Back after a long hiatus — Not for lack of content but lack of time to focus on letting it out – let’s see if I can get back on that horse and ride for a while…
I’ve recently worked with a client who demonstrated an organizational “stupid human trick” that was new to me, so thought it might be worth (discretely) sharing it.
They are a large diversified financial institution, with 5 or 6 vertically siloed business units, each with their own IT organization. The central IT organization is relatively small and historically weak. Most of IT development and support is outsourced, to several major vendors.
Several years ago, they created a center of excellence to manage all of the data movement interfaces between business units and data centers, using Datastage, JCAPS, Tumbleweed, etc. As an aggressively multi-sourced IT shop, they unleashed their Procurement folks on the vendors and negotiated rock bottom rates, on the order of a few thousand dollars to develop an interface.
The result? They now have nearly 2000 interfaces in the “COE”, driving a substantial sustain budget and exploding to as much as 3000 in the coming year as several large projects kick in. The interfaces are so cheap to acquire that the IT teams and their customers don’t see much point in considering opportunities to reuse or rationalize the existing data feeds.
The historically weak IT center has not established or enforced any meaningful control over the architecture (until now), so the COE just does the best it can to cope with what are now highly redundant, unstable interfaces over multiple software platforms and versions.
Which led me to this new “insight”: Highly Effective Procurement Can Lead to Crappy Architecture.
Perhaps this is a fancy way of saying “you get what you pay for”; it was certainly also a powerful illustration of the way that internal market forces drive the behavior of decision makers inside an organization — make something cheap relative to the alternatives and people will buy a lot more of it!
And as any Economics Professor might add, if the true cost of something is not reflected in the price (like air pollution in gas prices) then people will over consume it. The cost of cleaning up this “environmental hazard” (2000 interfaces) will be borne by the client for some years to come…
Ring any bells?
