In my last post I wrote about the offshoring phenomenon among the larger technology companies, like IBM, and I touched very briefly on legacy costs that have many components wrapped into them, aside from the usual assumptions of retiree commitments. In this post, part 2, I will cover the diseconomy of scale aspect of them.

In Technology Strategy 101, the high level goals are to reduce labor requirements, and to reduce and avoid unnecessary complexity. It encompasses a few different things, like making sure the elements within the infrastructure coexist well with minimal overlap and manual integration points, and is made up as much as possible of solutions that can be shaken out of the box and customized without interfering with the broader infrastructure.

There are a few ways to go about this, but as I found out, it’s a lot harder than it sounds in a large corporation with lots of politics and elements of diseconomy of scale such as fiefdom building and job security built into solutions (a million lines of spaghetti code) that results in inefficiencies such as technology for technology’s sake, solutions that have to be forced to coexist and can’t be cost effectively upgraded, or render the rest of the infrastructure unable to be upgraded to take advantage of automation features, and needed information cannot be provided on time and in the format required for the business to be successful. It is quite costly, mostly because it requires far more labor and overhead than necessary, and runs the risk of reaching a critical mass when it cannot change fast enough to support the direction of the business.

And what I was leading into in the first part is that a large part of the offshoring phenomenon we see in the infrastructure technology space is what I call “We’ll run your mess for less.” Because way back when in the little fiefdom, someone made a choice of a technology solution that required x10 the care and feeding of other possibilities because it matched their minion Bill’s skill set. They want to keep Bill. Bill hadn’t gotten the chance at his management Band as he’s been groomed for, and it gave them a chance to finally put Bill in charge of something; and imagine this happening with 10 different Bills, at least.

But then, after a few years, something strange happened. The bean-counters noticed that IT expenses, which are generally considered SAG, are astronomical and unaffordable; and the business has been complaining that they are missing opportunities because IT can’t deliver. So, the inevitable happens anyway. Bill and all of his direct reports get their pink slips as the entire thing is outsourced to Big Blue. Big Blue says, “We’ll run your mess for less,” and instead of throwing American labor at the problem, they throw offshore labor at the problem on virtual machines as the bad choices merely change hands. But the problem is still there, and in some sense, it is now worse because the contract is a contract. And depending on the size and weight of the customer, they are likely to have a disinterested “partner” (and I mean disinterested it the broadest sense of the word).

I remember interviewing at Kodak for a strategy position about six years ago, and they were worse off as far as silos in IT than where I was coming from. I chose not to take them up on their offer because while they recognized there was an efficiency and manageability problem, they were unwilling to take necessary steps to gain control over and centralize key infrastructure components. There was little chance at success being unwilling to stick a stake in the ground and consolidate. And the company just emerged from bankruptcy last year…

The basic point here is that many of the companies that outsourced to offshore outfits thought they were escaping the cost of diseconomy of scale, when they just made it cheaper to maintain. It is like optimizing waste; and I think there is an enormous opportunity in the IT services space to standardize and centralize 80% of infrastructure requirements and run multiple customers with the same infrastructure.

It can be done domestically and still be competitive with the offshoring big boys, because it would take only a fraction of the labor while providing customers with automation, AI, all with a consistent response that makes support incidences rare. How about doing a much better job and providing customers with what they really need with 1,000 people than the big boys can do with 5,000?

As far as I know nobody is doing this. Microsoft Azure comes close, but they provide disparate virtual infrastructure for customers to do whatever they want with; and is not the kind of IT partnership and leverage of economies of scale like having the same infrastructure and SQL as a service, etc…, that I have I mind. Where are they?

PS: I am not much worried about letting the cat out of the bag. I was looked at for strategy, sort of a casual glance that I wasn’t supposed to know about (but the walls have ears); and I haven’t heard anything to date. Besides, it takes seasoning in having dealt with the issue of diseconomy of scale in IT as a career to be able to prepare the justification, architect and engineer, and establish appropriate governance processes for such an undertaking properly.