example: table relationships in the ExpertSender system.
UML Entity relationship model of the table relationships created in the ExpertSender system.
UML Entity relationship model of the table relationships created in the ExpertSender system.

Moving forward, you should store all of your customer’s future transactional data, so you can continue using it in your marketing efforts. This could be in real time via the ExpertSender API or you could have it synchronised on a predefined schedule according to your business needs (be that, hourly, daily or at another cadence).

Analysis & Segmentation
So, now that you’ve got your data, let’s make use of it. We imported some transactional data that we’ll analyse on the basis of the RFM model.

Number of purchases
Analysing your customer’s purchase history is very important in retaining your customers. When segmenting your customers, you should treat customers who made 2 purchases differently than those who made 5. Increasing the number of purchases made by each of your customers will give you a higher customer lifetime value over time.

Most recent purchase date
Knowing the details of your customer’s last purchase will allow you to build a customer lifecycle, and targeting users who you know are just about due for another purchase makes for an ideal segment. Furthermore, sending your campaign at the most appropriate time is paramount in trying to engage your customers as much as possible.

Source
Knowing the source of your customer’s purchases will help you build a long-term strategy, allowing you to analyse your source traffic value and compare sources against one other.

Product & category
Making use of product and category purchase history in your campaigns can have a profound effect on your numbers. Customers who purchase “Le Coq Sportif” shoes again and again might just want to buy them again. You should give your customers what they want to see, based on their historical purchases.

This data will help us build a customer journey, and hopefully convert one-time buyers into long-term repeat buyers. It’s therefore in our interests to make use of their purchase history. So, let’s analyse it!

Transaction Analysis

When we analyse our transactional data, and cross reference it against other variables, we can identify patterns in our customer’s behaviour and correlate it with other variables that provide us with insights into how, where and when we should communicate with our customers.

Propensity Modelling

Calculating our customer’s propensity to buy, and understanding the touchpoints we need to engage them with to increase their propensity to purchase is very important.

By targeting these customers who have a higher propensity to purchase, and by sending them sales propositions at the right time, we’ll increase our campaign’s conversion rates and improve our ROI.

Mapping our customer’s journey

Analysing our customer’s transactions allows us to segment customers based on how much money they can bring to our business. This will help us to create a customer journey and convert a low value and unengaged customer into one that’s high value and highly engaged.

Our customers have different characteristics and we should have personalised dialogs with each of them. We should send a different message to someone who’s just begun their customer journey and to someone who’s a regular buyer.

It’s by analysing our customer’s transactions that we’ll learn important insights about our customers and improve campaign performance and ROI. Without performing this analysis, it would be much more difficult to know who our most valuable customers are. It’ll also allow us to fine tune our campaigns for further improvements. We’ll find our best performing customers and that’ll allow us to better target our customers at a lower cost.

Who are your most valuable customers?
Now we’re going to identify who our most important customers are and decide how we should divide our efforts amongst them.

It could be argued that recency is the most important metric because it’s an indicator of a customer’s activeness. A recent and frequent customer is our ideal customer, whether their spend is large or a little less so. Those who had spent big and regularly in the past, but haven’t done so for a while, may also be worth spending some time on.

Segmentation based on your customers propensity to buy
Let’s see the RFM model in play, where we’ll segment all the customers whom we classified as our most valuable. I.e. they made a purchase on avearage at least once each quarter,within the last 3 months, with an average order value of $150 or $600 in total over the last 12 months.

UML Entity relationship model of the table relationships created in the ExpertSender system.
UML Entity relationship model of the table relationships created in the ExpertSender system.

This segment could also be easily modified to take into account other types of customers. E.g. customers who have a lower propensity to buy, and have only made one purchase. Depending on your line of business and customer lifecycle, it’s possible to target your customers in tailored ways that’ll be more profitable for your business.

Make use of your transactional data to grow your revenue

Well, there you have it, if you’ve got historical transactional data about your customers, then you already know their spending habits and preferences. Add RFM as an additional layer to your segmentation and you’ll know who your most valuable customers are, when to contact them and at the most appropriate cadence. Are you already making the most of your customer’s transactional data?