1. “More than 50% of business intelligence projects fail to deliver the expected benefit” (BI projects failure)
2. “Two thirds of executives feel that the quality of and timely access to data is poor and inconsistent” (reports and data quality)
3. “Seven out of ten executives do not get the right information to make business decisions.” (BI value)
4. “Fewer than 10% of organizations have successfully used business intelligence to enhance their organizational and technological infrastructures” (BI alignment)
5. “those with effective business intelligence outperform the market by more than 5% in terms of return on equity” (competitive advantage)
The numbers reflect to some degree also my expectations, though they seem more pessimistic than I expected. That’s not a surprise, considering that such studies can be strongly biased, especially because in them are reflected expectations, presumptions and personal views over the state of art within an organization.
KPMG builds on the above numbers and several other aspects that revolve around the use of governance and alignment in order to increase the value provided by BI to the business, though I feel that they are hardly scratching the surface. Governance and alignment look great into studies and academic work, though they alone can’t bring success, no matter how much their importance and usage is accentuated. Sometimes I feel that people hide behind big words without even grasping the facts. The importance of governance and alignment can’t be neglected, though the argumentation provided by KPMG isn’t flawless. There are statements I can agree with, and many which are circumstantial. Anyway, let’s look a little deeper at the above numbers.
I suppose there is no surprise concerning the huge rate of BI projects’ failure. The value is somewhat close to the rate of software projects’ failure. Why would make a BI project an exception from a typical software project, considering that they are facing almost the same environments and challenges? In fact, given the role played by BI in decision making, I would say that BI projects are more sensitive to the various factors than a typical software project. It doesn’t make sense to retake the motives for which software projects fail, but some particular aspects need to be mentioned. KPMG insists on the poor quality of data, on the relevance and volume of reports and metrics used, the lack of reflecting organization’s objectives, the inflexibility of data models, lack of standardization, all of them reflecting in a degree or other on the success of a BI project. There is much more to it!
KPMG refers to a holistic approach concentrated on the change of focus from technology to the actual needs, a change of process and funding. A reflection of the holistic approach is also the view of the BI infrastructure from the point of view of the entire IT infrastructure, of the organization, network of partners and of the end-products – mainly models and reports. Many of the problems BI initiatives are confronted with refer to the quality of data and its many dimensions (duplicates, conformity, consistency, integrity, accuracy, availability, timeliness, etc.) , problems which could be in theory solved in the source systems, mainly through design. Other problems, like dealing with complex infrastructures based on more or less compatible IS or BI tools, might involve virtualization, consolidation or harmonization of such solutions, plus the addition of other tools.
Looking at the whole organization, other problems appear: the use of reports and models without understanding the whole luggage of meaning hiding behind them, the different views within the same data and models, the difference of language, problems, requirements and objectives, the departmental and organizational politics, the lack of communication, the lack of trust in the existing models and reports, and so on. What all these points have in common are people! The people are the maybe the most important factor in the adoption and effective usage of BI solutions. It starts with them – identifying their needs, and it ends with them – as end users. Making them aware of all contextual requirements, actually making them knowledge workers and not considering them just simple machines could give a boost to your BI strategy.
Partners doesn’t encompass just software vendors, service providers or consultants, but also the internal organizational structures – teams, departments, sites or any other similar structure. Many problems in BI can be tracked down to partners and the ways a partnership is understood, on how resources are managed, how different goals and strategies are harmonized, on how people collaborate and coordinate. Maybe the most problematic is the partnership between IT and the other departments on one side, and between IT and external partners on the other side. As long IT is not seen as a partner, as long IT is skip from the important decisions or isn’t acting as a mediator between its internal and external partners, there are few chances of succeeding. There are so many aspects and lot of material written on this topic, there are models and methodologies supposed to make things work, but often between theory and practice there is a long distance.
How many of the people you met were blaming the poor quality of the data without actually doing something to improve anything? If the quality of your data in one of your major problems then why aren’t you doing something to improve that? Taking the ownership over your data is a major step on the way to better data quality, though a data management strategy is needed. This involve the design of a framework that facilitates data quality and data consumption, the design and use of policies, practices and procedures to properly manage the full data lifecycle. Also this can be considered as part of your BI infrastructure, and given the huge volume, the complexity and diversity of data, is nowadays a must for an organization.
The “right information” is an evasive construct. In order to get the right information you must be capable to define what you want, to design your infrastructure with that in mind and to learn how to harness your data. You don’t have to look only at your data and information but also at the whole DIKW pyramid. The bottom line is that you don’t have to build only a BI infrastructure but a knowledge management infrastructure, and methodologies like ITIL can help you achieve that, though they are not sufficient. Sooner or later you’ll arrive to blame the whole DIKW pyramid - the difficulty of extracting information from data, knowledge from information, and the ultimate translation into wisdom. Actually that’s also what the third and fourth of the above statements are screaming out loud – it’s not so easy to get information from the silos of data, same as it’s not easy to align the transformation process with organizations’ strategy.
Also timeliness has a relative meaning. It’s true that nowadays’ business dynamics requires faster access to data, though it requires also to be proactive, many organizations lacking this level of maturity. In order to be proactive it’s necessary to understand your business’ dynamics thoroughly, that being routed primarily in your data, in the tools you are using and the skill set your employees acquired in order to move between the DIKW layers. I would say that the understanding of DIKW is essential in harnessing your BI infrastructure.
KPMG considers that the 5% increase in return on equity associated with the effective usage of BI is a positive sign, not necessarily. The increase can be associated with hazard or other factors as well, even if it’s unlikely probable to be so. The increase it’s quite small when considered with the huge amount of resources spent on BI infrastructure. I believe that BI can do much more for organizations when harnessed adequately. It’s just a belief that needs to be backed up by numbers, hopefully that will happen someday, soon.