Presentation tools don't skew data, people do.
The job of information technology is to inform. This means that IT professionals must be involved in the debate over how IT tools shape the presentation of data and steer decision-making processes.
The coming years IT budgets are sure to see proposals for the cinematic video and other multimedia technologies that were prominent on the radar at last months Comdex, but those budgets should also include end-user training in how to present data and state recommendations; without those skills, adding bandwidth to enterprise collaboration tools is a proposition with low or even negative returns on investment.
Im not condemning the current state of the art in presentation tools, although there is a nice little punditry storm taking place these days on the subject of PowerPoint. Infographics guru Edward Tufte has analyzed, in particular, a PowerPoint
chart from the briefings that assessed the risk to the space shuttle orbiter Columbia; you can see his comments at www.edwardtufte.com
by clicking on the hyperlink labeled, "ET on Columbia Evidence: Analysis of a Key Slide."
Tuftes core points have merit: The space limitations of PowerPoint promote a telegraphic style that paves the way for ambiguity. The chart in question also makes the common mistake of betraying a presenters thinking sequentially, using the title of the slide as an introduction. Better would be to display that thinking hierarchically, showing first the conclusion and then the support for it.
Bad presentations result from people learning to write with a model of "topic sentence, body, conclusion," instead of a journalistic model of "lead (conclusion), significance, supporting details."
Additionally, the person who prepared the Boeing chart in Tuftes analysis probably ran out of space; otherwise, the presenter would most likely have finished the point that only begins to be made in the fourth and last of the slides third-level bullets. Talking about a mathematical model of damage done to an orbiters wing by falling chunks of insulation, the chart concludes with numbers that show the Columbias reality being 640 times worse than the models assumptionsbut the chart never actually states this ratio.
Getting to the bottom of that chart, literally as well as figuratively, requires an uncommon determination to seek out bad news. Presentations like this are a weakness in NASA and in many other organizations.
I part company with Tufte when he blames this kind of sloppiness on PowerPoint itself. He compares it to a drug with "frequent, serious side effects" of inducing stupidity, wasting time and degrading "the quality and credibility of communication." Hes wrong. PowerPoint doesnt corrupt; it concentrates. If you have something useful to say, it helps you say it in a more effective way; if youre ignorant or confused, PowerPoint makes it more obvious, but only to an audience that isnt in the same condition. Moreover, Id argue that its easier to be deliberately obscure, and to cover ones self against every possible outcome, in a document laden with footnotes and appendices than it is in a 40-word chart.
As I say, its important for IT pros to take part in discussing the effects of the tools they provide because media dont just transmit facts; they alter both selection and emphasis, creating different realities in the process.
As to selection, we can see every day how TV news looks for images while print media seek out details and context: A car chase doesnt play as well on the front page as it does on TV, and the Pentagon Papers made an absorbing book but would have been deadly on the tube. As to emphasis, the same events covered by different media result in different perceptions. The Kennedy-Nixon debates were won by Kennedy, according to those who watched them, but by Nixon, according to radio listeners.
When enterprise communications can include a video clip as easily as they can include a table or a graph, people may fall into the trap of presenting data in the way their tools encourage them. Its part of the enterprise IT charter and should be part of its budget, to train users in following one anothers thinkingnot just softwares instructions.
Technology Editor Peter Coffees e-mail address is firstname.lastname@example.org.