FREE DM Review Site Registration!
Sign-up today and access DM Review on the Web!

Your FREE registration entitles you to:
FREE email newsletters

FREE access to all DM Review content

FREE access to web seminars, resource portals, our white paper library and more!

   

BI for All, not BR for the Few

Open Comment

DM Review welcomes analysts from Quocirca as bimonthly online contributors. Quocirca is a leading primary research and analysis company with native language research capabilities across the whole of Europe, North America and the Asia Pacific region.

"Knowledge is power" as the old saying goes, and many an organization has been built and prospered on that pretext. In the old days, information percolated through highly formalized structures to a group of people who made decisions based on the information provided to them. This was all fine when the decision process could take months without impacting the business - but times have changed, and only those organizations that respond rapidly to the knowledge contained within changing information can now be successful in the mid to long term.

The first issue that businesses are facing is the need to react faster to market changes - today the competitive windows of opportunity have shrunk alarmingly. Internet reach has led to the fact that organizations of all sizes now operate in a far more competitive global environment, so consequently decision cycles are shorter - that more people need to have greater access to underlying business information, presented in a meaningful manner.

Secondly, stovepipe hierarchies have had their day. The perceived wisdom has always been that no one person should have managerial responsibility for more than 15 direct reports, preferably as low as five. Therefore, in a company of 10,000 employees, you have a minimum theoretical hierarchy of four levels with a preferred structure of seven levels. However, the majority of organizations this size will have more than10 levels, due to inefficiencies in the structure. Moving information up such hierarchies is slow and can lead to issues with perception - the knowledge that one person perceives to be contained within a specific data set may not be perceived in the same manner (or seen at all) by the next person up the chain. This leads to the problem of "Chinese whispers," where the original thought becomes so changed as it moves through the hierarchy as to become useless when reaching the decision-maker. Again, the old British army tale of the original message being ,"Send reinforcements, we're going to advance" getting to HQ as, "Send three and fourpence, we're going to a dance" rings true for many a missed opportunity for organizations.

Thirdly, existing vendor offerings have been of a variable quality, yet have all been messaged as the universal answer to an organization's business intelligence needs. Quocirca's research points to the market leading tool for data analytics being Microsoft Excel - a fine personal tool, but one that lacks organizational capabilities in areas such as collaboration. It is also the one part of the standard Office package that the user has the least familiarity with - and the least capability to utilize to its full capabilities. Functionality such as pivot tables and multiple trend analysis are beyond all but the most Excel-savvy users, and so large amounts of useful information lays unseen within the masses of data available to the general user.

For those who can use Excel to its best abilities, corporate intelligence then becomes decentralized, and a failure of a user's desktop or laptop can lead to the loss of intelligence that has not been formally backed up.

Many other solutions fall under business reporting (BR), that is, the capability to present a snapshot of where the organization is or has been but add little real value to these findings. Although BR has a place to play in reporting the status of a business to certain audiences, these same audiences are increasingly demanding to know what this means for the future and how will the business respond to dynamic threats? Pity the poor businessperson standing in front of an audience of stakeholders who can only look backward at what has happened, with no solid background to support what may happen in the future.

Why have organizations arrived at a position where the technology is so patently failing to support the business? Again, Quocirca has found that there are multiple reasons why business intelligence (BI) has not achieved mainstream adoption, ranging from perceived complexities and costs, through to managerial levels not entrusting critical decision-making to those closer to the coal face.

The organizational approach to business intelligence has to break free of these constraints. At the highest level, organizations have to recognize that knowledge in the hands of the few is not power - it is commercial suicide. Once this has been recognized, workers whatever their hierarchical level, throughout the organization, should be provided with decision-making capabilities, enabling them to rapidly feed information into other levels so that decisions can be made more effectively and rapidly.

The next areas that need addressing are data availability and data quality. With much of an organization's data residing in databases that are dedicated to specific applications, there are issues of how to get to all the data required at any one time, and also whether application A's Client_ID field is the same as application B's Customer_Name. Even when this has been decided, the data needs to be cleansed, for example, checking against various metrics to see if "Louella Fernandes" and "Louella Fernandez" are one and the same person.

The need for broad scale adoption and usage of BI means that many existing BI approaches are no longer valid. Large, complex, expensive standalone BI systems will never be pervasive or gain mainstream recognition, for all the reasons that are in the first part of the sentence. To make BI a core component of an employee's life dictates a need for BI to be always available, to be available from within the employee's day-to-day environment and for it to be intuitive in use. What is now required is BI as a service - intelligence and reporting that can be called from within another environment as needed to add value to the task in hand. This service needs to be easy for users to utilize, and here we need to look at what can be expected from various types of workers.

Existing BI systems tend to fall in to one of two main camps - either the tool of choice for the business analyst or for the statistician. Neither approach works when we are looking for pervasive BI - something which is simple and flexible to use, fits within the framework of everyday applications and doesn't require the need to understand first or third order trending or whether a Monte Carlo would fit in better than Bayesian analysis. Shifting the emphasis from the few to the many should not be a threat to the existing specialists' jobs - the business analysts and statisticians will still be required, it will just be that their findings will be understandable by more people, who will have been utilising similar approaches to dealing with information on a more regular basis.

Visualization

Many newer approaches are highly graphical, enabling the user to drag and drop any given variables against each other to see the underlying interactions. Many users work far more effectively with immediate graphical output, as the underlying rules for statistics, trending and data analysis have not been taught to them during their training for core tasks. Whether the graphics are bar and pie charts, heat diagrams, lines or size blocks is neither here nor there - a picture paints a thousand words and definitely replaces a lifetime of macros and equations. However, even modern BI systems can still be confusing for the occasional user, and overall usability can still be improved through the provision of advanced wizards, through better previewing of results and through better advice on what may be the best way for certain results to be presented.

Collaboration

The capability for groups to work together on a problem should also not be overlooked. BI is subject to a modified Metcalf's Law - the value of the resulting information grows exponentially as the number of people working together increases, but only up to a level depending on the type of information. Too many people can just confuse the end result, but a single viewpoint can be too easily swayed by existing perceptions.

What is needed is a system that fits in where it needs to be used - an integrated, collaborative means of rapidly seeing the interrelation between variables for the masses, along with a way of pulling out basic information that can be rapidly sent through to others for further discussion and for decisions to be made based on the findings. The system has to have access to all data that the end user has authority to access as well as information held outside of the business. It should have the capability to be used as a simple drag-and-drop visual system or to be statistically programmed by business analysts. The system has to be able to look forward as well as backward or else it is just a reporting tool. In looking forward, it must be able to provide guidelines as to the statistical accuracy of the information trended, otherwise decision-makers will rapidly lose faith in the information fed to them, and such lack of trust will have a trickle effect back down to users who will no longer bother utilizing the tools. It needs workflows to ensure that necessary sign offs are obtained before any business impacting action is carried out, yet also needs to be empowering enough such that decisions that make an end-user's life easier on a day-to-day basis can be carried through immediately and captured as best practice as necessary.

A tall order? Not really - but vendors will have to realize that selling thousands of seats of a system at a low cost can be far more profitable than a few seats at a high cost. Organizations will have to change their mind-sets, accepting that hierarchical decision-making is more often than not too slow and that capabilities to bypass the hierarchy are necessary. Knowledge in the hands of the few is too dangerous in today's markets, while the old vision of the totally empowered workforce has not demonstrated any real advantages. Empowering the workforce to make decisions within their realm of capability is fine, but empowering them to rapidly find valuable information and get this to the right person at the right time for a corporate decision to be made is what the new generation of BI has to be about.


Clive Longbottom is head of research for Quocirca Ltd, a leader in business and IT analysis in Europe. Longbottom is an internationally recognized analyst who brings together large organizational experience with extensive IT knowledge to cut through current flavors of the month, ensuring that clients concentrate on the technology required to support business needs. You can reach him at clive.longbottom@quocirca.com.

For more information on related topics, visit the following channels:



Industry Vendors