4 key considerations when building a lifetime value model

Understanding the lifetime value (LTV) of your customer is one of the most important capabilities your Big Data programme can, and should, offer.

Simply put, you cannot have a meaningful conversation about business threats or opportunities in a customer-centric context without a detailed understanding of what the long-term profitability of your relationships are. 

In fact, I would argue it makes little sense to advance a Big Data road map focused on all the best data science and technology has to offer without doing the heavy lifting (and thinking) required in developing customer lifetime value metrics and implementing a plan to embed the required vocabulary into business processes.

Customer lifetime value is the net profit attributed to a customer relationship over the predicted tenure of the relationship with your organisation.

It is important because it clearly represents how each and every customer relationship – more broadly speaking, every reader or consumer relationship – contributes to the overall profitability of the business.

It is important because it allows a business to understand the long-term impact of decisions that alter the cost or revenue structures of the business.

It is important because it provides clarity and focus regarding the significance of having sustainable reader relationships and how much they contribute to the bottom line.

And, it’s a metric you can take to the bank (literally).

No self-respecting customer-centric business should be able to live without it.

There are several LTV models to consider working with but essentially they all include the same fundamental inputs:

  • Revenue: recurring revenue per customer relationship.

  • Cost: one-time and recurring costs per customer relationship.

  • Tenure: expected tenure of the relationship.

In general terms, the model is intuitive: It calculates how much you revenue you generate, the cost necessary to acquire and maintain the relationship, and the expected tenure against which those revenues and costs will be realised and incurred.

When you set out to build an LTV model, it is worth knowing from the beginning that, more so than other data exercises, LTV is an exercise in building a new corporate culture. With that in mind, these are some things to consider:

  1. Start high level and work towards granular. You can start with a generic aggregate LTV calculation, but eventually you will have to get granular to drive its adoption and truly evolve the business culture.

    A smart thing to do is to develop multiple LTV models for various customer segments. For example, you would calculate the LTV of the average customer with a product profile of X versus customers with product profile Y and so on.

    By using this product combination approach, you will gain a material understanding of the long-term profitability in maintaining or migrating customers of varying profiles, and you will be able to anticipate the direct financial impact these decisions and strategies will have on the business.

    This kind of insight will likely also lead to the development of new product or service offerings to protect or maximise profitability. It might also lead you to abandon existing product profiles where the long-term value simply isn’t acceptable.

    Very advanced LTV-focused organisations invest in individual level LTV scoring (where scores are assigned to each and every customer). Such firms are able to leverage this understanding across all their outbound and inbound channels to differentiate service, messaging, and pricing models, and even leverage the insight within the product experience itself.

  2. It will require a combination of basic arithmetic and advanced data mining. Sourcing and calculating all the various cost- and revenue-based inputs within the LTV model can be grueling work. And while things can get complicated quickly, these challenges do not exactly represent a “Big Data” use case.

    However, the cornerstone of an LTV model is the expected tenure of the customer relationship. Make no mistake, this will require advanced analytics practices and large data sets more amenable to Big Data skills and tool sets.

    There are several statistical methods that can be used to determine the expected tenure of the relationship, but essentially they all involve analysing past behaviour to anticipate future behaviour by looking at the sets of patterns and constraints that circumscribe the customer relationship.

    If the cost and revenue structure represents the “value” in lifetime value, then this is where the data scientists earn their keep to deliver the “lifetime” portion of the equation.

  3. It’s not just about “paying” customers. Things get very interesting when consumers (non-paying customers) are accounted for in LTV modeling.

    A good example of this sort of relationship is in the digital environment. Even non-subscriber (anonymous or registered) Web site visitors contribute lifetime value to the organisation.

    Costs are still incurred to service the readers and revenues are still generated through advertising – these are in fact essential relationships media organisations must understand to fully evaluate and optimise their business models: scale versus engagement; advertising versus subscription.

    You cannot truly have an informed and intelligent conversation about these things without first understanding the lifetime value contribution for each type of relationship.

  4. Practical use cases are critical. Do you want to understand the net result of a price increase? Do you understand the balance between the revenue gained versus an uptick in churn? Do you want to understand the impact of an investment in customer loyalty meant to extend the tenure of a customer relationship?

    Do you want to know which product profiles are more profitable than others, so that focused migration paths are established in your marketing campaigns? In strategic planning mode and wondering if there are areas of high return you should be doubling-down on?

    These are just a few examples of fundamental questions LTV analytics can answer. Ensuring the philosophy is adopted and leveraged appropriately will require some smart thinking to determine where to focus on first, gain credibility, and add practical value that just makes sense.

So, why is this as much an exercise in building a new corporate culture as it is something about a bunch of numbers that just need to be crunched? The answer is that this insight should drive decisions ranging from routine marketing to strategic pricing decisions as well as product development and the sun setting on things that just aren’t worth it anymore.

While this intimate understanding of what a reader is worth can be the catalyst for new, creative ways to generate increased engagement and optimise cost structures within the business, there is a flip side.

This level of transparency and objectivity can be uncomfortable for stakeholders to come to terms with – in some cases it will even tell you to kill the sacred cow. If you have built the model correctly, the evidence is going to be very compelling.

It is critical, then, for any LTV effort to be accompanied by a well thought out roll-out plan and for it to have executive sponsorship at the highest level, including your CFO’s support, which is essential if you are going to succeed.

About Greg Doufas

By continuing to browse or by clicking “ACCEPT,” you agree to the storing of cookies on your device to enhance your site experience. To learn more about how we use cookies, please see our privacy policy.