Customer Lifetime Value modeling (CLTV) is a well intentioned approach to prioritizing a company’s marketing spend against their high value customers. CLTV models look at the overall value of a customer over a number of years, and generally deduct from the revenue the cost of acquiring that customer and the ongoing cost-to-serve. In doing this, a CLTV model should help a company optimize its marketing spend to have the greatest return.
However, the implementation of CLTV models is often so poor that it destroys value and de-optimizes marketing spend, and marketing dollars are spent where they are least able to make a difference. CLTV models also do not adequately capture the complexities of how to value a customer – for instance, a customer who is highly networked and able to be a high volume referrer. But this is potentially complicated math.
To take an example from targeting consumers through Facebook, it is wrong to only target high noise/high edge individuals on the basis that they are well-networked. The ability to get your communications to appear in the newsfeeds of high noise groups is going to be greatly compromised for the very reason that you are competing with all their other edges. Instead of hoping to show up on their newsfeeds, optimize targeting of the sufficient network to make the effort worthwhile.
It is also worth noting some context around product, service and industry life cycle. Let’s think about the evolution in Mobile TelCo and the relationship with potential consumers and commercial strategy. At the growth phase, the acquisition and retention costs were biased towards customer acquisition. Through the late 90s and early 00s, mobile operators were only concerned with getting new customers. No one cared about retention and CLTV. Today, the market is saturated and strategies are all about high value customers and retention.
Overall, the mathematics of CLTV is complicated, but unless you do it right, CLTV models will destroy value – not create it. It seems that for the last decade, I have fought a single-handed battle to address both the misuse and under-use of CLTV. Now, I’m setting the record straight once and for all. You’re with me or against me; this is my 8 point CLTV Manifesto.
1. CLTV is incorrectly used to identify supposedly high lifetime value customers, and creates a self-perpetuating model of actually targeting CURRENT high value customers as opposed to FUTURE potential. Companies take current customer values, slice off those of the highest current value saying they are the best customers and call them high value customers. Hang on. Don’t we think marketing actually works? If we do the right thing to low value customers, maybe we can turn them into a high value customer? CLTV models need to take account of potential marketing effects.
2. The very notion of a high value and a low value customer is actually misleading. The two have much in common – namely that the cost of losing either of them is massive, on the basis of the considerable cost of recruiting a new customer. Low value customers should potentially be seen as MORE valuable because the cost of recruiting them is lower than a (supposedly) higher value customer, and a CLTV model should factor in differential acquisition costs. The true, and far more equal value, of customers (when you think about the cost of replacing them) makes you turn many retention priorities on their head, realizing that the difference between a low value and high value customer is over-stated.
3. In any case, low value and high value customers make up part of the jigsaw puzzle that a company can profitably serve all of. A company should be capable of profitably harvesting from all segments of the market with appropriate products and services for every niche. Supposedly low value customers can contribute considerably to overhead, create scale economies, and ensure an overall enhanced profitability business model.
4. However, CLTV is under-used as a dynamic model output for assessing different scenarios. Many of my clients have thought of CLTV as a fixed number; an input into a marketing scenario. This is wrong. I’ve helped them to re-think of CLTV as something that is a variable number; an outcome, depending on what they do. If you increase retention, your CLTV increases. If you reduce your acquisition costs or cost to serve, your CLTV increases. If you increase your cross-sell, your CLTV increases. All of these can be captured in a dynamic CLTV model. For example, we have seen some great prophet examples in US health insurance, where a focus on reducing acquisition costs has had a profound effect on our client’s success. Thinking of CLTV dynamically can help drive reduced acquisition cost through a business.
5. CLTV can cause some companies to under-value the aggregate segment of low value customers, i.e., a much higher overall value than the small pie of high value customers. Many low value customers outweigh in value a smaller number of high value customers. In other words, there are more low value customers than high value, so the aggregate segment value is higher. Because the individual value is lower, one more purchase per year by each of the low value group, makes a huge percentage difference to aggregate segment value.
6. Marketing to higher value customers will have a far greater proportionate impact than marketing to low value customers. Basic arithmetic would suggest targeting supposedly high value customers (who are actually saturated) with marketing efforts can give a lower ROI than targeting the low value customers, e.g. persuading a low value customer to take just one more policy has a more dramatic impact on their value than a high value customer being cross-sold an additional account. Also, because (as we’ve seen in 5) there are so many more low value customers than high value customers, the aggregate benefit of cross-selling the low value customers can be dramatically greater. As a GROUP, low value customers are worth more than high value customers, and marketing to them can give a far higher ROI for that marketing spend. When I see CLTV models used incorrectly, it drives me mad because it’s destroying so much value. In the majority of cases, CLTV models are used so badly that they justify marketing strategies completely counter to what a correctly applied CLTV model would suggest.
7. CLTV models sometimes incorrectly focus attention of marketers on activities to the high value group. But think of how political elections are decided by a small number of swing voters – this is where the marketing attention needs to be addressed. Similarly, there is no point in marketing to these folks because they’re not going to buy any more insurance, cell phones, mops or whatever. Target your marketing dollars where they can make a difference.
8. The concept of “brand” and CLTV are closely related. Why? Because loyalty most dramatically impacts CLTV. In the markets where CLTV is most often used (TelCo, Financial Services) the level of saturation actually means that “brand” is a CRM concept. Who is the brand re-positioning communications really aimed at? Potential switchers? That’s a pretty small part of the population. It’s the installed base of existing customers we’re trying to prevent churning and trying to up-sell. It’s the power of the brand that helps up-sell from auto insurance to wealth management. Once you realize that in most markets, brand is a CRM concept, the close relationship between brand and CLTV becomes clearer. Brand spending influences CLTV models (and values), but because in so many companies there is a separation between a marketing function (looking after brand and communications) and a commercial team (looking at CRM activities and CLTV type models); this powerful connection is lost. Dynamic CLTV models can be a powerful way of bridging this divide.
In conclusion, CLTV models can do a lot of damage; but if done right, they can be an incredible tool. As I work with my clients to help develop and implement CLTV models, I also see the way in which clear thinking dramatically helps clarify what marketing is meant to achieve. That alone is surprisingly powerful in mobilizing an organization.
We have to begin again with the assumption that marketing actually works, and therefore; the potential (not current) high value customers need to be the focus. CLTV models can then help address issues such as “We have lost this high value customer, how much I can afford to spend to win them back?” or “Which customers should I be targeting to maximize my return on investment?”
In addition to answering these questions, correctly implemented CLTV models can bring logic and evidence to marketing decisions and elegance to the strategy of who we are targeting, why, when and with what offer. CLTV, done right, is one of the most powerful and commercially insightful tools at your disposal, whether you’re a coffee shop, selling insurance or both! But if done wrong, instead is a fast way to wreck marketing effectiveness and destroy a business.