Now these best practices and agile development methodologies are all great, but it's very difficult to implement them in the practice unless you also have all sorts of agile technologies. Next generation agile BI technologies can be summarized into four categories: automation, unification, pervasiveness and BI without limitations.
First and foremost, we do need to automate business intelligence because lots of steps in BI, such as discovering the data in the first place, or making BI contextually aware are not often automated. The next one, and it's paradoxical that for many years we talked about ending data silos or information silos in BI. But BI itself is very much siloed because technologies that handle data and content are often different.
BI technologies that handle batch and streaming real time data often are different technologies from those that handle historical and predictive analysis. Complex data structures really don’t work with nice traditional relational or multi-dimensional databases, and therefore we are forced to build multiple data marts or build multiple cubes to make that happen. Last but not least, meta data from our ETL processes for our data warehouse, and from our Business Intelligence and even metadata from our ERP applications is very siloed. That needs to be addressed.
#1 Ranking: Read how InetSoft was rated #1 for user adoption in G2's user survey-based index |
|
Read More |
The next point is how do you make BI pervasive. If you haven’t made it pervasive, well you can’t really compete on the analytic side. Key decisions your enterprise need to make aren’t enabled. The solution is to embed BI in ERP applications and business processes. Embedding BI tightly in your desktop applications such as email and an enterprise portal is important. I already mentioned a huge need to make BI as intuitive as possible, to make self-service by end users possible. Mobile BI is another trends and will drive BI pervasiveness in the next few years.
Early generation BI technologies had too many limitations such as depending on rigid data models which take tremendous efforts to change. They typically do not work the area well with hundreds of dimensions and very complex hierarchies. As we already mentioned the combination of exploration and reporting and analysis typically has been achieved in the past by different technologies. Before you can analyze anything you need to find it, but in order to find it you aren’t going to use a BI application, you use some kind of a search technology. But you can’t really use search technology for analysis and in some of the cases so you keep jumping back and forth between different applications. So hopefully this gives you a good overview of the three legs of the triad of BI: agile software development, agile organizational structures, and agile processes and agile technology.
So how do you go forward with this? Well first of all we definitely recommend making sure that all of your business stakeholders understand why is it that you need business intelligence. There are many wonderful publications out there that do go in-depth as to why information is really critical for competitive advantage. It is true that here are lots of BI applications that still need to follow the traditional approach. Those are anything that is mission critical, such as producing financial statements or printing customer bills or creating a regulatory reporting package that has to comply with thousands requirements.
Obviously these types of applications still have to follow the traditional approach, but elsewhere, agile BI is very applicable. When we talk about best practices, it is all about lessons learned, believe me everyone without exception makes mistakes in BI so we highly recommend that you learn from somebody else’s mistakes and not your own. The best way to do that is to work with professionals that have made those mistakes already, and now you can leverage their experience.
So we think those systems integrators and consultants who understand you and know the difference between where traditional BI approaches need to be applied and where there is room and a actually critical need for agile applications.