J. Rothman, Ph.D.
||Michael Rothman & Associates, LLC
President & Founder
MR&A is dedicated to creating intellectual capital in
an area, which may be broadly referred to as "data analysis."
This has taken the form of: traditional consulting, the filing
of patents, and joint ventures in creating, testing, and implementing
new products and designs.
Over the past two years we have been working with a leading
pharmaceutical company to identify the factors which determine
the shape of the sales growth curve for drug introduction.
The shape of the curve is a key element in determining cash
flow and profitability. We also created premier
explorer for analysts and executives to be able to
browse, to compare, and to understand product characteristics.
A shorter term project involved executive consulting services for a credit card association to review proposals for hardware, software, and consulting services. They were building a 10TB Data Warehouse. We also provided a critical review of system structure, design, contracts and cost.
In another project we evaluated the assets of a failing dot com in order for a bank to understand which pieces were worth purchasing, and how they might duplicate or compliment existing bank assets.
Pro bono work was completed for a guide dog school. We analyzed puppy mortality data based on birth order, birth weight, and labor time. The results changed birthing policies at the school and are expected to be published in an appropriate veterinary journal.
Patent filed for 2 new financial services products: Blend Card and CD Card.
||First USA Bank, Wilmington, Delaware
Senior Vice President, Strategic Information
Key thought leader for the Bank. In this capacity worked
with the CEO and EVP of Marketing to establish an approach
to interacting with customers, entitled the "Conversation
with the Customer." This will result in consistent,
non-duplicative, optimized, learning interactions with customers.
Designed and implemented a "Preference Database"
(patented), which contains detailed consumer profiles of 50
million First USA customer accounts. Every two days this
database is automatically refreshed, and scans transactions
with 15,000 rules, to determine our customers’ interests in
any of 2000 preferences. This database also acts to consolidate
data from multiple data sources and to transform it into a
simple to use, uniform information source for marketing, risk,
collections, retention, and other Bank functions.
Created a "Match Engine" to effectively determine
the appropriate Bank action with the preferences of our customers.
This allows for a customer-focus rather than a product focus
in Bank communication.
Designed and implemented a database access tool, "Profiler,"
which enables large numbers of users to access databases via
the First USA Intranet, with drag & drop query capability.
The tool generates SQL against a sample of First USA's portfolio,
allowing users to bring together data elements regarding credit,
account information, preferences, triggers, segments, and
card type with response times of less than two minutes.
||IBM Consulting Group
Manager and Executive Consultant, Global Business Intelligence
Consulted to most of the major banks and financial institutions.
In 1998 the practice was the most successful in the IBM Consulting
Group. Focus was on business problem solving based on neural
and statistical methods. Special emphasis was placed on transforming
mathematical results into meaningful business strategies.
Specific projects included:
Also, has been an invited speaker on business applications of
neural technology at IEEE symposia. Spoke about emerging technology
at the MIT Sloan School of Business. Presented a paper at the
22nd International Conference on Very large Databases, "Applying
Data Mining Techniques for Effective Health Insurance Administration."
patterns of spending behavior in credit card transactions which
could be related to likelihood to default or attrite (for a
- Analyzing patterns of credit line utilization as relates
to bankruptcy and write-off (for a major credit card issuer).
This discovery used a method, "Time Series Segmentation,"
developed by the practice.
patterns of revolving credit utilization (for a major credit
bureau) that revealed persistent characteristics of consumer
credit usage, and significant opportunities for Banks to refine
their marketing efforts
- Built the major components of a credit card transaction-based
targeted marketing system for a major credit card association.
We developed specialized segmentation models (overall, regional,
travel, retail, brand-driven,...) and scored over 100
million cardholders. The system generated significant lifts
in consumer response to targeted campaigns.
In the non-financial area:
- Analysis of sales tax payments for a state government,
which identified "types" of payers. This information
could be used to direct enforcement efforts.
- Discovery of patterns of food purchasing behavior (for a
major packaged goods company) involving analysis of Nielsen
- Market basket analysis (for a major grocery chain) revealing types of shopping
- Analysis of video tape sales (for a major entertainment
company) revealing geographic and demographic preferences
for different types of videos (e.g. live action, adventure,
- Analysis of patterns of medical test requests from 10,000
GPs in Australia to create better monitoring programs for
appropriate medical practice (for the Australian government).
In development customer gained an immediate $1 million return
in avoided costs.
- Developed algorithms later incorporated into IBM's Intelligent
IBM Microelectronics Division
Senior Engineer / Advisory Engineer, Failure analysis
Led a team, which uncovered both manufacturing misalignments
and fundamental design flaws with an IBM cache memory chip
for the 390 mainframe. As a member of the failure analysis
group, developed models, which revealed process sensitivities,
and worked with a circuit designer to link these a susceptible
circuit. This discovery led to a doubling of chip yields,
the creation of a re-designed chip, and its implementation
just in time to prevent a forced 10% reduction in the speed
of the machine. The machine was a key product at IBM, responsible
for billions of dollars of sales.
IBM T. J. Watson Research Laboratory
Advisory Engineer, Computer Science / Artificial Intelligence
Investigated neural network applications to decision-making
algorithms, wrote LISP code to implement the system, led IBM
meetings on neural networks, spoke to IBM customers and at
IEEE meetings on neural networks
IBM General Technology Division
Advisory Engineer / Staff Engineer / Senior Associate Engineer,
Led a joint engineering / IS team to develop algorithms
to control an automated plating line. This work was published
in the IEEE Journal of Systems, Man and Cybernetics and had
as a key design element the empowering of manufacturing operators
in the intelligent running of the tools. It was implemented
in the IBM plant in Hopewell Junction, NY and increased throughput
and product quality. This was followed by implementation at
IBM manufacturing plants in Bromont, Canada, and Manassas,
Earlier tasks included working with the Process Engineering,
IS, and Equipment Engineering groups to see that issues of
chemistry was being communicated correctly to IS and Equipment
University of Michigan
Ann Arbor, MI
Doctor of Philosophy in Chemistry, specialization in quantum
chemistry (physical chemistry), Kodak Fellowship, Rackham
Fellowship, other fellowships and honors, published in American
Journal of Physical Chemistry, Journal of Chemical Physics,
Journal of Computation Chemistry
Sc.M. in chemistry, Sc.B. in chemistry with honors