Over the past year, there has been a lot of talk about agile BI again. It seems to have become a buzz word for what is a requirement, a business requirement actually, that needs to be supported in IT and by pretty much everybody in the organization and that’s agility, time-to-market, the ability to adjust deal with dynamics in the market and so forth.
You can think of it more broadly than the original term for agile software development. Agile software developers have been around for a long time. And if you just take the software development side, you flip that over to business intelligence and data warehousing, you are probably not going to be real successful with it. You have to look at it much more broadly. There’s a difference between being agile versus doing agile.
If we just do agile, we are probably going to be less successful than if we be agile. How can you look at agile as a word versus a methodology? Where do we need to be more agile? Because one of the things that agile doesn’t do a very good job at is looking at the data model.
Next you can look at analytics. How do we be more agile and how we deliver analytics to our business users? We talk about dashboards and scorecards and static reports. Those probably aren’t really agile to our business users so how do we get more agile in our delivery of our analytics?
The next thing that is the very important is the agile organization. We talk about one of the pillars of agile is collaboration, and one of the things that I think makes it successful is true collaboration between business and IT. So how do we be more agile across our organizations and get them working together?
Then there is our infrastructure. How do we be more agile in infrastructure? Look what’s happened over the last three to five years. We have got cloud computing. We have got software as a service. We have got platform as a service. We have got all these different services components. So how do we be more agile in how we implement our BI data warehousing solutions?
And by doing that, the cost, if you look at the cost components of cloud and all these pieces, it actually kind of levels a playing field and allows more small to medium size businesses actually be more competitive more quickly because they don’t have to build that infrastructure internally.
Part of the answer has to do with emergent soft skills and crossover skills between business and technology. We have always taken the model of business plus IT where maybe we can just separate the I and the T also and focus around the information. So I think it’s based on a lot of that foot in two camps, the collaborative states.
There is the older agile methodology, agile development, and there is the agile manifesto, and that’s been with us 12-15 years now, and there are lots of especially large organizations with agile development teams, but they are not doing BI. They are typically writing a lot of code. It is procedural code. The procedure is typically automating business processes and quite often automating transactional systems.
It’s very code heavy, and there was a reaction years ago to try and streamline a lot of the code based methodologies. If you think about things like code review, peer review, if you think about the way that code is promoted from development to test to production and all the bureaucratic processes that go with that.
So a lot of people in operational applications came up with their own methodology, and a lot of people pushed back and said no, agile is not really a method, it’s really a set of goals. And I will go with that, I mean let’s not belabor that too much. But the goals are very relevant to us even though we are not doing the same kind of heavy hand coding they are doing. A certain amount of heavy hand coding we do, yeah sure but nowhere near what those guys are doing.