Return to Enterprise Table of Contents


A roll-your-own network is a perilous project, but it saves big bucks-and even rouses the ranks

It isn't every day I blast The Cranberries' song Zombie from my computer's CD player. But today we inaugurate our local-area network, and a carnival atmosphere fills the office. Co-workers gather around each other's desks, marveling at the miracles wrought by word processing. For decades, the clerical workers labored over typewriters. The bookkeepers can't believe how much easier Excel looks compared with the old spreadsheet program. And these computers are so much faster and better-looking, I keep hearing, than the five-year-old models they're replacing.

This enthusiasm quickly spills over into action. Staffers type up their own letters, freeing customer service people to handle more calls. Our plant engineer hatches a plan to use E-mail to buy replacement parts from suppliers that post catalogs on the Web. That will reduce order errors, since we'll get to see more of the parts before we purchase them.

I'm grateful for the good beginning. At our family business, Emerald Packaging Inc., a 33-year-old plastic-bag maker, we had mulled over a local-area network for three years. The idea was to save clerical time, feed sales, costing, and production information into all parts of the company in real time, and create new efficiencies--all the things computer networks are supposed to do. My sister, Maura, who initially spearheaded the project, backed off in 1995, frustrated that she couldn't find the right software package. When I joined the company a year ago, she dumped her project files on my desk.

EASY PACE. I took the challenge. But since a little bit of knowledge can be dangerous, my first step was to find a good consultant. I tracked down a friend who worked for a large software firm, but the rapid timetable he proposed was certain to overwhelm non-techies. I called big consulting firm contacts--too expensive. I settled on Tom Li, a local consultant who had helped us keep our ancient PCs working for the past three years.

Good choice. Tom understands our company, knows a lot about technology, and hasn't pushed the project faster than we've been willing to go. He spent months, for instance, helping me decide what kind of server we needed. One consultant had suggested we'd have to spend at least $15,000 for a high-end Hewlett-Packard server. Tom helped me figure out that a midrange Compaq Computer model would do the trick for only $6,600. And we can afford him: He charges $150 per hour vs. $250-plus at, say, Anderson Consulting.

But cracking the software conundrum hasn't been any easier for me than it was for Maura. Unfortunately, there is only one package written for companies in our business. We could have it for $110,000, no less. That's a staggering amount when you consider the typical software package for manufacturers our size costs around $30,000.

Even so, Tom and I gave the software serious consideration. It provided everything we wanted--pricing, job costing, inventory control, shop floor data collection, and the full suite of accounting modules. Unlike off-the-shelf products, it did so using industry-specific language. Worse yet, Tom and I concluded, we would have to customize any other package so severely that future upgrades would erase our modifications.

But $110,000 just seemed too steep. So about three months ago, Tom and I determined to go it alone. We decided to buy the accounting modules and inventory system off the shelf but create our own job costing, sales order, pricing, and shop-floor programs. We have budgeted $35,000 toward the project and hope to have it done by mid-1997.

So far, we don't have much to show. Fear of acting before we know exactly what we want to measure has slowed us down. But I think we have broadened our chances of success by seeking input from as many sources as possible. We've corralled line workers to aid in setting up the shop-floor reporting system. That's helping us figure out what we should measure and how the screens should look so they're user-friendly. So far, we've mapped out the graphics on storyboards, which are on a table in the conference room, where anyone can go and attach yellow Post-It notes about cutting or adding fields.

Buying the hardware before we have the software may seem like putting the cart before the horse. But many of our officers don't have much computer savvy, so this way they'll have time to get comfortable using the spreadsheet, word processing, and E-mail programs. And since we're looking to hook 14 computers together, building a local-area network seems like a no-brainer. The enthusiasm and ideas generated by the mere presence of the new machines has convinced me that while computerizing may not be easy, it will probably prove to be one of our safest investments yet.



Updated June 15, 1997 by bwwebmaster
Copyright 1997, Bloomberg L.P.
Terms of Use