Were really good at building IT systems that remember what we told them: “If A today, then A tomorrow.” Were fairly good at building systems that can combine facts according to rules: “If A and B,” then C.” But were terrible at building systems that will tell us, “I dont believe you when you tell me A,” or systems that can tell us that “A doesnt matter unless B”—let alone systems that are usefully able to ask us, “Would you like to know more about Z?”
Until we can build those more skeptical—or even more speculative—enterprise applications, well merely drown ourselves in data that doesnt actually tell us anything we didnt already know—while virtually assuring the future recurrence of fiascos like the dot-com implosion (a result of people over-analyzing numbers that were, on their face, unbelievable) or the WorldCom debacle (a result of people lying to their systems, which faithfully repeated the lies).
In his 1968 novel “Stand on Zanzibar”, John Brunner envisioned what we would today call an ERP system based on high-level machine intelligence: In one crucial scene, the system refuses to make forecasts of a major projects outcome because the given data, the system asserts, are absurd. No one knows how to resolve the problem until someone with more insight asks, in effect, “What additional input would make it possible for you to believe what weve told you so far?” The answer not only enables them to go forward but also opens up promising new avenues for research into a previously unasked question.
If were developing source code, we can use a tool like MKS Code Integrity to compare defect rates against norms. If were forecasting sales, we can use a product like Business Forecast Systems Forecast Pro to give us more than just a simple trend line. If were developing an internal application, we should look for ways to put entered data into appropriate context; if were developing applications for sale to others, we can offer high-end network-subscription versions that provide a dynamic context from ongoing research.
What we want to avoid, at all costs, is the kind of error satirized by Stan Kelly-Bootle in 1995, when he wrote, “The PC soon blossomed as the Uzi of creative corporate accounting. The What-If moved to Why-Not, indicting the spreadsheet as the chief culprit in the 1980s S&L scandal.” We can build applications that produce whatever output we desire; we have to work harder, in realms of both technology and corporate culture, if what we want is the truth.
Tell me how applications have given you the right answer to the wrong question.