No, not that type of BI, I’m referring to the trend of information management and technology towards what is called Business Intelligence. Business Intelligence, or BI, is not simply software but rather an entirely new way of thinking about data mining. Traditional forms of data mining usually involve relational data models and custom generated reports. The intelligence of the information these systems produced was limited to the forethought put into the logical data model and the way it anticipated the questions management would ask of the data once that model was set in place.
Unfortunately, the drivers of data modelling were, and still are today, driven by the issues that are facing managers today, not what might be lurking around the corner tomorrow. Hence time and cost quite often out-muscle quality as the prevailing wind almost always includes the standard phrases “we don’t have time for that”, “what do we look like we are made of money”, and “just do the minimum necessary to deliver X before month-end”. It is usually these same organizations that then wonder why they are constantly reinventing their data-warehousing models every 9-18 months. Seems there is never enough money to do it right but always enough to do it over.
What BI brings to the table is a more holistic approach to data analytics. The standard method data analytics is to work from the data model layout outwards to custom, usually one-off, type reports. The BI way is to start with the key management questions that need to be answered and work backwards to the data model.
This is a substantial difference in that it requires leadership to really think about what is important to the strategy, tactics, and operations of the organization. The process involves thinking about the analytics required to answer those questions and in turn what data is required to drive the analytics. The word “report” is something that should never be mentioned in a BI planning session because in a well-designed BI environment, there are no reports. There are only views of the metadata and methods of interpretation.
By focusing on methods rather than specific outcomes, the permutations of data views can be near limitless. Most likely only a few dozen may be required at a specific moment in time, but so long as the class of management questions that need to be addressed can be adequately described early on in the process, there is very little need to have to redesign the data warehouse.
Beyond this, the BI method becomes more strongly tied to the organizational strategy. Instead of asking “this is what I’ve got what can you give me”, the more competitive question in the BI world is “what am I missing and how do we get it”. This is again because the driver is focused on business outcomes not IM/IT expediency.
So enough of the background stuff – where is all this going
Gartner got a write-up in CBR Online on January 6th as to their prediction that
By 2013 33% of BI functionality will be consumed via handheld devices. Initial mobile BI capabilities will focus on existing reports and dashboards ported to the mobile device, Gartner said, but by 2012 organisations and vendors will develop mobile analytic applications for specific tasks or domains. This should help take BI and analytics to a more mainstream audience, which in turn will mean more investment, according to Gartner.
Nice pipe dream but there are three very important points that Gartner failed to address. 1) Audience, 2) Culture, and 3) Security.
Audience: First off – which organizations? There is a very big distinction between the capabilities of small and medium sized enterprises (SMEs) compare to those of multi-national enterprises (MNEs). Bi is not a simple matter of just waiving a magic wand and it happens. Poorly designed BI systems are simply fancy reporting engines. If an organization wants people to think they are on the cutting edge the easiest way to give that appearance is to put some nebulous statistical report through Cognos, create a couple of rudimentary analytical cubes, and then stream to a mobile device. Technically they have delivered functionality using BI but what they have really done is taken an elephant gun to shoot a fly with corresponding similar results.
Culture: If anyone is in a good position to make a prediction of this nature come true it is SMEs. Primarily because niche players tend to substitute entrepreneurial spirit for capital in the production of goods and services. SMEs are better capable of driving innovative change as there are fewer hierarchical barriers to get through in order to make real change happen. The leadership of these organizations are more in tune with their operations and as such are more dynamically involved at a level that makes true BI applications work. Plus they have a greater vested interest as they are more strategically driven rather than tactically driven.
MNEs by contrast have to break down barriers of resistance to new ways of doing things. Large organizations have established bureaucracies which view change in very long cycles. While the implementation of the technical aspects of BI may receive very little resistance within the workplace, the cultural change required to develop BI solutions that take best advantage of these methodologies requires strong buy-in and leadership from the top. That means typically a 3-5 year cultural shift being driven by a consistent business plan and message to the troops. Considering that most executives at that level will typically move placements every 18 months or so, that is a very tall order for an MNE to overcome. In short, the odds that MNE’s are going to represent the bulk of those moving to this ‘new BI-order’ are somewhere between slim and none unless they are already on that path today.
Security: What is going to make the job even harder for MNEs to embrace this new BI-order on mobile devices is that security of transactions is still in its infancy. A well functioning BI system exposes a much broader spectrum of metadata to the outside world than do the simple transaction processing apps that have just in the last year made their way onto mobile devices. Protection and confidentiality of personally identifiable information is much more difficult BI systems because it requires substantially more layers of security than does a typical reporting mechanism. More layers means a larger number of points of failure in the system which could, if an organization was not careful, risk the exposure of hundreds or thousands data points.
Since BI specializes in methods of access, not just the raw data, that could potentially mean would be thieves of such information could not only steal gigabytes of raw information but also use those same methods to provide more relevant information than what raw data might be able to provide on its own. Consider that most SMEs take substantive short-cuts in security and development by comparison to the lengths a MNE will, all it would take is one high-profile theft to put the brakes on all development globally. Not necessarily stop but certainly it would slow things down dramatically.
It is true that as more organizations start to find product means of putting BI-rich applications onto mobile devices, the more the consumer will demand of other companies to do the same. My feeling however is that, 2013-2014 is an extremely aggressive target. It appears to ignore the types of organization that is leading this charge and the cultural shift that will be required to get there. It is, in my mind, a best case scenario of which any project manager will tell you the best case rarely happens in information technology.
Could I see these numbers being surpassed by the end of the decade? Sure but if it were me I’d probably drive out the predictive analysis to range between 2014-2016 after which we should start to see an exponential rise as public confidence in these types of technologies solidifies. Public confidence which btw also follows its own timelines and not that of Gartner.
Still – will be interesting this next decade me thinks.
– Kevin Feenan