How To Kill A Business: Let Business Guys Do Technology

Business people are a lot better at cutting business deals than they are at creating the technical infrastructure to make those deals work. That's one of the points readers made after hearing about the new mobile contactless payment alliance between AT&T and Verizon. Telcos, in particular, have a long history of having good ideas but failing to deliver on them because of conflicting objectives or simply poor execution of the technology.

Even when telcos get an industry-wide standard right and make it a profitable business--with text messaging, for example--there's usually a reason. "SMS was a technical feature of the GSM standard which was baked into every operator's infrastructure," one reader wrote. "Engineers from various European operators collaborated to build the spec. Business people weren't in the room. It wasn't even originally intended for consumer use. Only later did unanticipated revenue opportunities arise, when it was already built. Mobile payments are being led by business people at the U.S. operators--very unlikely to be able to agree to the commercials, let alone standards."

That sounds harsh. It's not. In fact, it's true. And telcos aren't alone. Retailers are also a lot more successful when they don't try to create both a standard technical platform and a new business at the same time.

Consider RFID. Suppose that when retailers first began looking at RFID as a way of tagging inventory there was already a single kind of tag, with an established format for tag content and a well-established protocol for securing and reading the data. In that case, all retailers and suppliers would have used the tag in the same way. Suppliers would still moan over a Wal-Mart mandate and retailers would still grumble about the cost of the tags. But 90 percent of the early confusion over RFID would never have been there.

Or think about the number of different variations in EDI documents--not just the number of documents themselves but the fact that for years every retailer seemed to use a different set of options for each document. Retailers embraced this approach because it was easier to customize EDI documents to match the requirements of their existing business systems than it would have been to change the business systems to match EDI forms. But it was miserable for suppliers, who needed a different set of electronic paperwork for every different retailer.

And those examples are just the ones that managed to survive. The recent history of retail is littered with technology-dependent business efforts that collapsed or never got off the ground because too much technology had to be invented along with a new business model.

On the other hand, when there's an established, well-standardized technology for business people to build on, there's far less pain and a much smoother ride. Think barcodes, Wi-Fi and, of course, the Web.It's all a lot easier when the business people have fewer technical choices.

Standards are supposed to reduce many of those choices. But how do you get everyone to conform to a technical standard? It helps a lot if there's an 800-pound gorilla in the game--an IBM, a Microsoft, a Wal-Mart--that's so big other players are dragged along in its wake. Then there's a single de facto standard for everyone to conform to.

Otherwise, there are official standards created by multivendor committees that meet for years in identical hotel ballrooms. Those standards tend to have so many options and variations that there's no guarantee any two "standardized" products will work together.

But there are ways of forcing standards to actually reduce choices. The most useful tool: shame. The most useful way of going about it: a "connectathon," like the ones Sun Microsystems sponsored starting in 1986 to make sure all the products that implemented its Network File System would work together. Sun's solution was to lock all the vendors in a big hall--er, invite all the vendors into a big hall--and then have them connect their machines to a single network and test how well they worked together.

If something didn't work, it all happened out in front of competitors. It's amazing how well shame--and fear of the competition--works to encourage interoperability.

Network products have been tested with connectathons for 25 years. The electronic healthcare records crowd finally turned to a connectathon a few years ago, when it became clear there was no other way to get the fiercely competitive and incompatible vendors to interoperate.

Connectathons work. They should be used a lot more; for example, an RFID connectathon might have cut years and billions off the time and cost of implementing tags.

Of course, not all interoperability is technical. A connectathon might not be able to get telcos to work together effectively on mobile payment, simply because the people who really need to be connected with them in the virtual hall--the retailers--haven't been invited to the party. At least not yet.

But if everyone is at least on the same solid technical ground, an awful lot of finger pointing vanishes because the technology all (mostly) just works.

And business people have a lot less risk of putting themselves out of business.