Lesson From The Market: Software

I go to an upscale grocery store that is like Whole Foods, but without the preachy liberal nonsense. They sell all the fashionable stuff and they make sure to create an atmosphere that flatters their clientele. It has a coffee shop and bakery, so shoppers can have a latte and a muffin while they browse the organic food items. Of course, everyone uses the silly canvas sacks that are supposed to be good for the earth. It’s all pretentious and silly, but unlike Whole Foods, the prices are great so i tolerate the other stuff.

Anyway, I was at the checkout and I noticed they still use a character based computer software. Judging from the interface, I’m going to say it is written in COBOL. You can sort of tell by the blue screen with red highlights. It could be some form of business basic or maybe even Clipper/Foxpro. Regardless, it is old stuff, probably written for them two decades ago. They have scanners and scales, but that’s not revolutionary. Integration to these items was done long before the graphical user interface.

Checkout is fast and efficient. The clerks never struggle to price an item or look up some obscure code for a weird item. For the grocery chain owners, the point-of-sale system is perfectly adequate and they obviously see no reason to “modernize it.” From their perspective, this tools does the job that needs doing, so there is no reason to replace with a new tool. In the grocery business, margins are small, so anywhere they can reduce costs, they will, even if it means using a twenty year old software system.

I know of at least a dozen goodly sized companies that use legacy system for their business. In one case, they use a system written by a firm long defunct in a language long abandoned. The other use systems that are a decade or more out of date and no longer supported by the developer. I know of a large importer using character based software written in the 80’s. Most banks in this country rely on software written in the 1970’s, patched up over the last few decades. There’s lots of old code out there.

The point here is that good enough is good enough. As much as we think technology is a constant driver of innovation and change, it does run its course. The lesson from the market is that we may have picked all of the low hanging fruit from the computer revolution. While automation will continue, it will be much slower than the futurist would like us to believe. That growing pile of legacy code that is out there in the world will only make the process slower. We are quickly reaching the point of diminishing returns.