Ten years ago, there was a strong longitudinal study performed of IT spending. This wasn’t the typical “how much is this industry spending” type: it looked under the covers of IT spending, large and small, across all industries to understand whether there were some common threads.
There were. At that time, the typical enterprise — regardless of its choices for packages, its choices of technologies, whether it was outsourced or not — was spending 80 per cent and more of its money just keeping the lights on. (Remember, 10 years ago, we weren’t in the cloud, hadn’t virtualized and the majority of our portfolios were depreciating client-server applications and residual mainframe or midrange services.)
When four out of every five dollars you command goes to standing still, you can’t do much (and the project request backlogs of the time proved it to be true). Getting more of the spending delivering the future, and less of it delivering the past, was an essential next step.
Enterprises cleaned house. They retired applications, removed duplication, got rid of whole platforms. They ruthlessly consolidated servers and virtualized heavily. Contracts were renegotiated.
Today’s stars manage to achieve 65 per cent (plus or minus five) on keeping what exists operating well and in compliance. Some, to get that low, even undertook the multi-year journey that is removing all the mods in major packages to make maintenance and updating the environment cheaper.
Still, getting to a basic half-and-half split — half of the spending going to run and maintain what exists, half going on the next new things — is still a long way away for IT.
The trouble is that in-the-business spending often looks like it reaches that 50:50 split, or even better. (“Looks like” is the operative phrase here, since the business always has the option of dumping operating expenses off onto IT by “graciously” allowing IT to take the reins of some system or other they built for themselves — and most business departments don’t categorize their expenses cleanly enough to do side-by-side comparisons with ease, anyway.)
Why is a 50:50 split so important? It’s the key to being an agile enterprise, flexing and shifting to seize opportunities.
Enterprises can grab onto all the insights of a Euan Semple (Organizations Don’t Tweet, People Do) on weaving themselves into social cues, organize around the Wirearchy principles of a Jon Husband, have the team capabilities enabled by Steven Forth’s new start up Nugg, ruthlessly manage cost using Mike Rogers’ Counterpart offering, and do a host of other recommended things to be an agile enterprise, and still find they can’t get there from here thanks to the boat anchor holding them back of their existing systems (and their share of cost and effort).
If it takes you three calendar years and a few million (to retrofit mods) every time you want to get your ERP implementation to shift gears, you literally can’t be an agile enterprise.
There’s no magic in striving for a 50:50 split. Having half of all your IT spending available for new things doesn’t guarantee anything. But getting from best practice and performance in managing the cost of standing still — 65 per cent or so — to 50 will require you to shift your thinking about your portfolio of applications and services far beyond doing the usual steps we’ve been doing.
You may, for instance, find that instead of consolidating around just a few behemoth packages you’ll need to return to tightly focused point solutions. You’ll invest in standardizing your connections between them, so that change ripples can be isolated and controlled. You’ll be resisting any and all mods, to keep maintenance and upgrade costs low.
Almost assuredly you’ll be in some sort of hybrid cloud solution, with work slipping and sliding from one point to another. You’ll probably also be configuring for continuous operation and “dial up and down” capacity based on loads rather than configuring for “normal” conditions.
You may never reach getting half of your spending going toward new things in your lifetime. That’s not the point. The point is that your enterprise will need to be much more agile in the future simply to stay alive. (Yes, that’s true even for government: the restructuring of ministries, division of work into outboard agencies, recombining piece parts of recent years will accelerate. So, too, will be the burden of providing services with less. It’s as disruptive to doing your job as global competition is in the private sector.)
Aggressively tackling your legacy (a good working definition of “legacy” is “in production” these days) and shrinking the drag it brings to enterprise agility is one of those jobs that no one’s clamouring for, and no one will say thanks for. But getting the kudos for the big data implementation that made a billion, or the ten points of added market share gained from decentralized field devices, or the millions saved thanks to an Internet of Things sensing system, or any of the other “wins” you might get, will be built on the back of lightening the burden of the past and the cost of daily operations.
Can you get to half your spending being on new things? Maybe a better way to put it is “how long can you go on not trying”?