 |
Sprint Calls on Crystal Ball to Properly
Provision Customers
|
APPLICATION: Financial analysis of product provisioning.
SUMMARY: Sprint used Crystal Ball to identify customers
who were incorrectly provisioned on residential plan types.
RESULTS: Sprint plans to save $2.8 million by better aligning
customer needs and product offerings.
When Sprint began to offer residential CLEC-based (Competitive
Local Exchange Carrier) phone service in New York City and New York
State in December 1999, the response was extremely enthusiastic.
The introduction of local residential services brought a significant
response, both from the target market segment of residential customers
and from a newer market segment of small business owners. But the
number of small business owners presented Sprint with a dilemma.
Because many of the products Sprint offers in NY are only tariffed
for residential use, Sprint felt it needed to redirect certain customers
to business products that better fit their calling patterns.
Jeff Blase, an Associate Financial Analyst in Sprint's National
Consumer Division was tasked with determining which customers would
be better served on business plans. At Blase's disposal was an abundance
of profile information for small business customers, but he knew
it wouldn't be enough to simply compare those profiles with the
profiles of their New York customers. To convincingly demonstrate
to management which customers should be moved to different products,
he needed to create a spreadsheet model that would independently
forecast the calling patterns of residential CLEC customers. At
the time, Blase was relatively new to Crystal Ball, having only
worked with it for a few months on applications such as discounted
cash flow analysis. Now, he turned to Crystal Ball and Monte Carlo
simulation to forecast calling patterns of residential CLEC customers.
Blase first created a deterministic spreadsheet model that represented
the expected calling patterns of local service residential customers.
He then enhanced the model with ten to fifteen normal and lognormal
Crystal Ball probability distributions. Some of the assumption variables
included number of local calls, local minutes, long-distance minutes,
and custom calling features such as voicemail and caller ID. To
make his distributions as real as possible, Blase derived the parameters
(minimum value, maximum value, and standard deviation) from actual
customer data. He then ran two Monte Carlo simulations of 10,000
trials each, one based on customer data from Sprint, and the other
based on Bell Atlantic customer data obtained from an independent
vendor. Finally, he used the Crystal Ball overlay chart to compare
the actuals against the results of the simulations.
His results displayed a significant difference between the tails
of the fitted and simulated distributions, more than enough to support
the hypothesis that some customers were strong candidates for business
products. By better aligning customer needs and Sprint's product
offering, Blase was able to identify potential cost savings of $2.8
million through more efficient product provisioning.
For Blase, one of the software's strongest selling points is its
ability to view complex ideas and data in a meaningful way. "Being
able to portray probability distributions graphically was absolutely
huge in selling this project's potential to management," he said.
Chad Lander, Blase's immediate supervisor and an avid Crystal Ball
user himself, supports Sprint's coordinated efforts to increase
Crystal Ball's acceptance and use within the company. Lander notes,
"Crystal Ball is truly an invaluable tool for creating persuasive
and rigorous analysis, especially where traditional forms of analysis
leave the analyst and management stranded with no clear direction."
In upcoming projects, Blase and his division will continue to apply
the risk analysis power of Crystal Ball 2000. "My use of Crystal
Ball was far from sophisticated. I'm anxious to explore the extensive
array of tools available in Crystal Ball," he added. In addition
to simulation, he expects to apply the Tornado Chart tool for sensitivity
analyses and the OptQuest optimization tool as part of a flow-through
analysis to monitor the progress of the project.
Visit our
Telecommunications industry page!
|
