A large investment bank is hoping to create new business
opportunities by providing its highest-tier clients access to
more sophisticated analysis tools. This bank already has dozens,
perhaps hundreds of analysis tools written in Excel by corporate
programmers. Some examples include:
- Collaterized Bond Obligations models
- Complicated Annuity calculators
- Customized pricing applications in Credit Derivatives
However, these tools are only used for internal purposes - and
the underlying Excel files are treated as company confidential
information because they contain proprietary equations and programming.
Java programming is expensive
The bank starts down the path of hiring Java programmers to
convert 6 of the most popular analysis tools into interactive
Web pages. This requires creating Web pages in HTML that replicate
the core user-interface of the Excel applications, and to write a Java
Web-backend on the server that replicates all the calculations that
are done within Excel. Due to the complexity of the Excel
spreadsheets, as well as the added overhead of creating HTML pages
for a client-server model, the company Java architects predict that
each of the 6 spreadsheets will require approximately 4 months to
rewrite as Java applications with HTML user-interfaces. In addition,
there is a significant risk of programming errors resulting in buggy
conversions -- and the bank does not feel entirely comfortable about
exposing potentially incorrect trading-analysis tools to its clients.
XL2J provides immediate productivity gains
Instead of hiring the Java programmers to convert the Excel
spreadsheets, the bank decides to try a new application called
XL2J that promises to automate this effort. To everybody's astonishment,
XL2J performs the conversions automatically. Each Excel application
takes less than a day to publish on the Web site. Furthermore, XL2J
provides a full regression test harness that allows the bank to
double-check the correctness of the conversion from Excel to Java.
The cost savings are significant
Not only did the bank save 4 months of programming time, it also
saved a lot of money. For converting 6 complex spreadsheets, the
bank calculated its savings as:
6 applications
x 4 months each
x 200 hours/month
x $120/hour/programmer
$576,000 saved
New opportunities for the future
The bank has realized that XL2J can fundamentally change how
its IT department uses Microsoft Excel - not only in client-facing
applications, but also in internal applications. Specifically, the
bank is working on a risk-management initiative to quantify and
manage risk across many trading desks. Each group stores
trading information in Microsoft Excel files, and there is no easy
mechanism for accessing or querying this information from a
centralized computer. However, using XL2J it is possible to
automatically convert the decentralized Excel repositories into
Web-based Java objects with XML/J2EE interfaces. While this
project has only begun, the bank is certain that it will replicate
the success story above but in even greater proportions.
