Let's start by answering the titular question. Well in Wikipedia this is how it is described (shown in the chart below):
In statistics, a bimodal distribution is a continuous probability distribution with two different modes. These appear as distinct peaks (local maxima) in the probability density function.
My interpretation of this is it means by spreading your options you are more likely to cover off all possibilities and a more literal reading of the graph below would mean there are bumps in the road ahead.
In a recently released analyst report it was recommended that a bimodal approach to BI implementations was the way forward: so how does this relate to our world, and more specifically Qlik? Is it as relevant as it was when I joined Qlik over eight years ago?
When I joined Qlik in 2007 we were so successful because of the fact that organizations were not getting results from their existing BI stack vendor. We were welcomed with open arms by the business but greeted with barely disguised hostility by many IT departments. The reason for this was that when you dig under the surface these departments expected their incumbent technology to be a one stop shop. They did not want to support another technology as this could potentially be seen as a failure on their behalf and another tool for them to support.
Things have moved on a lot since then, IT have had to evolve and be more open to this bimodal approach that supports both the traditional approach and the need for speed and agility. An orchestra cannot be one person, and in a time that we need flexibility and the ability to react in hours not days: the bimodal approach is essential to satisfy these needs.
Now I will insert a note of caution because what we don’t want to see is a proliferation of tools that can be (and often are) organically triggered by a BI implementation and is symptomatic of a badly performing legacy vendor. There are plenty of BI tools out there that are agile but include no governance and can pose a serious security risk. Just because a solution is agile doesn’t mean it is so stripped down that it just becomes a liability. No one would drive a Formula 1 car to work. There is only one thing that causes a proliferation of agile, but ungovernable, tools and that is a tipping point where the agility of the data discovery vendor outweighs the advantages of the incumbent legacy vendor. From there, analysis anarchy ensues.
In summary I believe that a blended or bimodal approach is most definitely the way forward and can potentially achieve the nirvana of keeping both the business and IT happy. In effect, a perfect marriage of agility and governance.