Maybe the most fun aspect of my work is that I have many colleagues from both the online and the offline marketing worlds. So I get to learn and compare the view points of both sides. Sometimes, both sides have different names for the same thing (e.g. ad view through vs. response attribution). Yet, things get really confusing when both sides share a term yet attribute different meanings to it. This is the case when it comes to “Life Time Value”.
Web analysts know all about Life Time Value, don’t we? Clearly, it is a metric. We judge which of our online campaigns or paid keywords are worth their investment by this metric. For this assessment we don’t just consider the returns achieved in the initial visit after the ad clickthrough, not even just the returns from a delayed conversion in a subsequent session, but we sum up all of the subsequent repeat purchases from all of the customers that were initially attracted to our site via that particular ad. So, Life Time Value is one of the most valuable metrics that we use for our decision support.
Offline marketers, especially direct marketers, smile to that. Offline marketers have been at this a bit longer than Internet marketers. They haven’t just started getting serious about metrics seven years ago when our Internet bubble burst. To offline marketers Life Time Value is a call to action! Namely, the call is for maximizing the value of each customer by not just measuring what they have done, but by conducting right-time, event based marketing. Call it direct marketing, call it one-to-one marketing, call it personalization, or call it customer-centricity. Offline marketers aim to get in front of their customers with offers and messages that are so relevant that they feel like a service rather than an interruption. They spring into action across channels in order to maximize each customer’s loyalty, wallet share, and life time repeat purchases. In that last sense, offline marketers use Life Time Value as a metric after all, namely to monitor the outcome of each action that they take so that they can maximize their success.
How do you use Life Time Value?
Hello Akin
Great point. Often within marketing, marketers do use the same terms to mean different things. It’s one of the reasons it can be difficult for marketing teams to agree on metrics and measures. The adoption of marketing solutions such as marketing resource management (MRM) aim to bring marketers together with a standard and shared understanding of key marketing terms and metrics. MRM is just one part of Enterprise Marketing Management solutions (EMM). Marketers are increasingly adopting EMM to fuse online and offline marketing and to deliver a consistently personalized experience across channels for each customer.
Another important aspect of life-time value is predicted LTV. Direct marketers often use predictive modeling to determine which customers are likely to have a high life-time value if the company succeeds in relevant right-time marketing for those individuals. Direct marketers use this predicted metric to determine which customers they should spend more time and money with because the marketing can have the most impact.
Carol
Akin, Carol
Most marketers are not delivering messages that are so relevant they feel like a service. They are interrupting customers instead. An average response rate of 2.61% from over 1,100 campaigns studied by the DMA is testament to this sad fact of marketing life. But it still works well from a marketing ROI perspective, so marketers continue to bombard customers with spam, both off and particularly on-line. Marketers will only move on to other marketing approaches, for example permission marketing or open-source marketing, when it stops working well enough for them to meet their marketing targets. This is normal and to be expected in the evolution of business models.
However customer lifetime value is about much more than just off-line, on-line or even integrated marketing. It is about the big customer picture, not about the smaller marketing picture. It is about all the things that a company, its staff and its partners do that contribute to the economic value of the customer to the company over their lifetime as a customer. For example, a particularly dumb marketing campaign that alienates swathes of customers, produces negative press headlines and aggrieved customer lawsuits will destroy customer lifetime value, even for customers that were not targeted by the campaign. It may even destroy the company. Remember the Hoover free flights fiasco? But deft action by PR, or by customer service may not only recover that customer value but may even increase it still further. Sadly in Hoover’s case it was too little, too late.
Only by taking this systemic, long-term view of what drives the economic value of customers can we hope to use customer lifetime value sensibly. Just looking at the ROI of a marketing campaign without looking at all the other impacts it has is not enough. And Customer lifetime value is only one part of the larger metrics picture. It is a backward-looking financial metric. That is why Kaplan & Norton developed the balanced scorecard in the early 90s to provide a more balanced view of business metrics.
Before we get carried away with micro-managing campaign-driven ROI, let’s look at the real customer lifetime value as part of a balanced scorecard of metrics.
Graham Hill
Graham,
Thanks for elevating the conversation.
In addition to thinking about all aspects of the relationship and interactions that affect value, we need to expand the concept of value as well. It’s not only the value customers bring to us directly through their purchases but also how they influence others as to whether or not to engage with us. With the growth of social networking, and the speed and reach of word-of-mouth marketing, your comments regarding the need to consider the full relationship are of even greater import.
Carol
I couldn’t have put it better myself. Customer lifetime value as currently measured is too simplistic. As you say, the value measure needs to cover not only the future economic profit that customers themselves generate, but also some of the value that they deliver through influencing others. The latter part of the extended value equation is still very much work in progress.
But even this extended view of customer lifetime value is inadequate. It only covers the value of the customer to the company. We also have to look at the other side of the coin too, at the value of the company to the customer. Only by understanding how customers receive value from a company, can it hope to understand how it can create more economic profits from customers through a mutual exchange of value. And guess what, this is still very much work in progress too.
Graham Hill
Beautifully put! So for example, a prospect or customer may participate in a company’s web 2.0 investments, say a viral marketing campaign or a blog. This prospect can be of high life time value (LTV) even if they never purchased anything but simply by engaging other prospects/customers through their participation. Conversely, your largest customer can have a highly negative value if they are dissatisfied and blog to the world about it.
In both cases web 2.0 participation should flow into the LTV model. It is easier to measure precisely which users clicked the button on your web 2.0 site to forward your viral marketing campaign to others. But when it comes to unstructured blog posts there are some AI tools seeking to automatically assign positive/negative points to text. Yet, one can imagine that this technology has quite a challenge to figure out natural language. For that reason, I would recommend adding a feature such as “rate your satisfaction” to the blog so that bloggers can translate their sentiment into a numeric value that we could funnel into the LTV model.
Reading Carol’s and Graham’s comments I dream of a big LCD display over the entrance door of our office. The LCD display shows the current potential LTV of all our customers and the change from yesterday. Of course I can slice & dice the value (just like in our web analytics solutions) so that we can a.) see immediately what caused the change, and b.) segment & act to improve the value.
Akin
The value of a customer’s relationships with others is a new concept and is not at all well understood. We must be careful not to over-extend early research to the whole networked world. Sunil Gupta’s work you quote in the first paragraph – How Do You Value a “Free” Customer? – was a simulation of customers in a multi-sided market where the value accrues to the platform provider, not a real-life example from a conventional one-side market where the value accrues to the selling firm.
And this warning against over-extending simple real-world models (or worse, simulations) to entire business models applies generally. There is so much that we still do not know about conventional customer lifetime value (the sort measured through DCF), let alone real-options valuation for those harder to forecast futures, customer betas to account for customer segment risk and customer portfolio valuation. The value of relationships is a more complex matter entirely. As the work of Barabasi, Watts, Newman and others shows, we are only just starting to scratch the surface of social networks and their impact on customer bahaviour. It is much too early to be applying their early findings as generalisations.
Management is better off sticking to the customer lifetime value basics, like measuring costs accurately, knowing how far forward they can forecast reasonably and how to share the results appropriately with key staff, than worrying about a developing a Sheldonesque customer lifetime driven future.
As a sign hanging in Einstein’s Princeton office put it, “Not everything that counts can be counted, and not everything that can be counted counts”. Not for the time being anyhow.
Graham Hill
Hi Graham,
Always good thoughts from you. For the web analytics practioners that I work with day to day, what counts is that what they can use in their day to day work for achieving better results for their business and for their customers. So the point of my initial post was to say A.) don’t just use LTV to judge what investments are working well, but B.) also think of LTV as something that you can maximize by doing event based marketing with individual customers at the 1:1 level using web analytics as the learning machine on each customer. Then Carol widened the horizon by adding C.) also take into account potential future value of each customer. I know from web analysts at Ford and Volkswagen for example who are doing just that based on models that they run.
Then you elevated the discussion even further but commented that much of what you were mentioning was still work in progress. Do you see any way in which a web analytics practitioner can put your thoughts to work today? How would they go about it? Curious to see what practical advice we can derive form your valuable strategic thinking. Does the current state of affairs permit using your recommendation or would that also be overextending early research for the time being? Either way, they are thought provoking comments for which I am very grateful.
Akin
Akin
You point is a very good one. Interactive CRMers need simple tools to help them decide what works and what doesn’t.
My simple advice is to work through their organisation’s through-the-line customer strategy to decide what strategic interactive initiatives to follow (see the recent McKinsey study on How Businesses are Using Web2.0 for details of general strategic initiatives being planned), but to carry out as many tactical experiments within the strategy to see what works best, i.e. what attracts people to the site, what makes makes them do whatever you are trying to get them to do(!) and ultimately, what drives economic value.
One big strategic challenge is getting locked onto a low-peak on the marketing fitness landscape through over reliance on champion-challenger. This is where experimentation comes in. Another big tactical challenge is making statistical sense of the clickstream noise that interactive CRM so easily generates. This is where keeping things statistically simple comes in.
Some interactive CRM can be measured in a way that contributes to a more rounded picture of customer lifetime value. Interactive CRMers should feed this into the bigger customer lifetime value picture. But a lot of interactive CRM still cannot be reliably measured. Or the measurements still cannot be reliably interpreted. These should not be fed into the bigger customer lifetime value picture. Not yet.
To quote Einstein again, “Make everything as simple as possible, but not simpler”. Interactive CRMers should bear this in mind when thinking about applying customer lifetime value to their work.
Graham Hill
Wow, I love this post: So often you only hear about the “soft tissue” from marketing folks and this post is steeped in analytics. What’s important to me being in customer care software is better understanding CLV so that we can do a better job in optimizing allocation of support resources to people who in the end contribute (indirectly or directly) the most to my customers’ bottom line. However, this is a very slippery slope since CLV metrics generally only assess what the customer is currently doing with the organization rather than their potential, or their influence with known high CLVers, like their parents. That’s why I think it’s ok and actually prudent to differentiate customer service levels based on CLV, as long as the service levels for everybody else are still reasonable and competitive to the marketplace. It goes without saying that customer care provided to every person should be provided by friendly and engaging staff who understand feedback on their performance is faciliated, encouraged and acted upon. Customer loyalty has a big impact on CLV and service and support can often be the differentiator.
Well said Chuck! Sounds like CLV or LTV is a great compass for the service team to ensure looking beyond ones plate. Still tricky to measure there too.
Well said Graham! Especially about the local optimums.