Scribd
Upload
62 views2 pages

CRM Mining Returns Paper

Data mining tools to support customer relationship management (CRM) efforts have been growing in use by firms. However, quantifying returns on these efforts is challenging due to small sample sizes and companies' reluctance to disclose strategic information. Nevertheless, some examples show potential returns: Sprint reported saving over $1 million in costs and increasing customer retention and upselling using SAS CRM services. Bank of America identified $4.8 million in savings over two years from a SAS credit risk model, a 400% return. Data mining technology helped Liverpool Victoria increase campaign response rates by 35%. These examples provide an idea of returns possible from CRM systems using data mining analytics.

Uploaded by

Shantharam Kamath
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
62 views2 pages

CRM Mining Returns Paper

Data mining tools to support customer relationship management (CRM) efforts have been growing in use by firms. However, quantifying returns on these efforts is challenging due to small sample sizes and companies' reluctance to disclose strategic information. Nevertheless, some examples show potential returns: Sprint reported saving over $1 million in costs and increasing customer retention and upselling using SAS CRM services. Bank of America identified $4.8 million in savings over two years from a SAS credit risk model, a 400% return. Data mining technology helped Liverpool Victoria increase campaign response rates by 35%. These examples provide an idea of returns possible from CRM systems using data mining analytics.

Uploaded by

Shantharam Kamath
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

Predicting Returns from the Use of Data Mining to Support CRM

The number of firms that have successfully used data mining tools to support
CRM efforts is growing. However, quantification of the returns to these efforts is
hampered by the relatively small numbers of organizations who have had such efforts in
place for a sufficient time to assess results, by difficulties in sorting out the causes of
changes in customer response, and by the natural reticence of companies to publicly
disclose this strategic information. Because of these limitations, it is not possible to
establish a typical return to the use of data mining tools in the CRM arena.
Nevertheless, a number of firms have been willing to describe in quantitative terms,
results they have achieved and these results provide a picture of the type of results that
might be expected from the application of data mining tools to CRM issues. A few
recent examples are summarized below.
Recently SAS institute provided CRM services to Sprint with emphasis on
reducing churn. Sprint officials reported requests for information that used to take weeks
to process are now completed in hours. Customer Retention and up-selling are
measurably higher. In addition, Sprint has already (in the first few months) saved more
than $1 million in processing costs. (www.sas.com/news/preleases/042301/news3.html)
Bank of America identified savings of $4.8 million in two years (a 400 percent
return on investment) from use of a credit risk management system provided by SAS
institute consultants and based on statistical and data mining analytics. This model was
used to allow Bank of America to find eligible low-income and minority customers to
ensure B of As compliance with the Fair Housing Act.
(www.sas.com/news/success/bankofamerica.html)
Liverpool Victoria, an English firm providing financial services to over 2 million
customers, has used Quadstones OLAP and data mining technology for campaign
management and customer retention efforts. Early results have included increasing the
response from a campaign by 35 percent. A senior database marketing analyst at
Liverpool Victoria noted that with appropriate use of this technology I have the
opportunity to combine my business instincts with the realities of the data.
(www.quadstone.com/customer/studies/Liverpool.html)
Customerlinx, a knowledge-based marketing services company providing
customer-tailored marketing programs, has used Quadstone technology to improve the
speed of in-campaign modeling. It used to take us between 3 and 4 weeks to accurately
profile a customer. Quadstone takes the time down to 4-5 days. the quick turnaround
time is great and allows us to focus on profitable customers.
(www.quadstone.com/customer/studies/Linx.html)
BMG Direct has used the SAS solution for CRM in an effort to improve customer
loyalty. A BMG executive indicates: We have significantly improved our performance
goals with the telemarketing programsresponse rates rising from 30 percent to more
than 50 percent. (www.sas.com/news/feature/28may01/dmreview.html)

The results cited here relate to specific vendors and systems involving much more
than just data mining tools. They are, however representative of the types of returns that
can be achieved from CRM related systems built around data-mining analytics.
The CDI itself does not have a basis for estimating dollar returns to the customers
for which it has provided CRM related analytics. We do have anecdotal evidence in the
form of both return engagements and actions taken as a result of engagements that
suggests that clients have achieved valuable actionable results form projects completed
here. While the data required to rigorously identify costs savings or revenue
enhancements from a data mining project are rarely available, it is often relatively easy to
establish the gain from using data mining as compared to a making decisions with no
information.
In identification of best customers we have typically found that a cluster of 20
percent or so of customers can be identified which account for 80 percent or more of
sales and/or profitability. In models designed to identify the best half of customers we
have been able to build models that are 80 to 93 percent successful. That is, the odds of
identifying a good customer have been increased from a 50-50 guess to an accuracy of, as
much as, 93 percent.
Another type of problem frequently associated with CRM is identification of
defaulting customers. A recent study of this type by the CDI was able to produce a
measure that successfully flagged over 80 percent of those who actually defaulted as
defaulters, while flagging less than 20 percent of non-defaulters. In other words, Over
eighty percent of the target group was be found within only 20 percent of the customer
population. This type of identification is similar to the problem of identifying the set of
customers most likely to respond to a particular marketing campaign.
The values cited above are fairly typical examples of the ability of data mining
analytic models to identify target groups that are frequently important in CRM. As one
would expect, results will vary depending upon characteristics of the customer base and
the adequacy of the available data. Where measures of the selectivity of current customer
identification techniques are available and costs customer treatments are known, the
measures above can be used to produce approximations of the returns that might be
expected form a successfully data-mining based analytic project.

You might also like