There is a contradiction, if not a paradox, running through a lot of thinking about BI at the present: On the one hand (as books such as Tom Davenport's Competing on Analytics demonstrate), BI is increasingly being seen as a powerful tool that the best businesses adopt with passion and commitment, to great effect. On the other, repeated surveys and analyst reports show that BI often fails to live up to the business objectives of those who implement it.
At a time when CIOs are investing more and more in business intelligence – with the market growing at around 10% a year – and when research shows CIOs have put BI at the top of their list of major concerns (Gartner), this contradiction is a major worry.
It appears to imply that there is a powerful elite that are investing tens of millions in BI and analytics, doing it right and getting huge benefits. And then there is the rest – failing in terms of business objectives, backward looking, and repeatedly making big mistakes.
But are these perceptions accurate? Some recent research by Information Age suggests that BI is much more effective – even for those outside the so-called elite group of users – than many experts generally suggest. But Information Age’s research also shows that if BI is to work effectively, and live up to expectations, it must be made more flexible.
The first set of figures come from Information Age’s annual Effective IT Survey, which asks end users to go beyond all the supplier hype and give their rankings on what really works.
Of 699 polled, 43% said they use BI tools in their organisation. Of those that use it, BI gets a 92% “approval rating” – meaning that the technology either improved service levels, or had a financially positive impact, or both. 18% of BI users said it did both.
The conclusion from this large survey is that BI is not failing in most organisations – even if it does sometimes fail to fulfil ultimate business objectives.
In a second study, sponsored by BI software supplier Hyperion, Information Age asked businesses that had implemented BI if they actually trusted the information they were using. Over 90% of the 354 respondents were either confident or very confident that the data they are using is accurate. A similarly high number are happy the information they are using is consistent with that used by colleagues in other departments.
The report also explores the issue of relevancy – there is widespread belief that many BI reports are of little day-to-day use. But again, four out of five BI users said the information they get is relevant (although they don’t necessarily read or act upon it).
However, there are some problem areas. About two-thirds of organisations said that the structure of their report had changed significantly in the past two years – yet most also suffered delays changing reports. If a user wants to see the data in a different way, just 10% can change it themselves. A quarter will have to wait three days. A further quarter have to wait a week or more.
While there may be good technical reasons for the delays, such as the underlying data schema or the need to change fields, the users themselves say the main reason is that the IT department either gives the need to change reports a low priority, or it suffers from a lack of resources.
Inflexible tools are also a problem – which may be a consideration for managers looking to buy a new BI system. Half of the respondents say that their decision making and agility suffer because of delays in changing reports.
When asked which features they would most like to see in their BI systems, users most wanted “the ability to interact more with my reports” – again a demonstration that they want more control. The also want more customised reports, and more forward looking reports.
The research also uncovered one finding about BI in practice that is perhaps not seen in the hype surrounding the subject and the technology. Nearly 60% of respondents reported that they receive at least some of their reports in an Excel attachment, with large percentages getting their reports either in the body of an email or as a pdf (Acrobat) file or fax. It is possible for BI to be simply effective without being predictive, real time or embedded in a portal, dashboard or application.