A popular outsourcing rationalization has it that companies should “keep the core and outsource the rest.”

I call it rationalization rather than rationale because:

• There isn’t a shred of evidence to support it.

• It solves nothing: Outsource something you don’t know how to manage and you still don’t know how to manage it, only now, you’re badly managing a company that wants as much of your money as it can get.

• It suffers from recursion failure.

KJR has already covered the first two points—click on the links or buy yourself a copy of Outsourcing Debunked (Bob Lewis, 2011). But what’s recursion failure? Glad you asked.

Outsourcing is like anything else in business—to succeed, you have to be good at it.

But as anything that isn’t core must be outsourced, and as the notion that managing outsourcers might be a core competency is absurd, the only logical conclusion is that companies that outsource must outsource outsourcing management to an outsourcing management outsourcer.

To succeed at that, the company must be good at outsourcing outsourcing management. And so on, ad infinitum—recursion at its finest.

Okay, I wouldn’t want to try that logic on a business executive weighing the pros and cons of outsourcing, but it was fun, wasn’t it?

What isn’t fun: Why an increasing number of American business managers are receptive to the even-more absurd arguments in favor of IT outsourcing, both the traditional kind and its current commodity end-point, cloud computing.

Look, even those unenlightened IT shops that thought they had internal customers usually involved themselves in the business processes and practices they were helping improve through automation to some extent. Few business analysts strictly limited their conversations to “what do you want the software to do?”

IT outsourcers, in contrast, deliver software that fulfills requirements and meets specifications. If they do that, all is good with the world, whether or not the software does anything useful.

Deep down inside, every business executive who ever endorsed an IT outsource understood this difference, and yet it didn’t matter. They considered overseeing IT to be an aggravation, and so they willingly “kept the core and outsourced the rest.”

Now we have the cloud, and software as a service (SaaS). The “new” question is, “Why should we spend lots of time with IT on a CRM implementation when we can call Salesforce and be up and running the next day?”

What’s sad is that they know the answer to this seemingly rhetorical question: If they do this they’ll be up and running with what we used to call, in more enlightened times, an island of automation.

Multiple islands, really — as many as they have sales representatives configuring Salesforce as they prefer. Add to that a database that’s completely unusable for reporting and analytics, as each sales rep stashes the data they want in whatever data fields appear convenient for that purpose.

Heck, IT could do that in a day, too, if it was amateurish enough to be satisfied with an implementation that banal. It could buy Act! licenses for a fraction of what SalesForce would cost, too, installing the software on individual sales rep laptops with no attempt to integrate them.

Nothing to it.

We in IT have failed in at least three respects, and we’d better fix all three soon, or we won’t be around to say “I told you so.”

The first is that we thought business executives long-ago absorbed the islands-of-automation argument, so we stopped making it. They had absorbed it, but ideas have a half-life, and because we stopped repeating it, this idea long ago lost its potency.

The second is that we argue rather than discuss. Faced with a sales executive who is thinking about Salesforce, too many CIOs say, “You can’t do that. Here’s why …” instead of, “We can do that … in at least three different ways, depending on what you want to accomplish and how much you’re willing to invest to get it.”

Then there’s the third — failing to focus everyone in IT, from the CIO on down to every help desk analyst—on the importance of managing relationships throughout the company. Without this, nobody will give the CIO or anyone else the time to have these discussions, or the patience to listen to the to-them complex engineering issues we need them to engage in.

So of course they outsource, and go to the cloud without involving us.

We’ve given them no reason not to.