Your IT Department is not
an Honor Student at any School
It’s
estimated that the Fortune 500 spends, in aggregate, over $300B on information
technology annually. If you’re part of the core readership of WITDW, an
operational executive at a $1B+ company, you’re handing your CIO 3-10% of your
gross revenue. Simplistically, that money is used for two things:
-
Keeping your existing systems up and secure
-
Designing, building, and deploying new applications and improved
IT infrastructure to either save or make money
I’ve
spent a lot of time looking at how large enterprises do these things, and here’s
a report card:
A
typical company get about a “C-“ grade on the first item. While there is usually
a lot of room for improvement, the good news is that any competent manager
should be able to keep your systems alive and even enhance them a bit through
straightforward incremental improvements to processes, training, planning, and
staffing. This task is not trivial but is clearly doable. I can’t count the
number of wall charts and presentations I’ve seen with line graphs showing
ever-increasing system uptime. Someone is doing their job reasonably well.
Damned impressive.
But
let’s get to the interesting part: Development and deployment of new IT
infrastructure and new strategic applications to meet the real needs
of the enterprise. In this category, I’d give most of the enterprises I’ve
analyzed a big fat “F.”
My
database for this incendiary opinion is unscientific but broad, being based on
extended work I’ve done with Fortune 100 companies in the telecom, financial
services, transportation, energy, and technology sectors, as well as public
sector agencies at every level.
Maybe
I’m just a tough grader, but I have trouble giving a passing grade to an IT
organization when:
-
Applications are months or years late, or cancelled outright
-
Initial application releases are scaled down from what was promised with
many business functions lacking
-
Applications routinely (almost universally) fail to meet their so called
non-functional requirements in initial releases (performance, scalability,
availability, operability)
-
Applications do not integrate adequately with existing applications,
requiring ridiculous Rube Goldberg processes to transfer data that incur
real risks of data corruption. (What? Your IT guys forgot to mention that?)
-
Applications are built on hardware and software components selected ad hoc
with little or no enterprise-wide strategy
And
so forth--with the result that perhaps 80-90% of the dollars spent on these
development tasks are wasted.
Something and Someone
Now
there are many causes behind these outcomes. These causes, are, indeed, the meat
of WITDW. But there is one overarching root cause, and I’m going to start you
thinking about it in today’s screed. I’ll explain it by analogy:
-
Corporate IT is a $200 million movie production with an executive producer,
three producers, two cinematographers, four stars, and hundreds of craft
workers, but no director and an unfinished screenplay
-
Corporate IT is a skyscraper project with no architect and no blueprints,
only an “artist’s conception” of the final building
-
Corporate IT is the SR-71 team at the Skunkworks without Kelly Johnson
Is it
clear where this is going? Information Technology at the Fortune 500 level is
lacking something and someone. The something is an overall
blueprint that is designed to guide application development and to define the
infrastructure to integrate those applications. This something is called an
Enterprise Architecture. And, for all practical purposes, it’s MIA. Why?
It’s pretty subtle:
There’s no someone at most large enterprises with
the charter, skills, creativity, experience and guts to create the
Enterprise Architecture. This someone is your Chief Architect.
Now,
if you’re still in denial, you have two possible reactions to my opinion:
(1) We do so have an Enterprise Architecture! (2) We don’t need no
stinking Enterprise Architecture! Wrong either way:
The
“we already have it” error:
-
Your company almost certainly does have an IT Vision, an IT Strategy, a
Strategic IT Vision, a Visionary Strategy, and so forth. These documents
come from the senior IT management in response to the hallowed Business
Strategy.
-
You also certainly have numerous mid to low level documents dealing with IT
principles, standards, practices, and processes, to say nothing of white
papers from your Advanced Technology guys telling you what may (or may not)
be really cool in two years when the bugs are fixed.
-
You also have a dozens of project-specific documents chock full of technical
details.
Guess
what? None of these is an EA. And (common mistake here) an EA is not somehow an
emergent property of these documents in aggregate.
The
strategy level documents are akin to sketches of a proposed building. The PSPP
documents are like the mandated electrical and plumbing codes. The project
documents are like an interior designer’s floor plan for one room.
None of them tells
you what construction materials go into the building, how deep to dig the
foundations, how many elevators there are, where to put the wiring and plumbing,
where the interior doors and halls are, and on and on. The Enterprise
Architecture does. It defines the building in detail.
I
like this analogy a lot, but note closely where it breaks down: The architecture
of a building is static. The architecture of your IT is dynamic.
You would expect it to evolve for, perhaps, eight to ten years, and then be
substantially redone. Does this mean it’s not needed? That brings us to denial
number two:
The
“we don’t need an it” error:
Let’s
see—Is there any business value in common platforms, systems that are integrated
when delivered, systems that don’t lose or corrupt data, systems that use 20-80%
less hardware, systems that are developed in less time, systems that are more
reliable, systems that interact with zero latency instead of daily or monthly,
the simplicity of dealing with three or four primary vendors instead of twenty,
the capability to integrate systems rapidly after business mergers?
Gee,
this question is too hard for an old IT guy; I’ll let the business guys decide.
* * *
Maybe
you think that corporate IT, with all the maladies I’ve listed, is as good as it
gets. Maybe projects are supposed to fail outright or use five times the
hardware expected or corrupt customer data or be years late-let’s
just declare victory and go home early.
Still
here? Good. Come back next time and we’ll talk about IT heroes and villains.
Can’t
wait ‘til then? Call up your CTO and tell him you want your Chief Architect in your office Monday to explain your Enterprise Architecture. Tell
him to make sure to explain how it supports common platforms, global transactionality, end-end systems management, disaster recovery, avoidance of
niche solutions, sharing of computing capacity across projects, scalability, and
standards-based integration. Tell him to bring along briefs on the last three
major projects and use them to illustrate how the EA impacted their development.
Have
fun....
|
