Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ Changing of the Guard ]   |  Turn Your Head and Cough  |   [ Land Grab ] >>

Weekly Column

Turn Your Head and Cough: It Isn't That Microsoft Is So Healthy That They Have an Advantage in Healthcare IT, It's That the Current Players Are So Sick

Status: [CLOSED]
By Robert X. Cringely

One definition of insanity is doing the same thing over and over, failing each time, but always expecting a better outcome. Call me crazy, but I'm not insane, and that was the point of last week's column about Microsoft elbowing its way into the fight for outsourced IT services at the UK National Health Service. The same big IT consultants have been proposing the same tired solutions for decades now, and I just think it is time for someone else to take a shot at the problem. We've been told over and over that these same companies are the only ones that can take on such huge projects, yet look at their astounding lack of success doing just that. The failure rates of these big projects are horrible, as are user satisfaction and that biggest of all fantasies, increased productivity.

It's time for a new approach, I say, but some people disagree. So here's an alternate view from one very experienced and persuasive reader. After he's finished, read on to see why I think he's wrong and what we all can do about it.

"To begin," he writes, "the big boys are not exactly head count driven, as you state. While it is true that a consultant sitting on the bench is not billable, you do not win a contract just so you can bury the unlucky client under 1,000 consultants. Contracts are won and lost depending on head count, because if you bid too many, your bid will be too costly, and the majority of IT contract costs are people costs. So your profit margins are actually driven by reducing people and increasing other margins like special hardware and software, or by setting performance targets that net you a bonus for delivering ahead of schedule, etc. Needless to say, this does not make the consultant's job a happy one!"

"You are correct in your viewpoint that the big boys usually just brush off and trot out the same old IT project management processes that they have used for every other mega-government project they have worked on in the past. Familiarity breeds contempt. But, to their credit, these titanic-sized Project Management Offices (PMO) that manage and document the IT project are titanic and unchangeable for a reason: They need to manage huge budgets, reams of requirements, matrices connecting requirements to specifications to programs to hardware to documentation to deliverables, etc. Without the supporting management infrastructure, a big project will quickly lose its way and never get completed. Unfortunately this also means the projects tend to move in straight lines, driven by inertia, and they suffer when icebergs pop up and scratch the paint job before course corrections can take effect."

"Now compare this with what Microsoft has to offer. Unfortunately, their idea of Configuration Management processes, Capacity Planning, Security, and other basic IT management processes are very weak, and reflect their focus on the desktop and small enterprise. Take security as an example (please!). MS is not known for their strength here, and as a health system project manager for security, I can tell you that all the competing, conflicting and draconian security requirements on health information can make your head explode when you try to model and document it. Microsoft's documentation on these subjects is available on their web site and appears almost comically naive in comparison to what is needed for a government project the size of the NHS. Microsoft still has a way to go before they can think as big as the major government IT players. However, a nimble player who does one thing and does it well certainly deserves a chair on the sun deck with the other contractors."

"Will Microsoft succeed? Given your example of letting MS do the GUI work for the NHS project certainly allows MS to strut their stuff and take on a role that the traditional guys are often bad at -- interface design that meets the real needs of the end user. However, I would not put it past the big boys to try and cripple Microsoft's initial foray before they get a foothold in this lucrative market. I worked on one project where the lead contractor took $10 million worth of work completed by a subcontractor, terminated their contract for vague but acceptable reasons, and then successfully stiffed them for their money owed! Microsoft could find itself in a dangerous spot if the major contractors feed them manure and keep them in the dark. Wandering in the night over unknown terrain, Microsoft wouldn't even know what was about to happen until they heard the click from the pressure plate under their foot..."

Bob here. That was a fun read, wasn't it? And it sounds so convincing. Maybe Microsoft isn't right for this job (though I think they'll do fine), but does the sheer size of these contracts mean that nobody except a half dozen huge companies are qualified to even try? I don't think so.

For one thing, his algorithm for how head counts relate to bid pricing isn't the way these things are actually done. The real rule is to bid aggressively low and interpret the requirements to mean the least possible work, ignoring reality as much as possible. Once you win the deal, you then begin revisiting the requirements, increase the scope of the job, and increase the size and cost of the job sufficiently that it now becomes profitable.

What's key here is that the deliverable isn't a new Buick or an action figure or even a TV show but a SERVICE, which is very difficult to measure. That's what makes IT such a profitable business. If you are building Buicks or carrying freight there is some overhead to manage, but it is not a significant part of the overall cost of the job -- typically 15 percent or so. In IT projects is often over 50 percent.

But how many of those people are contributing REAL value to the effort?

Yes, Microsoft doesn't do configuration management and capacity planning well, or a bunch of other things. But it is helpful to step out of the IT mindset and understand what these things really are.

Configuration management is simply knowing what hardware and software is on one's computer and what options and parameters were used to set it up. There are two camps who need and use this information -- the bean counters (asset management) and the support teams. Strangely, the two never align their needs into a single tool. The two never openly share their data. What you usually get is two half-done applications that barely work. It is fairly trivial to write a script that could run once a day, or once a week, or maybe once a month to collect data on each system and upload it to a database somewhere. A simple comparison of before and after would show changes in the environment. Adding a couple fields for purchase order numbers, serial numbers, etc; and some related queries and you'll have what the bean counters need. To this date, this whole process is still mostly manual, it's rarely current or accurate, and for the admins who need the data to support their systems the data can't be trusted.

In the non-IT engineering world every item is given a part number. Every item has a published specification, purchase order, paperwork, etc. When you want to change something, you pull the documentation and amend it. In the IT engineering world what was originally purchased and why is usually a complete mystery. Config management is important in IT because they do it so poorly. So let's not be scared into using that as an excuse to keep using the same folks who screwed it up the last time.

Now look at capacity planning -- the other area where the EDSes are supposed to have an insurmountable advantage. And again the contrast to make is with plain old non-IT engineering. When you build a factory you know to the widget how many can be produced in any given amount of time. The factory is DESIGNED to produce a specific amount of product at specific production costs. In contrast, IT is amazingly primitive in how it predicts system sizing. Part of it is money -- you make more money if you sell more hardware. So you don't always get the leanest, most efficient design. In fact, you never do.

I keep coming back to the contrast of IT with other types of engineering. THIS ISN'T ROCKET SCIENCE. If Donald Trump, of all people, can build a skyscraper and make money on it, why can't IT projects come in on time and budget? The difference is that in a big construction project everyone involved understands the big picture. Everyone understands their assignment. Everyone understands their interface with the other parts of the project.

IT used to do this, decades ago, but no more. Today, IT is based mostly on selling general ideas. The requirements gathering process is superficial. The project proposal is mostly a "shell of an idea." IT projects have lots of meetings, lots of issues to be resolved, lots of management to coordinate things. The workers rarely spend more than 35 percent of their time actually accomplishing something FOR the project. In an engineering project the workers are generally over 80 percent productive.

It would seem obvious that one or another of these big outfits would figure this stuff out and use that knowledge to defeat the other players. Why don't they? Well, that brings us back to billable hours. Making their operation more efficient leads to lower revenue and lower profits. The same goes for asset management and capacity planning. There is no true incentive to do things differently. Ah, but what about those bonuses for beating deadlines? Why try for a five percent bonus you MIGHT earn by beating your brains out when there is, in contrast, an effective 30 percent overcharge bonus you are assured of getting by missing the deadline by a typical amount.

There just is no incentive to do things smarter, nor is there a culture prepared to embrace that incentive should it appear. And that's why I think whole new players are required to effect any change.

Maybe the best solution isn't Microsoft, which once entrenched would probably learn to play similar games. Certainly, they'll do their best to build Microsoft products into any final solution and will eventually become as hard or hard to dislodge than the current group of suspects. An obvious possibility is Open Source software, but to the bigshot business folks who buy this stuff, Open Source is still too scary and Microsoft isn't, even if they should be.

If you have a very complete and well thought out design, a small team of people can accomplish more than an army with just a fuzzy project idea. What Microsoft does in the first six months for the NHS will determine if this idea is crazy or not.

Comments from the Tribe

Status: [CLOSED] read all comments (0)