How Much Is Enough?

Beware the technology that makes us more efficientit actually may increase inefficiency and defeat the business purpose.

BY PAUL ROLICH

Call Me Ishmael

Stately, plump Buck Mulligan. It was the best of times, it was the worst of times. I saw the best minds of my generation destroyed by madness. What do we have here? We have some of the best known opening linesthree from novels, one from a poem. Authors and poets work for years to find the perfect opening line, the line that not only will grab their readers attention but immortalize the author. So, after endless agonized hours, do these bits of glory spring forth from a gin- or drug-induced genius insight? Or are they just happenstance? Is inspiration really necessary? What if we could automate the whole process? You knowuse a computer to write some great literature. After all, we live in an age in which we are asked to create computer systems to automate all human tasks.

Automate This Process!

The longest quoted line above is only 60 characters long. To make things easier, we will allow only 30 distinct characters (lower case alpha, space, comma, apostrophe, question mark). That means if we created every possible 60-character opening line, we would have 4.2391288 combinations. Now I know that is a pretty big number. But in present day computing terms, it isnt all that big. Last November, the 40th known Mersenne prime was discovered. (A Mersenne prime is a prime of the form 2P-1. The first Mersenne primes are 3, 7, 31, 127, etc. There are only 40 known Mersenne primes.) That number is 220,996,011 -1. Now that is a big number6,320,430 decimal digits.

So, lets assume I have designed a system that will create a measurably infinite number of opening lines. I have developed a very sophisticated algorithm that eliminates a lot of the meaningless nonsense a true random character generator will produce. We could be, of course, really pedantic about this whole thing. We could populate columns number 1 through 59 with a and then run column 60 through the 30 possibilities. Then kick column 59 over to b and continue the process. Such a process ultimately would produce all possible combinations, but even in geek think it isnt very efficient. There are literally thousands of rules we can apply that will make the task much less onerous. For example: We will not allow any character to be repeated more than twice in a row with the exception of the space character. The opening line of Moby Dick will contain more spaces than alphas. As the finished opening lines come rolling out of the system, they will be fed into a spell checker (for right now, we will deal only with American Englishnot to be jingoistic but practical). They then will be indexed and stored in an Exabyte Raid 5 array. (Exa means 1018 just as Tera means 1012.) Now I am in business. I am ready to market my product.
Some well-healed poet who desires to write the definitive poem for his or her generation will come to me for help. He or she gives me a few key words, and for a fee, I return a reasonably perusable finite selection of opening lines. He or she can spend a month or so sorting through them and pick a winner. If the first batch isnt fruitful, we always can deliver another few hundred thousand.

The uses of this thing actually are much more significant. In fact, we can recreate all past, present, and future English literature with my little machine. In The Library of Babel, Jorge Luis Borges used a similar premise to create the universal library. Each volume in his library contained 410 pages. I am dealing only with 60-character segments, but we theoretically can string those 60-character segments together to create it all.

Gimme a Break

OK. So, now youre thinking I finally have gone over the edge. How can this possibly relate to your need to maintain an efficient and useful data processing (or information technology) division in a 21st-century insurance enterprise? The answer lies in the analysis of how we can use technology efficiently. We are operating in times where we rush to quantify and digitize everything that remotely can be considered useful data. Then once we have that data, we struggle to build automated computer systems to do something useful with that data. The business folks, the nontechies, look to us for solutions to their business problems, and it is our responsibility to provide those technical solutions when they make sense. It also is part of our job to push back occasionally and point out circumstances where applying technology may not be the best solution. My first- line generator may seem ridiculous, but in this business, we all are asked routinely to provide similarly ridiculous technology solutions to common problems or processes.

Sometimes Manual Is Better

How often does someone in your organization crank up MS Word, type in an address, select the envelope and label tool, and figure out how to feed an envelope manually into a printer all for the sake of printing an address on an envelope? Is that really a good use of technology? If an old IBM Selectric were available, the task could have been completed in a quarter of the time. It also is possible to write addresses by hand. Too often we want to use a cool tool when we dont need to. I think it makes perfect sense to use Word to create bulk envelope addresses or labels. I dont think it makes any sense to use it for a single mailing.
You think I am being trivial? My group recently was creating a series of high-value opt-in e-mails that were being delivered to different groups of customers and distributed via an in-house bulk e-mail system. There are a half-dozen steps involved in using the system from creation of the message to selecting lists to testing to release. Some of the customer groups had thousands of users. One group had seven users, and one had a single user. Did it make sense to use a complex system to send e-mail to eight addresses? I dont think so. I could be wrong. (We used the system anyway.)

Enabling Technologies

I think we all agree e-mail is an enabling technology. It often has been called the killer application for the Internet. Enabling technology has become a terribly overused buzzword. It is applied to everything from XML (as in XML is an enabling technology) to corporate intranets. Lets use a high-level business definition and say an enabling technology allows us to fulfill some business requirement more efficiently than we did before the technology was applied to the process. By that definition, we probably all will agree e-mail is an enabling technology. It provides for documented, instantaneous communication both within and without our organizations. We all agree we could not do our jobs without e-mail.

So?

So, I picked a day at random last week. During my regular working hours, I received a couple-hundred work e-mails and found it necessary to send 70 of my own (mostly in reply to the inbound messages). That is insane. And it doesnt even account for the thousands of spam and virus-infected e-mail our IT guys automatically filter out for us. Believe me, I am not that important. Seventy people do not need to hear from me today via e-mail. Nor do I need to be sent an e-mail that is actually a train ofe-mail that goes back for weeks and that I am expected to read, digest, and respond to in the next few minutes. Then there are the individuals who feel it necessary to copy at least six people on every e-mail they send to me (always making sure to include my boss). I presume the intention here is to tell everyone this particular e-mail is really important (and, by presumption, all the others arent). This isnt enabling technologythis is irritating technology. Dont get me wrongwe need e-mail. It should be a true enabling technology. The problem is we havent yet learned how to use the technology properly. I suspect in reality we waste a significant amount of work time with this and other enabling technologies. How about the guy who always responds to your e-mail in five seconds? Do you think maybe he isnt as productive as he could be?

1984

Then there is another use of technology automation that lets us see what our employees really are up to. Say you publish all procedures and policy information on a corporate intranet. Lets assume this information is updated regularly to comply with the myriad government, regulatory, and compliance rules we work with. All employees are expected to be current with new regulations. Senior management executives decree you need to start tracking who is reading these things regularly. Moreover, they want to know which sections are read the most and at what time of the day. Is that really a useful business process, or is it playing big brother? If your organization finds it necessary to spy on its employees to determine whether they are doing what they are supposed to be doing, maybe the organization has the wrong employees (or maybe the wrong managers). Cameras in cell phones seemed like a pretty good idea for a minuteuntil they started being used for immoral and illegal purposes.

Once a Geek

Dont get me wrong. I love technology. I was that kid who could build a radio or a rocket (real rockets, not those off-the-shelf safe rockets) from junk I had in my workshop. Technology is in my blood. But that doesnt mean every use of technology is good. Try to keep your organization from using technology to enable inefficiency (or foolishness). My first-line generator was a stupid idea, but it really wasnt all that wacky. Lets accept the fact not everything can (or should) be digitized and processed. I have said it before, but I will say it again: There are no IT decisionsonly business decisions. Make certain you are a part of the business decision-making process, and then you can ensure the technology solution is the right one (or, at the very least, not the wrong one).

Its Up To Us

We may be technologists, but we shouldnt be just technologists. We should be leading our business in the proper use of technology. Sometimes all we need is to suggest guidelines: If you cant say what you need in two or three lines, then maybe e-mail isnt the way to go. If the technology solution costs more in time and resources than the process it replaces, then it definitely isnt the solution.

Special Reports /

Technology Decisions

Related Stories

Recommended For You