Search the website

How a Customer Model Will Help Drive Your Business Forward

A more intelligent and personalized marketing strategy is in your reach: Track changes, behavior habits, and generate the best insights from your customers

Guess where the following seven words can be found: Personalization. Needs. Relationship. Communication. Analysis. Retention. Growth. All are mentioned in the Wikipedia page for CRM value – most of them in the first paragraph.

It’s easy to understand that everything from personalization to analysis is seen as the foundation for better retention and growth, but it’s not that easy to understand how to achieve it. One of the most fundamental tools customer centric companies (and others) should use to understand and know their customers better is a customer model.

In the CRM & marketing world, we usually face three major questions:

  • Who are my customers?
  • How much are my customers worth?
  • How much of my resources should I spend on each customer?

If built intelligently, a customer model can answer two of these questions, and slightly start touching on the third. In this blog, we will try to explain how even a simple model can help you build a better marketing strategy and achieve your CRM goals.

So, who are my customers?

The first thing a good model should do is separate between different types of customers. This step is based on the understanding that not all customers are the same, and should be treated as individuals. This separation can be achieved by segmentation and doesn’t need to be very complex – it can be based on a healthy combination of business understanding and reliance on evidence from the data. Build the model in a layered structure – starting from high level segmentation, and then breaking down each one of the segments to more granular ones.

Each customer, from their first interaction, will go through several states which are defined by their activity. For example, in most industries we can define at least three main customer states – new, active and churn customers. From this point onward, we will refer to these states as lifecycle stages. While you probably have an idea how each lifecycle stage is defined in your eco-system, it’s best to segment the lifecycle stages using the most important and basic attributes that describe the customer behavior. For instance – time since first activity, time since last activity, and of course the activity volume, preferably in terms of money. If you are experienced in conducting segmentation in the marketing world, you’ve probably crossed paths with the abbreviation ‘RFM’, which stands for Recency, Frequency and Monetary. The following graph illustrates the possible movements for our lifecycle stages:


Even now, after separating your customers into only three lifecycle stages, you can start mining information. It would be interesting to find out how many customers are in each lifecycle. Or, what is the conversion rate from new to active. And even what the churn of new and active customers is. Answering these questions will provide a lot of value, which could be translated into future marketing actions.


Once this step is done, we can increase the granularity level of the model by breaking down each lifecycle into sub segments, according to additional behavioral or demographic attributes. A simple example might be the customer product preference. An e-commerce company that sells clothes will want to know if the customer usually orders men’s clothes, women’s clothes or kids’ clothes. All three types can be active, but obviously each one is a different customer persona, and requires different communications, discounts and recommendations.
Creating the sub segments is slightly more complicated and requires using machine learning techniques such as cluster analysis, which have some advantages over defining fixed thresholds. In the next table we can see a common segmentation for active customers, looking at their purchase activity:
We can clearly see there are three groups; sub segments 1, 2 and 3, which are very loyal to only one clothes department, while sub segment 4 purchase clothes from both women’s and kids’ departments.

Going back to our original question, a good customer model will enable you to understand your customers from the macro level; by monitoring and hopefully affecting the funnel each customer goes through, all the way to the micro level; by revealing different customer personas, each might require different marketing approach.

How much are my customers worth?

After all the segments are in place, it is essential to understand how much each customer is worth to the company. Getting these numbers right will allow the company to plan its growth, resource allocation and marketing strategy.


Let’s say that in our example we have three lifecycle stages, and under each we have 15 sub segments, totaling 45 sub segments. Calculating how much money each sub segment generated in the past month, six months or even a year, can teach us a lot about the worth of the customers in each sub segment. With this information, we can estimate the average value of the customer in each segment. Imagine that now, using this method, you can easily see that customers who purchase kids’ clothes during the weekend, are worth double than customers who purchase men’s clothes during mid-week. By using a table similar to the one below, we can identify these hidden customer personas and get a better understanding of the value of each one:
Despite the fact that each sub segment had purchased a similar number of items – 22, 26, 26 and 26 – we can see that sub segment 4 has a much higher value in comparison to the others. This simple piece of information can impact your campaign’s timing, offer types and offer generosity, making them relevant for each customer persona.

How much of my resources should I invest in each customer?

This is the question that we cannot answer fully only by using our model above. As we know, marketing budgets are limited, and affected by many other factors that we have little or no control over, such as shifts in the company strategy, market changes, competitors, etc. However, by using your customer model you can identify some of your weak spots, or on the other hand – your superstars, and make better decisions regarding where to place your marketing efforts. Once all the data is available, questions like “should we invest more in new customers or churned customers?” or “which group of customers will probably generate more value?” are suddenly much easier to answer.

In conclusion – In this article we explained in a nutshell why we should use a customer model, while describing a very simplified way to do it. At the end of the day, the customer model should allow you to derive business insights, track changes in your customers’ behavior and make better decisions. Hopefully, these insights could be translated into more intelligent and personalized marketing strategies. So, if you don’t have a customer model yet, or you do, but it’s not taking its rightful place in your agenda, we urge you to make it count.

Published on

Nimrod Ifrach

Nimrod (Nimi) Ifrach, Director of Data Science at Optimove, leading all the onboarding and data integration processes in the North America market. Nimrod has vast experience across different business verticals – retail, subscription, iGaming and more. He delights in turning statistical power into business value, always keeping the data at the center. Nimrod holds a BSc in Industrial Engineering and Management from the Tel Aviv University.