IT cost, ignorance, and the race to the bottom

Look at the new crop of "smartphones" in the context of what happened with the PC - and you'll predict a race to the bottom as claims for the things escalate and both price and actual functionality fall.
Written by Paul Murphy, Contributor on

There used to be a high end Scandinavian furniture store in Edmonton rejoicing in the slogan: "For those who can tell the difference." Unfortunately most people either can't or won't - and one consequence is that the current store of the same name sells mid range products at high end prices.

We see that same phenomenon all the time in IT: almost everyone outside IBM's data processing markets wants to focus on cost, but price shopping across architectures implies a value equivelance that does not exist.

The most obvious reason for the cost focus in IT is that it's easy to attach credibility to costs measured as the sum of checks written, and very difficult to attach credibility to benefits estimated on the basis of airy fairy stuff like user seconds saved per transaction. The more subtle reason, incidently, is that senior management's responses to IT are conditioned by data processing expectations in which hardware costs millions and leasing makes it a pretty good proxy for operating costs.

Underlying the errors this produces is, I believe, a mistake most of us make: assuming that the factors driving personal decision making apply equally to decisions in larger organizations. My favorite political example of this involves brick recycling: because if you have a use for some bricks and your neighbor is throwing some out, then it's just stupid not to use those -but when cities extend this logic to forcing developers to recycle bricks from building demolition, the cost burdens imposed generally produce the opposite effect: largely shutting down inner city redevelopment in favor of greenfield work in the suburbs.

In the IT version we see lots of real people believing that the unit cost of a PC as advertised in a drug store flyer says something about the cost of a PC used in business - and one of the saddest stories I know illustrating the stupidity this produces features a senior government guy blowing a few million taxpayer dollars. The situation was that his agency wanted to issue a time and materials development contract covering about 120 people for an estimated eight months - but government wide PC purchasing preferences meant that the approved purchasing process for the workstations needed would bring in product from a vendor he didn't like - and possibly raise questions about the application development contract too. Faced with this, the preferred vendor offered to provide 120 older workstations for the project duration without mentioning this in the paperwork; and the ADM, who noted during the meeting that both his teenagers used two year old laptops surplused by his department, happily agreed to what he thought was a bargain - with the result that about 120 hourly contract developers got to spend nearly two years using 32MB/600Mhz PII gear for code compiles and testing in an era during which the PC transitioned from the 128MB/Ghz P3 to the P4.

People will tell you that if you can't tell the difference between two products or choices, then there is no difference, but that's just wrong - it's a celebration of ignorance, not a justification for failing to do your homework. However, there's an unfortunate consequence to this: because once people stop seeing differences in the product, they seek differences in pricing and so drive a race to the bottom - a race that drives suppliers to distinguish themselves more in their advertising than their products. Today's major PC companies, for example, all deliver the same products, made by the same people, from the same parts, and loaded with the same software -while selling the public on the belief that their product is the same as the other guy's but magically also better and cheaper.

In reality, of course, this is ridiculous - but if you've ever wondered why today's 3Ghz, four core, Windows 7 PC doesn't do much more than an early 90s Mac AV, this is the answer: price competition coupled with marketing differentiation on indistinguishable products makes it cheaper and more effective to add cores and wait states than to increase memory throughput or change the software in any fundamental way.

Where all this comes together is in business adoption of the iPhone and iPad clones now being rushed to market by the same companies that decried the Apple products as useless consumer ornamentation - because what we're seeing there is exactly the same ignorance driven rush to the bottom, albeit this time with Unix variants (Android, Linux, Qnx, etc) in the MS-DOS role, that we saw with the PC.

This is a frightening prospect - and it's not helped by remembering that Microsoft's first OS product was actually a Unix clone. The saving graces, however, could be that this time around IBM's closed community isn't driving purchasing and Apple's consumer market share is well enough established to ameliorate IT's drive to the bottom by ensuring that at least some people continue to be able to tell the difference.

Editorial standards