In tribal societies shamans tell the chief what the future holds, and where the evil spirits are lurking. The tribe of IT directors has such shamans too - they call them analysts.
"Analysts' reports can either be prepared for a direct fee or be within the public domain through the press," says Peter Spens-Black, IS programme director at the Ministry of Defence. "What you read in the press can range from mildly uncontroversial to uncompromisingly eye-opening - you don't pay much money and you take your choice."
Alternatively, "Firms of professional analysts can prepare reports on generic subjects, for example, mobile business communications, and then sell them to organisations," says Spens-Black.
These cost from hundreds to thousands of pounds - reports tailored for an individual organisation and its specific circumstances can easily run with a price tag of tens of thousands of pounds.
But is the tribe getting good value for money - or even good advice from its shamans?
Not always. "I have seen a multimillion-pound project set off in the wrong direction because an initial analyst's report was fundamentally flawed - and £80,000 was paid for the privilege," warns Spens-Black.
So how can IT directors ensure that what they get is what they need? The first essential, says Spens-Black, is to know what business benefits you are looking for and make sure the analyst knows too - as opposed to hoping that he will divine it.
"You've got to have a vision of where you want to be - the analyst is there to show you which paths are navigable. The analyst has to understand your requirements so you must put them in a business context - discuss with them what options to explore," he says.
Make sure your analysts understand anything that may affect your requirements, such as whether you have limitations on your infrastructure or training capabilities, which may exclude some solutions. There is no point getting information about solutions that may be excellent, but which are quite unsuitable for you.
Also make it very clear whether it is just information you are after, or advice. "Do you want them to make recommendations, or just report on the state of the nation in, say, customer relationship management software?" asks Spens-Black.
Take time to make your requirements clear and above all, try to establish a partnership or rapport with the analysts. "Sit down with them and explain what you want, and get on the same wavelength," Spens-Black says. "The analysts must be in no doubt what you want - they may not know all the answers off the cuff, but they must know the questions that need answering."
Think of analysts as a substitute brain, he suggests. You could do your research yourselves, but you have another job to do, so you buy it in. Besides, the need for analysis is sometimes instant - for example you may need to inform a chief executive who has just read an article on a new technology in an in-flight magazine, or dined with the president of an IT company with a new product in search of customers.
An analyst's report of any kind - individually commissioned, bought off the shelf or the results précised in the press - must be carefully assessed against various factors. Spens-Black suggests you should ask yourself whether the perceptiveness of the analysis is "coloured by the fact that suppliers are happy to provide information in exactly the format which an analyst would wish to use - and therefore sometimes you receive a replication of the supplier's sales pitch without an iota of active thought" from the analyst.
Some analyst firms have the reputation - justified or not - of being in the pocket of the IT suppliers, eager to recommend anything, for a price. But even if impartiality is not the name of the game, this need not necessarily be a problem for an IT director trying to machete his way through the solutions jungle. "Although you must select an analyst on the basis of trust or recommendation by people you trust, provided you know where they are coming from and they declare their interest, their information may still be valuable and useful," says Spens-Black.
"You must be prepared to challenge the provenance and validity of the information in the report - has the [report] met the requirement?"
Are the recommendations driven by the analyst seeking to put forward the solution you had accidentally mentioned you would prefer, one he just has a feeling for, or one that is within his own "comfort zone" because he has used it twice before and is happy with it?
"So long as he has explained his recommendation and it maps to your requirements, his report can be valid," says Spens-Black.
Although it is undoubtedly wise to cross-check with a report on the same subject from a different source, in practice that is a luxury that few can afford. "It is optimistic to expect to be in a position to be able to spend another £40,000, say, on commissioning a parallel report on the same subject," says Spens-Black.
"But you can network with other IT directors and get a feel for whether the proposed solution is valid or high-risk."
Sometimes, he acknowledges, the money is spent simply as protection money. "In some organisations it can be the default so that you are not held accountable - you can say you were professionally recommended to implement a particular solution," he points out. But for all that the buck stops with the IT director, not the report.
"We must not abrogate the responsibility for making a decision," says Spens-Black.
Judgement is all. "You need judgement in choosing an analyst, judgement in setting the parameters for the research - and judgement in your response to the report," he asserts.