% fortune -ae paul murphy

Reversing Outsourcing

A few years ago I blew an interview at Gartner Group by refusing to give them the answer they wanted to hear about out-sourcing. The question was along the lines of: "what analyses have you done to help clients understand the financial benefits of outsourcing?" I said first that there usually weren't any, and second that very few out-sourcing decisions were ever made for financial reasons. They said goodbye.

They weren't interested in hearing it then, because that's what they were hiring to sell: name brand financial justifications for out sourcing decisions. I wonder, however, if they wouldn't have a lot more empathy for my position today because the next wave of out-sourcing consulting is going to focus more and more on terminating existing contracts and bringing IT back in-house.

The big driver on insourcing, of course, is exactly the same thing that drove outsourcing: dissatisfaction with IT services. The real outsourcing sales pitch is that it lets senior executives put a positive spin on firing the IT department without forcing them to face their own responsibilities for IT. Insourcing decisions have exactly the same basis: it's just that the target isn't the internal IT group, it's that same group working for the outsourcer.

Thus yesterday's cnet headline: Sears ends $1.6 billion deal with Computer Sciences isn't the first and won't be the last. Fundamentally most outsourcing contracts substitute tweedle dee for tweedle dum and it should be no surprise, therefore, when interchangeable people and technologies produce interchangeable disasters.

So what can you do? Simple: don't be tweedle dum. Focus on reducing the upward flow of user complaints about the quality of IT support -but remember that expectations set by the perfect systems on television dramas like NCIS, by IBM's ads, or even just the latest oxymoronic Sunday Supplement on secure wireless, cannot be met.

One totally cynical option I've been growing fonder of lately plays directly to the biases induced in executives by the contradictions between what they're told by the mass media and what they're told by user management. What I do is argue that adopting open source applications on Unix (usually Linux on x86 or Solaris on SPARC) will produce monetary savings at the cost of some reduction in service quality.

I'll argue, for example, that it's cheaper to convert a building full of Microsoft Office users on Windows/XP to OpenOffice on Linux or Sun Rays than to upgrade them to the latest PCs and licenses -and then humbly shuffle my feet while letting them force me to admit that doing this will require the users to give up some of the productivity benefits they'd get from the power and flexibility of the real thing.

The higher the executive level, the better this pitch seems to sell. The only catch, really, is that both halves of it are lies. On the cost side I always try to spend every nickel available on getting the biggest, clearest, monitors, and the fastest servers, I can while getting rid of every switch on their network. As a result, the cash benefit usually disappears, but so do user complaints because the quality of service goes up quite dramatically. With 19" Sun Rays and SPARC servers, for example, reliability approaches 100%, the virus threat disappears along with the weekly upgrade traumas afflicting PC users, while things that take a long time on PCs, like loading a big email file, become virtually instantaneous.

Ironically, this process is just what that Gartner guy wanted me to swear allegiance to: just done backwards. They think they're paying for some cash savings by giving up functionality but actually get a breakeven on short term cost, a tremendous improvement in service quality, and the opportunity -just by keeping their hands off- to get productivity gains and operating cost savings over the longer term.

Paul Murphy wrote and published The Unix Guide to Defenestration. Murphy is a 25-year veteran of the I.T. consulting industry, specialising in Unix and Unix-related management issues.