Living with the status quo means overlooking contradictions and failures, but sometimes, even oblivion cant shield us from a view of chaos. It was arriving from sunny 70-degree Philadelphia into blustery 30-degree Milwaukee last week that started me thinking again of how quickly things devolve. But it was driving past the donut shop into the company lot and watching automatic lawn sprinklers spritz through snow flurries that got me depressed about how quickly things fall apart. I could already envision the frozen water lines and spreading ice sheets that would come next. I didnt stop to tell the storeowner about the problem. He was already looking at the scene from his own window and wasnt taking any action I could see.
I have to admit I was expecting something similar surrounding a much more important event - capturing the vote in this weeks midterm elections. The single most important data exercise in a democracy is an old, straightforward proposition: capture and tabulate voter data entry accurately, secure the identity of the voter, audit for veracity and safeguard against tampering. This process defines democracy and going into this election appeared to be as mistrusted as at any point in our history. But despite the fact that some 80 percent of the nation voted electronically, one-third of those for the first time, there is so far less upheaval over the polling process than in years past. Its still early, but by and large, the results are in and have so far been accepted. Pivotal outcomes remain in Virginia senatorial and in a few house races largely because the outcome is closer than margins allow. As one pundit paraphrased a tourism slogan last night, Virginia is for lawyers.
While this is worthy of a sigh of relief, lets not attribute the success to electronic polling, where a number of miscues occurred, some that were predictably preposterous. Thousands of calls came into watchdog groups complaining of polls that opened late because of malfunctioning machines. ABC News reported that many states, including California, Florida, Illinois, Iowa, Indiana, New Jersey, Ohio, Pennsylvania, Texas and Utah, reported electronic voting machines that were not working properly. Some simply failed to turn on, others failed to scan ballots, and software failed to function properly. The Chicago Tribune reported that some Denver voters waited 300 deep because of election workers unfamiliarity with the machines.
New symptoms, same disease. Three general paths to incorrect voting outcomes have always existed: poorly created ballots that lead to voters file mistakenly (entry error); mechanical and electronic errors, and messy tabulation (system error); and tampering (criminal malfeasance). So far, knock on wood, the problems appear to be more related to managerial and technical errors than to political fiddling. The breakdowns reported are numerically comparable to past elections. That is to say, so far, were not worse off than we were before.
But look at the higher cause that led us to believe electronic balloting would cure many ills after the 2000 elections, and you quickly run into two fateful assumptions. The first is that technology would deliver a higher level of precision where accountability was demanded. If those elections had not been so close we wouldnt be where we are now, says Bob Charette of ITHABI Associates, a frequent contributor to BI Review. Wed have said, well slowly get better with electronic voting and work out the bugs and use it when it works and has the audit trails. Youre missing that the data is corrupted, that more bad data doesnt make things better. Theres déjà vu in this: As a site( www .cs.uiowa.edu/ ~jones/voting/pictures/)
run by the University of Iowa department of computer science points out, the original 19 th
Century mechanical polling machines were among the most advanced developments of their time, and that didnt help either.
Bobs evolutionary point intersects with the second assumption, that wed be able to execute a critical project on a due date across fragmented data sets owned by different organizations. Remember that states control elections, including those with federal outcomes. States buy the equipment, train the people, and set the expected level of data quality. In the corporate world hardly anyone trusts their data and they cross check and go back, said Mark Piro of Affusion Inc., another BI Review contributor. This is a one shot deal, one event that requires an exact outcome quickly. No matter how much we buy technology, our intellect tells us whether to trust the people and process. The more we dig into the technology, the more well freak out about it.
Process, governance and training are words to describe the same balloting pitfalls weve always had. Then again, electronic balloting also provides new ways to circumvent and de-legitimize the political process. Now, a power failure can shut down a precinct and cancel out a whole set of voters leaning a certain way. Early balloting in Tennessee shut down mid-day because a processing hub couldnt handle more than 10,000 votes. The Texas secretary of state is being sued because of electronic voting machines that leave no paper trail. As a writer for The Nation noted, if an ATM can provide a receipt, why cant a voting machine? By comparison, slot machines provide a veritable track record of information.
We could chalk up the mess to fragmented bureaucratic ineptitude and ignorance to the facts of data quality and secure audit trails, but the federal government forcefully drew the line in these very areas as they related to corporate scandals. It hasnt come easily, but for all the reluctant time, money and effort spent complying with Sarbanes-Oxley legislation, four years later, enterprises are starting to reap the associated business benefits of data rigor. I was skeptical when this pitch was made early on because it was tied to unrealistic promises and timelines, and it certainly could have been done differently and more efficiently. But as many corporations can relate, that was then. No pain, no gain. Today, we are seeing examples of enforced data stewardship and accountability paying off elsewhere in the enterprise. You could say that regulation in some ways pumped up the current market for business intelligence, data integration and customer/product data management; I think it would be more difficult to argue the opposite.
Many politicians and bureaucrats however dont see government in the same light. Theres more than a little irony in the fact that Treasury Secretary Henry Paulsons pet project this year is to roll back key provisions of Sarbanes-Oxley because he believes they are too onerous on business. Certainly there should be wiggle room to address ill-conceived parts of that bill, but having come this far, the secretarys proposition is to steeply decrease accountability, cut back on prosecutions for malfeasance and surely earn the ire of the public one more time. As it applies to elections, the standards are already lower and the secretary must know that Democrats and Republicans have secured entire staffs of law firms to challenge every close voting outcome that goes against them.
In a better wor d, the government could benefit from the same medicine it prescribed in much larger doses to business. Ideally they would have the ability to have identity protected but validated so they could do a better job of what exit polls do now, to help capture the demographic and compare it to the vote, says Robert Abate, principal at RCG Information Technology. That way, politicians could better understand whom they are serving. The funny thing is, I dont think any of us understands what happens to data when we vote.
As the corporate world has learned, data discipline doesnt come easily but brings with it great rewards. It may be that federalization and standardization of the voting process is the ultimate answer to ongoing balloting problems, and it wont sit well with protectors of a flawed status quo. But similarly inclined executives in power already behave differently because the business parallels have already documented and proven the argument for rigor and diligence in the enterprise approach. In any case, technology by itself wont be the answer, no matter how much we hope it is.