A question that is posed by many people is, "what is Agile BI, what does it mean?" Because it is not blindingly obvious, many think it is just another set of weasel words, being casually used to try and dress up offerings, and make them appear more than just the same old stuff being trotted out again. Surprisingly, for those who knew me, rather than supporting this cynical point of view I wish to dispel it, and affirm what these guys are trying to do.
For those of us who have worked in BI for many years, it has always been an area of contention. It does not really fit in an IT department, we lack their Enterprise mind set, going for quick turn around, flexible solutions, we have always been used to users who cannot articulate clearly what they want, so we have been flexible and tried to be responsive. On the other hand we have hardly ever been welcomed with open arms by the business community, because we have still been seen as too concerned with structure, maintainability, documentation and testing to do things fast enough to meet their demands.
Those who are selling solutions that would fit into the Agile BI camp that I have looked at of late include Endeca, who combine search with traditional BI to seamlessly access all data, whether it be in a SQL database or schema less; Tableau, who present the results in a visual rather than a tabular way; and Oryx from Accountagility, who provide end users the tools to extract, validate, manipulate and present their data in a single tool kit without the need for IT intervention, and without moving so far from the end users skill base in Excel as to disenfranchise them from being able to quickly pick it up and produce results.
All of these tools must be doing something right because the established players, instead of just ignoring them, are very quick to point out that they are not enterprise ready, they are just desktop tools aimed at BAs, and other put downs which indicate that there is more than a grain of virtue in what they do. I am not going to say they are perfect and can directly replace your BO or OBIEE suites, but what they are doing is questioning the orthodoxy that is constraining us today., and meeting a genuine need.
So what is Agile BI? It is a fundamental rethink about how BI should work, it is going back to the basics of addressing how to enable someone in the business who wishes to make a better decision, or to understand how the numbers stack up against plans to help themselves, quickly, using skills that they are familiar with, without the need for IT intervention to access data, select the bits that they want, manipulate it to address their need, and present the results so that they and their colleagues can use it to address business problems. In so doing they are going against many of the sacred cows of the past-there is no more single version of the truth, no more having to know SQL to be able to pass go. They are making big strides to deliver results that are quick to achieve, flexible enough to adapt to changing perspectives, and fundamentally are targeted at non-BI professionals in the business.
Where does that leave those of us who work in BI? Well, I think we have to espouse this trend and make it as easy as possible for the business to access reliable data wherever it may reside, by ensuring that through very IT specific skills, like Master Data Management, that all data is part of a unified version of the truth, using things like data virtualisation, making it easily accessible to the business, and to be there to help the business understand the data and the results they are obtaining, so they can apply their business knowledge rapidly and with assurance; and that is what Agile BI is all about.
We automatically stop accepting comments 180 days after a post is published. If you would like to know more about this subject, please contact us and we'll try to help.
7th July 2011: 'Bill Cabiro' said:
I fully agree. Today it's finally possible to perform fast and interactive analytics based on data sources not necessarily residing in a data warehouses or data marts.
Depending on the vendor, this new generation of software is known by different names like Data Discovery, Advanced Visualization, Visual Analytics or Self Serve BI.
A great feature of this class is its data blending functionality. This means that the applications allow connecting simultaneously to disparate types of data bases or tables, whether in a data warehouse, Excel, Access files or even websites.
The results are instant since this process takes place in-memory, so there is no need to develop data marts or build cubes ahead of time like in the past. Additionally the visual interface, when used proficiently, permits to digest huge amounts of information and visualize trends and patterns in seconds.
For Business users this is a dream. It's no surprise that understanding their capabilities, the new software vendors are focusing their marketing effort on the business community and not just on IT like in the past.
I should point out that it's not a good idea to promote these new applications as âIT Freeâ because unless you are in a miniscule enterprise with only one analyst, you will need data governance to ensure a single version of the truth. In this regard, a strong IT-Business partnership early in the game is a must.
Iâve found close to 10 applications that fall into this category from companies around the globe. The leading ones are growing their acceptance at a very fast rate. Four of them have made it to Gartner Research's 2011 Business Intelligence Magic Quadrant Report and I think theyâre going to stay there for the foreseeable future.
In the short term the Visual Analytics packages will not replace traditional BI or Data Warehouses in large organizations but will be a complementary necessity to provide a competitive advantage to the different business units.
For small and medium size companies that havenât yet invested in BI, Visual Data Discovery can be a very fast and effective solution.
Reply to Bill Cabiro?
13th July 2011: 'Ric Ratkowski' said:
I think the your second sentence in the second paragraph sums up nicely the biggest issue with BI implementations (or CPM, CRM, MA, HCM, TM implementations as well).
"...we lack their Enterprise mind set, going for quick turn around, flexible solutions, we have always been used to users who cannot articulate clearly what they want, so we have been flexible and tried to be responsive."
The problem is IT knows the technology but doesn't have a good handle on the business case. It is not just a matter of bits and bytes it is a matter of understanding the character of the business case. ALSO end users can't articulate what they want or the details of the business case and they don't have a technology perspective. Add to that my belief that data is inherently dirty and you have your challenges.
I view terms like "agile BI" "visual Bi", user driven BI"...as just marketing hype created by software vendors (of which I am one) and industry analysts to try to make something that is relatively complex sound easy, when the crux of the problem is the two ideas you identify. IT focus with lack of business perspective and business perspective by folks that can't articulate it.
With that said I don't want to offend anyone, there are folks who have both perspectives, but they are few and far between. Gartner highlights this problem at the BI Summit each year.
Reply to Ric Ratkowski?