Strategy and Tactics – Page 201 – IS Survivor Publishing

A better TCO model (first appeared in InfoWorld)

If I were starting a business today, I wouldn’t dream of using Microsoft Office.

No, Corel hasn’t bribed me. This is simple economics. Corel has priced its office suite so much lower than Microsoft I can’t imagine enough of an incremental benefit to MS Office to cover the spread.

Very few of InfoWorld’s readers are opening a business today, but many of you have MS Office ’97 staring you in the face. There’s no concurrent use licensing, file formats aren’t backward compatible, and it costs a lot of money compared to its competition.

Besides, many of you have expressed strong interest in the NC, which means you’re open to non-Microsoft applications. If you’re going to make your move, now’s the time.

Last I heard, IS budgets were still pretty tight, so the savings should be pretty interesting to a frugal CIO. “Except,” I can hear some of you thinking (I have mentioned my telepathic abilities, haven’t I?) “software is just a small part of the total cost of PC ownership. So this will just be nibbling around the margins.”

The numbers you hear bandied about on this subject are ridiculous. If you really believe each PC costs you eleven grand a year, put out an RFP. Any number of outsourcers will be delighted to provision your desktops for at least 20% less than that (I’d make the offer myself, but I don’t like lines that long).

So here’s what I’m going to do. Over the next few months, as a bizarre hobby, I’m going to put a better model together. InfoWorld’s Features and Test Labs may join me … we’re still working out the details. The point of this model won’t be the “real” costs. The point will be to help you probe the issues you face managing this resource, so you can make realistic decisions.

Here are the key elements this kind of model needs:

Fixed and Variable Components: Any useful financial analysis separates costs into those that don’t vary with volume and those that do. Analysis often benefits from separating “semi-variable” costs – those that increase in large steps at certain volume increments) – from purely variable costs.

Partitioning of Benefit-driven Expenses and Overhead: Some expenses increase as employees take greater advantage of technology. Training is a good example: the higher your training costs, the more different applications employees know how to use, which means they’re gaining more value from their systems. On the other paw, you have to spend some money before the employee ever turns the system on. Include network connections in this category. Understanding the difference is critical.

Separation of Non-Technical Issues: I’m heartily bored with hearing how much time goes into managing hard disk space, doing backups, and all the related drivel. Take away the PC and employees will spend more time than that handling and filing paper, cleaning out filing cabinets when they get full, and wasting all the other time that gets used with non-technical alternatives. If we include these costs at all it will be as negative numbers, crediting the time saved from doing things the old ways.

Recognition of the Real World: PC cost models pretend we work rigid schedules at hourly rates. Time spent in non-value-adding activities like training is lost to the organization.

It’s a great theory that ignores our day-to-day experience. When most of us take a vacation, our coworkers absorb some of our work and the rest piles up until we get back. We deal with it when we return, and our vacation (or training class, or whatever) costs our employers effectively nothing.

Still believe each PC costs you $11K per year? Call me … I’m selling shares in a bridge crossing the East River, and I’m just sure you’ll want a piece of the action.

How not to keep the Internet from collapsing (first appeared in InfoWorld)

If you blow enough smoke, you’ll convince a lot of people there’s a fire somewhere.

Last year, Bob Metcalfe and other pundits predicted the imminent collapse of the Internet. I, on the other hand, courageously predicted its non-collapse, and even more courageously endorsed the free-market economic model on which it is based.

The collapse contingent cites big outages last year, especially at Netcom and AOL, as evidence of the coming collapse. Their recommendation: throw a big party for the Internet’s inventors, thank them, shake their hands, and send them home so professionals can take over the job, replacing the current “anarchy” with modern central administration.

Here’s where I get lost. Netcom and AOL are private corporations operating centrally planned and administered networks. Both suffered big, noisy, noticeable outages. So to save the Internet from collapse, we want to centrally plan and administer it?

What am I missing? Most Americans think communism collapsed because big, centrally planned and managed economies don’t work. Most Americans are right. The Internet’s stability comes from its decentralized model. Rules are established centrally and kept simple – hence “Simple Network Management Protocol (SNMP) and “Simple Mail Transport Protocol” (SMTP) – while implementation of the rules is decentralized.

(One other problem with the proposed solution: there are no other professionals to call in. Only the Internet’s designers and implementers have experience in designing, building and running a successful, global network that links millions of independent computers. Who else are you gonna call?)

Despite all this enjoyable gloating, I give Dr. Metcalfe a lot of credit for sticking his neck way out, and for making a self-preventing prophesy. When disaster fails to strike, very few people give credit to those who sounded the alarm. I wonder how much of last year’s investment in Internet capacity resulted from the publicity attending Bob’s predictions.

Any lessons for you in all this? Ya, you betcha, as we say in Minnesota. The Internet’s successful use of centralized rule-making and decentralized implementation, far from being anarchic, gives you a wonderful model for organizational design.

Awhile back one of my readers sent me Sahakian’s Rules (Corporate Partnering Institute, Skokie, IL, 1995). It’s worth buying. One entry: “In ancient times, in order to manage large masses of people in the battlefield, Commanders used gongs, banners and horns. If a plan of battle was not simple enough to communicate with gongs, banners, or horns, it was a bad battle plan.”

Manage your organization like the Internet, or like a Commander on an ancient battlefield: with simple, easily communicated rules, not close, central supervision.

One other useful lesson: when you manage, ignore currently fashionable theories, especially those that demonize some group that Isn’t You. The idea that current Internet management doesn’t hack it anymore comes from one such notion that I call the BIG/GAS theory (for “Business Is Great/Government and Academics are Stupid”).

Yes, you can certainly find waste in government. Some comes from its sheer size, more comes from having lawyers (most of Congress) designing work processes, and much of the rest is unavoidable. (The military, for example, is staffed to conduct wars. When there’s no war, it’s over-staffed. Big deal.)

While government isn’t as good as it could be, the Internet’s history proves that BIG/GAS is just … vapor. In the 1960s visionaries in the academic and defense community created it. It’s been an international reality for decades.

In the 1980s the Graphic Communications Association (where, coincidentally, I worked for a few years) began creating the Standard Generalized Markup Language (SGML). Commercial publishers ignored it; the defense community embraced it as part of its Computer Aided Logistics System (CALS). Early in this decade, Tim Berners-Lee, working at CERN (the international particle physics laboratory) adapted SGML to invented the HyperText Markup Language (HTML).

And business finally caught on three years ago. BIG/GAS indeed.