Devil's Advocate: More hazards of buying technology

Last week we heard about Martin Brampton's own problems with upgrading his PDA and mobile (http://www.silicon.com/a52996 ). But how about corporate buyers of IT? Is there any way around the same problems with multimillion pound price tags?

My ability to buy technology was trashed last week, as Ericsson disrupted my carefully laid PDA plans. Was that an isolated event? Or is there a more general problem in IT that makes integration prone to failure?

Best of breed is one thought with which one can approach the purchase of complex IT. The idea is to look at each requirement, deciding which is the most suitable product to provide the solution. Then just link the individual best of breed solutions together.

Claiming that applications should drive technology is another way of saying more or less the same thing. The argument is that no IT policy should be the deciding factor in selecting an application. The choice should be made purely in relation to the business need and how rival software products meet it.

Obviously, there is merit to this approach. Politically astute IT managers may be reluctant to follow any other path when faced with powerful interests within their user base.

Yet, as we have seen, however much care is lavished on the integration of the elements of a solution, sometimes they simply do not fit together. A possible solution is to build additional software to provide links between disparate systems. Unfortunately, applying my programming experience to the creation of a new interface between my Ericsson PDA and Windows 2000 is a bigger project than I can accommodate. And buying a new PDA would certainly be cheaper.

The problem of uncontrolled cost in technology integration has been met to some extent in the large scale. Middleware is now available to provide a layer that smoothes out the wrinkles between different systems. It is expensive but not as expensive as trying to build your own.

Which neatly brings us back to one of the major reasons for wanting to buy best of breed. Although it is commonplace that the initial purchase price of an item of technology is by no means its total cost, decisions are still all too often driven by the initial price. Some vendors encourage this. Inevitably, the full cost becomes all too visible during the life of the 'cheap' facility.

Quite often, the opportunities to be found in best of breed systems are to do with long-term costs. Take a simple thing like email. It is easy to get into email and users can be added to the system until everybody is connected to the world at large. But then what often happens is the cost of administering the system becomes inordinately large. 'Cheap' email turns out to be a significant drain on budgets and skills, limiting the introduction of other valuable capabilities.

Opposing best of breed is the policy that sets technology standards from the start, limiting choice of applications in the hope and expectation of easing integration. The extreme form of this is to prefer a single vendor for all technology.

Now only once in the history of IT has this really happened across both hardware and software, and that was in the period of IBM's dominance. In the early days, as IBM was working towards claiming 70 per cent of the entire world computer market, much software was 'free' with the rather expensive mainframe computers. In the end, of course, it was the customers that turned their backs on the 'nobody ever got fired for buying IBM' approach. IBM posted a landmark corporate loss.

Ironically, the rifts that appear even within large technology companies frequently defeat the single vendor approach. Successful IT vendors increase their reach by buying small, innovative companies. Or their divisions strike out on incompatible paths. And quite often, the vendor is no more able to integrate its range of products than are the hapless buyers.

So nowadays, the situation is complex. The unfortunate IT buyer is obliged to negotiate a mish mash of vendor claims, supposed industry standards and genuine attempts at integrative standards. Neither faith in one vendor nor best of breed is guaranteed to deliver results. But perhaps, as I suggested a few weeks ago, buyers have only themselves to blame as a result of their slavish and cowardly dependence on market leaders.

What do you think? Best of breed or one vendor - or outsourcing company - whenever possible?

** Martin Brampton is a director and founder of Black Sheep Research (www.black-sheep-research.co.uk ), an independent consultancy providing research, writing and speaking services on a wide range of business and technology subjects. Martin was previously a director at Bloor Research, and has worked with IT as a user and analyst for over 20 years. He is a frequent contributor to silicon.com's Behind the Headlines TV programme and can be contacted at silicon@black-sheep-research.co.uk.