Oh, what the heck. Best Buy isn’t going to bring me in as a consultant anyway — not after last week’s column about the retention bribes it paid to execs whose names are all over its steady decline.

So now, a few words on why its headlined 2004 IT outsource to Accenture (followed by its covert 2011 re-insource from Accenture) should have failed. Not, I have to add, why it failed. Best Buy hasn’t even formally admitted that the outsource was a bad idea, let alone explained in non-ManagementSpeak terms what led it to reconstruct its internal IT organization.

But a key reason it should have failed, and in fact why Best Buy should never have considered it in the first place, was right there in the press release:

Accenture advocates taking “packaged vanilla solutions and weaving them together in as simple a fashion as possible” and changing business processes, rather than heavily customizing software, as the most cost-effective approach for retailers, [Angela] Selden [Best Buy’s spokesperson] said.

Selden said she recognizes that the approach would represent a dramatic change for the legions of retailers that claim they had to heavily customize systems because of the unique needs of their businesses. But she said they must change the way they do business in order to be nimble enough to “absorb innovation quickly.”

What’s wrong with this picture?

Had Accenture’s pitch been that it had the expertise to help Best Buy leverage its bricks-and-mortar retail strength so as to beat Amazon.com at its own game, there might have been some sense in this arrangement. But that wasn’t the supposed logic. Understanding how flawed that was could help your company avoid a number of common mistakes. Here goes:

In 2004, Best Buy was at the top of the consumer-electronics heap, growing while its largest retail competitor, Circuit City, was in a state of steady decline.

Let’s pretend, just for a moment, that its success wasn’t entirely accidental. If that was the case … if Best Buy’s success was due to actual business competence … then its business processes were a source of competitive advantage, even if they did need heavily customized software to support them.

Which means Accenture’s sales pitch, rephrased for honesty, went something like this: “In exchange for hefty fees, we’ll rip out your sources of competitive advantage and replace them with generic alternatives.”

Accenture, that is, claimed it had more expertise in Best Buy’s business than Best Buy had … a sales pitch that might have made sense had Accenture pitched it to Circuit City. But to Best Buy? Seriously?

Even if Accenture did know more about retailing than Best Buy, its assertion that changing business processes is more cost-effective than customizing software strongly suggests that Accenture knew nothing about what it takes to change a business process. Changing a business process takes more than drawing the new one on a swim-lane diagram and teaching it to the employees who will have to live with the new way of doing business.

Because when the trainers have all finished, those employees will have to become practiced at the new process. That takes time, during which the whole company is less effective than it used to be.

But wait! It’s even worse than that! You see, the effort needed to customize software is a one-time investment. The inefficiencies that are the inevitable consequence of tailoring business processes to deal with the software’s limitations are costs the business has to absorb every day.

Which is why the whole plain-vanilla vs customized-software argument is the wrong conversation to have.

Best Buy fell for it. It’s no longer competitive. Draw your own conclusion about cause and effect, but please, don’t allow anyone in your company to debate the merits of plain-vanilla vs customization, because no matter which side wins the debate, the company will only follow the optimal course of action by accident.

Understand, I’m a theory-of-constraints guy. ToC says that for every business function, right after ranking the six dimensions of optimization (fixed costs, incremental costs, cycle time, throughput, quality, and excellence) in order of importance, the next step is identifying the most serious barrier to improving the top-ranked dimension, doing what’s necessary to remove it or reduce it until it’s no longer the most serious barrier. If that means customizing the supporting software, so be it.

Repeat ad infinitum.

It’s straightforward. It works.

Even better, you won’t need to renegotiate terms after the thrill of an outsourcing deal is gone.