CRM Segmentation for the Financial Services Industry

Liz wilson
Member

Posted 29-Jul-2005 01:59 AM
Graham can you help? I am looking to define customer value for a direct-sale financial services company. I have had a stab myself but would really appreciate advice from someone who is experienced in this area and has had positive results.

Thanks

Vladimir Dimitroff
Member

Posted 10-Aug-2005 08:44 AM
Hi, Liz-

I assume you are looking at more than just a definition of customer value, perhaps you want to create a segmentation scheme that will allow your company to differentiate customer treatment and resource allocation.

You start with a definition, of course, and there are a number of possibilities:

Customer Value, in its simplest form, can be the current or historic revenue/turnover with a particular customer. In financial services this could be the value of current policies (isurance) or loan/mortgage value (credit providers), or assets under management (investment services). In other words, a volumetric measure of 'how much business this customer is giving us'.

The difference between present (for the current period) or historic (since s/he became a customer) value is that the latter, if we have data to calculate it, is more accurate. It provides some idea of future value (despite the unreliable nature of linear extrapolations) to base decisions on.

With a little more sophistication, a present or historic value can be made even more accurate by deducting any known costs (cost of product, acquisition and retention costs, cost to serve). The true value of customers thus reflects their profitability, i.e. shareholder value. Individual costs per-customer are not always easy to obtain but where available, make the value score far more accurate.

An example can be the comparison of two 'low value' customers of a bank: a university student and a retired pensioner. They have equal low amounts in their accounts and should be scored the same. However, the pensioner makes frequent calls to the Call Centre with inquiries about recent transactions and perceived statement errors—each minute of call centre time costing more than the interest the bank makes on her sums. She visits the branch in person, always dealing face to face with the teller to make deposits or withdraw cash. The student, on the other hand, only uses ATMs and the Internet, never visits branches or bothers the Call Centre. Because of the 'cost to serve' the bank makes a small profit on the student and loses money on the pensioner, despite seemingly equal money sums.

Finally, companies with advanced segmentation capabilities do not trust linear extrapolations or the past as a sign of the future. They have sophisticated predictive models for future value. Often called LTV (lifetime value), sometimes CLV (customer lifetime value), this form takes into account the expected duration of the relationship with the customer. Within this duration, any future transactions and revenue streams (including broader portfolios, cross-sells and up-sells) are included as well as any known costs, and all this is discounted as net present value (NPV).

Back to our example with the two bank customers, the growth potential of the pensioner is obviously limited, as is the expected duration—while with the student, there is both a longer duration and growth potential—driving further apart their lifetime value.

Knowing each customer's value is only part of the task: for meaningful segmentation you need to look at the distribution of this customer value across your 'population'—an analysis typically resulting in a Pareto curve. Yes, it's that same guy with the 80:20 rule—in this case we have a continuous curve showing the portion of customers contributing each portion of value.

An important decision is into how many parts to split the customer base, and where to place the cut-off lines. Remember that you are doing this to optimise your resources while providing relevant levels of service. Segments have to be managed—sold and marketed to, serviced etc. To do this differently, in a profitable way with finite resources, it is wise to divide by comparable chunks of aggregate value. Thus, a lower segment will contain a larger number of customers, but they can be served (at their relevant level) with relatively less effort and investment. The fewer customers in the higher segment (who bring the same value despite their low number) merit higher service levels and generally more attention. At the top of the segmentation pyramid is a small but important segment who may even be treated one-to-one by personal representatives (in financial services—the equivalent of Private Banking), while the bottom can receive sevice through low-cost self-service channels (IVR, Web) and be addressed only through mass communication media.

And this is not all: you have to develop processes and procedures with modifications for each segment. You need to recruit and train frontline staff with different skills for each segment. Your campaigns will eventually be run with variations or even completely separately for each segment. Your information systems will contain business rules that trigger different processing based on the value segment fo the customer. Financial planning and performance reporting becomes 'per-segment', etc. etc.

Only when the entire business is run differently for each value group you can experience the benefits of value segmentation.

Which is not the end of the journey—come to the next session (on some forum or other) to hear about Needs sgmentation Smile

Vladimir Dimitroff
PRISM Consulting (UK)
vdimitroff@prism-gb.com

Exchange Synergism
Member

Posted 11-Aug-2005 12:01 PM
I sugggest taking a look at this article series on the Canadain Marketing Association's website.
http://www.the-cma.org/council/customerprofitability.cfm

Also feel free to ask me, I have implemented Customer Value Measurement at FIs throughout Canada and the USA over the last dozen years.

David McNab
David@ExchangeSynergism.com

David McNab
Exchange Synergism Ltd.
http://www.mscforum.com

Graham Hill
Guru
Member

Posted 26-Aug-2005 08:35 AM
Liz

Apologies for the delay in responding to your question. I have been on an extended holiday and have had limited access to the internet.

The question you ask is on the minds of many of CRMGuru's members. However, customer valuation is a relatively difficult thing to do well enough to gain real insight that you can use. The difficulty is not in the concept of customer value itself, but in getting enough of the right raw data to calculate customer value and then in using customer value SENSIBLY to inform everyday business decisions. I have heard far too many horror stories about completely inappropriate uses of customer value information.

In general I agree with Vladimir's response. The key thing is to get started with what you have. Most Financial Services Institutions (FSIs) have detailed information about product sales volumes, product revenues and sometimes product profits too. The product cost side is less often available, particularly the detail that typically comes from carrying out Activity Based Costing analysis.

The first challenge is in collating product revenue and cost information from across different divisions of the FSI and in translating whatever information is available into the equivalent customer revenue and cost information. This is the minimum required to calculate a simple, historical customer value.

Once you have a historical customer value, you can, as Vladimir suggests, think about extrapolating that forward to calculate a customer lifetime value. This requires customer retention data that shows how long customers stay with the FSI. The simplest extrapolation is based upon customers having the same annual value looking forward over however many years the customer stays. This is the minimum required to calculate a simple, historically-based customer lifetime value.

Unless there are contractual or other reasons why customers should be locked-in to an FSI, I would be careful about using retention periods more than 5 years, (even though FSI customers are often poor at moving their accounts to other FSIs). So much can happen in the FSI market 5 years!

As you gain more experience, you can improve the accuracy of the calculation by dividing the customers into segments to represent their different behaviour depending upon when they were acquired, the growing number of products that they typically use, the channels they use and so forth, house-holding and so forth.

You can also start to bring in an additional value uplift for those customers active in recommending the FSI to other customers. This is often forgotten in customer valuation and can be considerable (for example, studies into the importance of marketing vs. word of mouth recommendation of over 100 new product introductions shows that whereas mass marketing has a multiplication factor of 3%, word of mouth recommendation has an additional factor of 38%).

The quid pro quo for this increased accuracy in valuing customers is the additional cost of capturing and incorporating the information, and the increasing margin of error of the resulting customer valuation that the additional error-prone information brings to it (this is inaccurate business information we are talking about!).

More advanced customer valuation looks at future growth scenarios and real options to estimate future value more accurately. And at "customer value at risk" to understand what factors can destroy future customer value, for example, the lack of response to FSI marketing due to the dominance of "junk" in FSI-originated mail.

I recommend you look at the following books to get you started on thinking properly about customer valuation. They contain much more information than this or any article can possibly provide and provide a robust, tested framework to think through customer valuation for your own FSI. They are by far the best books on customer valuation that I have seen to date.

Robert C. Blattberg et al, (2001)
"Customer Equity: Building and Managing Relationships as Valuable Assets"
Harvard Business School Press

Donald Lehmann and Sunil Gupta (2005)
"Managing Your Customers as Investments: Are You Spending More on Your Customers Than They Are Worth?"
Wharton School Publishing

Let me now if you want more specific details.

Graham Hill
Independent CRM Consultant