Soft In The Head

I thought it would be better, and cheaper, to create our company's own software from scratch. I was so wrong

Just after I joined our family printing business four years ago, I made a decision that at the time looked like a no-brainer: to computerize the company by writing our own software tools. I wanted something that would fit our needs more snugly than an off-the-shelf package. Similarly, I wanted to implement the system at our own pace over a couple of years, rather than fitting into someone else's hurried integration schedule. I also thought it would be a lot cheaper.

Let's just say this wasn't my brightest moment. With nearly $50,000 invested, we're not even close to finished. Sure, we have the basic financial programs up and running. But the vast majority of our computer dreams remain, of all things, on paper. We can't get reports--on daily production, for instance--because we haven't had the time to create the necessary forms. Worse, just a few of our programs are integrated, which has caused huge paperwork bottlenecks now that business is booming. So last month, I decided to get out of the software business and bought a package designed for companies like ours that print and manufacture plastic bags.

How did I end up being so wrong? Largely, I underestimated the time and energy it takes to create a system from scratch. For example, to devise a production reporting system I had to get managers and hourly employees to list all the information we would want to capture. Then we had to create an interface that looked and felt accessible to the hourly workers who would use it. We couldn't even get so far as agreement on how to design the screen. Worse yet, even if we made a decision tomorrow, it would take our programmer months to write the code to create production reports.

We simply couldn't wait. Sales are up 25% this year. Our people, especially in the office, are swamped, and the current setup doesn't ease anyone's workload. Again, take production reporting. Since people can't enter their numbers on the shop floor, our receptionist spends part of the day logging them in. And since her system isn't integrated with our job-costing program, a customer-service person wastes several hours a week pumping the numbers into that program. Her time would be better spent making sure our new customers are happy.

As the workload has zoomed, my office mates have made known (loudly) their opinion of my software efforts. Worst of all, I tallied up our spending and estimated we were maybe halfway home. That would mean that we might spend close to $100,000 creating our own programs--roughly the price three years ago of the system we're buying today. When I found out the current price is close to what we would spend to finish off our own efforts, I went with the sure thing.

At least I think it's a sure thing. The company that makes what we're buying, Concord Business Systems Inc., spent two days with us recently and walked our office staff and several of our shop-floor personnel through the system. Everyone came away impressed. It speaks the language of our industry and is based on Microsoft Windows, which we use. And since the people who created it are in the software business, they've designed some versatility we never imagined, like letting shop-floor employees enter production numbers on a touch-screen computer. This breaks down a lot of the resistance to my system, which would have required typing skills few people in our factory possess.

Still, getting the whole thing to run won't be easy. We're going to have to enter lots of data and adapt some of Concord's forms, like job-specification sheets, that differ from ours. Plus, we'll be held to a serious schedule for implementation--with penalties for missed deadlines. Fortunately, though, I think our people are ready. I know they're ready for something besides my first--and last--foray into the software business.

Before it's here, it's on the Bloomberg Terminal.