Tight Couplings, Loose Couplings and the Knowable Unknown

A couple of years ago I read The Black Swan and was so taken by the subject matter that I wrote a column about it. The book had to do with the kinds of unknown issues that can strike even well-understood processes.

Now, Leo Sadovy, a vice president at SAS, has taken the concept further in a blog just posted Tuesday. Sadovy reminded me of the importance of “unknown unknowns” and risk, and in the process gave me insight into the power of social networking.

I don’t usually quote vendor blogs, but this is an exception. It’s real thought leadership without the self-congratulatory tone you might expect in a vendor post, so go read it. You might get confused with all the permutations of how things can mess up, but stick with it. I’ll wait.

Tight and Loose

A key point for this discussion is simply the definition of “tight” or “loose” coupling. As Sadovy explains, “Tight coupling means that A-causes-B-causes-C-causes-D is a given — the system is designed on purpose to automatically execute a certain chain of events once the initial conditions at step ‘A’ are met, and another chain of events for a different set of initial conditions. There are no ‘loose’ couplings where perhaps there is human intervention or where the system waits for confirmation from an independent source.”

From that definition, you can infer that loose coupling is pretty much the opposite of tight coupling. Now, for a long time, business processes were more or less tightly coupled. That’s especially true in the back office and manufacturing, and to a degree it’s true in the front office too. We like tight coupling; there is comfort in knowing that if “A” happens, everything else will fall into place. If “A” is an order and “D” is payment, “E” is business bliss. Tight coupling is what back-office computing is all about (almost).

In the back office, to grossly over simplify, you get an order and you go to the supply chain to get parts or components, or eggs, flour and sugar for all I know. Things show up on schedule, your manufacturing process goes into hyper-drive, and it all ends when something gets put on a truck. Actually, it doesn’t end until the invoice is paid and the suppliers receive their money, but that’s a nit here. That’s the back end of the business.

The front end is much different. A customer calls up and asks about a product or asks to speak with a salesperson. A meeting gets scheduled, you have the meeting but nothing happens. Maybe nothing will ever happen, or perhaps something will happen later when you aren’t even expecting it.

The back office is tightly coupled, the front office not so much, but here’s the fun part: Risk in the back office has, to the best of everyone’s abilities, been reduced to unknown unknowns. The raw materials may not get here on time, but we’ve done our best to eliminate the obvious sources of delay. For a shipment to not arrive, it will take an act of a supreme being that we cannot factor in ahead of time — in short, an unknown unknown.

The loosely coupled front office has no chain of connections that takes us from interested party to satisfied customer. It has many connections, each of which works sometimes. Many consultants and sales managers would like it better if the front office was more like the back office, stamping out customers with the efficiency of a manufacturing production line.

Weeding Out Known Unknowns

For many years we’ve tried to make the front office more like the back office, and we’ve always been disappointed. That’s why they’re still selling products that improve sales efficiency and effectiveness. But really, I have never seen anything that gets its arms around loosely coupled selling processes like socializing information through analytics. Together, they enable us to gather data from customers and use the resulting information to engage them in ways that don’t promise we’ll get to “D” or even “B,” but they are better than anything else I’ve seen.

As a practical matter, if we use our knowledge of loose coupling and apply prudent management to weed out as many of the known unknowns as possible, we can get very far. I’ve been a fan of modern forecasting solutions that do the weeding out for that reason. One that I am familiar with is Cloud9 Analytics, with products for pipeline management and sales forecasting.

Very simply by capturing data from a CRM system and applying analytics, these products can identify what’s changed — for better and worse — or stayed the same instead of progressing. It’s not rocket science, just efficient elimination of knowable unknowns — how often do we neglect to do that in our lives, and what are the results? You might think this has little to do with social media, and you’d be right, but at another level, applying analytics this way enables socialization of information, say, among the sales group and its management. That’s all it takes — you don’t always need social media to apply social ideas.