Knowing my High Value Customers with AI

As I get to know my High Value Customers, two (2) limited resources present themselves in any business: time and money. Or for some: money and time.

For this reason, it becomes a differentiating strategy to know how to obtain higher profits, while making efficient use of the company's availability and budget resources.

With customers being the beginning and end of the business value chain, it is logical to pursue efficiencies through a deeper and even more personalised knowledge of the people and companies that consume the company's products and services.

To paraphrase the British novelist George Orwell in his novel Animal Farm: "All customers are equal, but some customers are more equal than others". This witty phrase translates a business reality for every company, regardless of its product or business model.

There is a large majority of customers who account for a minority share of sales, with a small group of customers who shoulder the lion's share of revenues.

However, in between, there are also other customer segments whose investment of resources by the company deserves differentiated considerations.

Segmentation allows marketing teams to understand and profile the customer groups of greatest value to the business in order to enable better targeted contact, engagement, deepening and retention strategies.

Likewise, it is necessary to recognise the lower value customer segments and those with intermediate characteristics, as they also require alternative strategies for their attention and satisfaction based on their contribution to the company's sustainability.

What is RFM Segmentation?

Customer segmentation by Recency, Frequency and Monetary, commonly known as RFM segmentationis one of the most reputable and popular ways to identify the most and least interested groups in a company's customer base. The widespread use of RFM segmentation among marketing teams is due to:

It is objective. The way in which stakeholders and specific customers are classified responds to numerical scores that avoid interpretations and facilitate their prioritisation.
It is simple. Designer Stephen Few coined the expression "eloquent simplicity" to describe that which is simple balances functionality and aesthetics to the point of speaking for itself. RFM segmentation results do not require astrophysicists or orbital calculation software to understand and exploit.
It is intuitive. The elaboration of an RFM segmentation, from its input information to its final results, uses profiling criteria in terms of those responsible for the strategy, the marketing teams. In other words, it does not require specialised interpreters to act as intermediaries for its application in the strategy.

Speaking of strategy... a natural question is: How can RFM segmentation help my company's marketing and sales strategy? Well, businesses can find answers to the following questions at a high level of detail:

Who are my best customers, or rather who are the most profitable customers for my business?
Which clients are in imminent danger of defection or cancellation?
Who are the customers with the greatest potential to become the most profitable customers?
Who are the customers that consume the most of my business resources, compared to the revenue they bring in?
Who are the most suitable customers for loyalty or retention campaigns?
Who are the most loyal customers? Keep in mind that loyalty is different from profitability.
Which customer group is most likely to respond positively to the next ... campaign?

How to segment my customers?

Recall that RFM segmentation has promised to be: objective, simple and intuitive. In other words, it can be measured, it can be understood and it can be used.

Each customer is profiled according to the 3 criteria that identify segmentation:

Recency: Find what is the most recent interaction of a customer with my products, services or contact channels. Variations on what recency means depend on the dynamics of the business. The premise is that the more recent I have interacted with the customer, the more likely they are to respond to the business' communication strategies.
Frequency: Determines how frequently the customer interacts with my products, services or contact channels. The assumption is that the higher the frequency of interaction, the more receptive, loyal, sympathetic or dependent the customer is to my brand.
Monetary: Reflects how much money a customer has invested in my products or services over a given period of time. It is a fact that customers with higher spending and therefore higher profitability are ideal candidates for loyalty, deepening or retention strategies. On the other hand, customers with low profitability need a different type of communication to empathise with the brand.

Finally, the RFM segmentation takes into account independent profiling by Recency, Frequency and Monetary to generate a single weighted score.

In this way, each customer is qualified according to their behaviour on these 3 fronts, where we intuitively know that a customer who interacts recently, frequently and with considerable consumption is of greater value to the business than one who visits products and services constantly, but rarely buys.

Under this same logic, several combinations can be generated to profile a customer. Also, by stratifying the scores for each criterion, interest groups or segments are formed, which facilitates the application of communication strategies.

Without going to extremes, contacting with a single message without distinction of audience, or contacting each client with a personalised message. Either extreme ends up being fruitless or unsustainable.

The analysis of considerations, scores, stratifications and clients seems to be too much to develop and above all to update. And it is!

Fortunately, we live in an era where emerging technologies such as artificial intelligence are coming to our aid, as they take care of all stages of the value chain to ensure that the information is available and up to date for use by the company's marketing teams. segmentation of different interest groups will be available and up to date for use by the company's marketing teams.

What does customer segmentation look like?

Clarification 1: Every company is free to give the names to its customer segments that best reflect its corporate culture and values.

Customer segments can be named in such boring ways as: Seg 1, Seg 2, Seg 3, etc. In creative ways like the most successful and unsuccessful series of the Netflix era. Or in geeky ways, in honour of the special missions of the last 60 years.

In the end, the most valuable elements are the segment description and the suggested call to action.

Segment	Activity	Call for action
Breaking Bad	Recent, frequent buyers with the highest consumption.	Candidates for reward strategies, early adopters for new products and brand ambassadors.
Grey's Anatomy	Significant and frequent purchases. They tend to respond to promotions.	Candidates for cross-selling and up-selling strategies. They can provide product reviews.
Game of Thrones	Recent purchases, more than once and for significant amounts.	Candidates for promotional packages or subscriptions.
Lost	Recent large but infrequent purchase.	Candidates for personalised support with a trust-building strategy.
The Walking Dead: World Beyond	Large and frequent purchases. But a long time ago.	They should be brought back. Send personalised communications to return with renewal offers.
Fear the Walking Dead	Last purchase made a long time ago with low consumption.	Candidates for offers of other products or services and special discounts.
The Walking Dead	The most distant, infrequent and low value purchases.	Try to rekindle the flame with personalised communication. If it doesn't work, let it go.

Each company may have more or less number of customer segments as described in the example above, as it depends on the available ones, in terms of age, quality and variety. data available, in terms of age, quality and variety. For this an excellent news: The artificial intelligence also helps to determine the actual number of customer segments interacting with the business.

What data is required for RFM segmentation?

Marketing teams can identify which data sources and information they need to have in terms of the 5 W's: who, where, what, when and how. Customer characteristics are presented to the RFM segmentation solution with demographic attributes that indicate who, where they are located and under what conditions.

In addition, information systems such as CRMs or DMPs can provide the attributes that describe the ways in which the customer interacts with the products or services, and even record the reasons for one-off or recurrent consumption.

Now, let's see in concrete terms what variables are required to implement the RFM segmentation solution based on the 5 "W's".

Question "W".	Description	Variables
Who	These are the variables related to the demographic information of my clients.	Gender, age, education level, profession, income, marital status, dependents, etc.
Where	It tells us where they live, where they work or from where customers interact with the brand.	Urban/rural, local/international, city, department, state, country, postal code, among others.
What	It can be associated with customer behaviours in interacting with products, services and service channels in the past, present and future.	All date fields and types of transactions that the customer has made in the past, which is done in real time. Even services or products reserved for a future date.
When	They refer to commercial seasons that influence customer interaction with products, services or service channels.	Seasonal (Christmas, Love and Friendship, Mother's Day, etc.), working days, weekends, day/night.
How	It relates the ways in which the customer has known and interacted with the products, services or service channels.	Purchasing channel (web, telephone, chat, salesperson, etc.), customer service channel, how did you find out about our services?

Both customer segmentation and its results will also be presented in terms of these variables.

Individual customer scoring and segmentation into interest groups is based on the patterns that the solution manages to find among thousands or even millions of combinations.

Speaking of strategy... a natural question is: how can RFM segmentation help my company's marketing and sales strategy?

Well, in this way you can find the answers to the questions we mentioned in the first part of this blog and start planning marketing strategies for your company. So we invite you to harnessing your data and transforming it into knowledge.

If you have any questions or would like to know more about our solutions, please do not hesitate to contact us!