The first e-mail I received from the woman (call her Martina) sounded,
to say the least, distressed. She had just paid thousands of dollars to
have a company integrate her customer data and sales database with her
sales forecasting software, and it didn't work. That is, the software
ran, but the data didn't synchronize correctly--or something; she wasn't
yet sure what was wrong. All she really knew was that the forecasting
software made less accurate forecasts than when she was entering
customer data manually.
Martina works very hard at keeping her business together, and that
includes her business software applications. She doesn't run a big
business, so she doesn't need (and can't afford) an IT department, but
she owns several stores and a raft of software that helps her run her
business. The problem is, she's built her company over the last 10 or 12
years, which means she purchased the software over that period more as
less as needed, and with little or no consideration for integrating the
information that the applications generate. Now she wants to start
comparing and analyzing the data--and the ugly beast of application
integration rears its head.
I call it that because Martina has encountered one of the most serious
and enduring problems in computing. According to a study last year by
The Standish Group International Inc. in West Yarmouth, Mass., 95
percent of all application integration projects fail. That figure sounds
outlandish, even though a similarly gloomy figure puts the number of
successful application development projects at one in six. Still, a 95
percent failure rate is a strong indicator that it is commonplace to
suffer blown budgets, shattered schedules, and delivery (if any) of
software that fails to live up to expectations.
When it comes to computing, Martina is no quivering quail; she has a
background in computer science and spent several years working in the
computer industry. She knows better than most small businesspeople
what's involved in making her applications work together, but it was no
less shocking to see the project fail--and have to deal with the
situation--including a pending lawsuit she instigated against the
software development company.
Perhaps Martina is climbing into the cage with the beast a bit earlier
than many small businesspeople. It's only a little more than 20 years
since the personal computer revolution brought computing to businesses
of all sizes. For many businesses, that continues to mean installing
computers, networks, and using basic software like spreadsheets and word
processing applications. Integrating applications, an issue that
generally hit large corporations early and hard, has been more sporadic
for small businesses. To a certain extent there is satisfaction with the
way computing has made accounting, inventory management, and other
basics of business management more tractable. However, as the demands of
modern business increase and businesspeople become more sophisticated
about computing, the desire increases to have existing applications work
together. In Martina's case, she saw a potential competitive advantage
to using her carefully acquired customer data in new ways, especially
more timely integration of her sales figures and forecasting. She wanted
to be more responsive to trends in her business.
What she didn't expect was how difficult it can be to integrate her
software. In her second e-mail, in which she outlined her case, she
expressed surprise that after "all these years" the computer industry
still couldn't deliver consistent results for application integration. I
replied, somewhat flippantly, that it was a condition of too many kinds
of software meeting too many ill-defined requirements, abetted by too
many "solutions" and the lack of standards.
Quite rightly, she objected: "What you describe is not a natural
landscape. It's been shaped by vendor competition, polluted by deception
and ignorance, and allowed by too many people to continue. Application
integration may be difficult, but it's not quantum mechanics." Martina
comes down on the action side of her convictions; she began to research
the subject of application integration to see if there were other small
businesspeople who suffered from the pandemic incompetence of the
software industry.
As she put it, "That search required a single Google." The search also
revealed to her the name of the Integration Consortium
>www.eaiindustry.org
