Business Intelligence, in the heads of people, not the software

Many times I've been asked, whether it be at an event or on a call, "What skill sets are needed to use your tool?". These questions usually come from the IT departments of businesses as well as prospective partners. All within their rights, its understandable that they want to have a general idea of what talent they may need to hire or have to learn in order to deliver a successful BI implementation. (The following information, in general, can be used for most BI related topics but it will be positioned from the Pentaho Open Source BI angle.) The information within really does relate to the fact that Business Intelligence is in the heads of people, not the software. The software itself is a commodity, a bunch of 1's and 0's. However, without the proper skills or expertise a BI implementation is bound to fail no matter what software you use.

Usually in response before answering the question directly, I would delve in deeper to understand who asked it and why it was asked. It's important to understand the roles of the individuals. One person/department might be responsible for creating a data warehouse or using data integration/ETL tools; another may be responsible for creating reports, templates and dashboards; and another might be a casual user or a business analyst who will be the consumer of what is developed and offered. These are all very important roles in a BI implementation. Depending on the role, my initial response would be to ask what skill sets are currently available as well as what technologies are already being used and are familiar to the business.

Based off that response I am better prepared to position the software and services accordingly. For both development types and user types, the Pentaho Open BI suite has flexible development and deployment options that offer simple tools and interfaces that can accommodate the novice to the most advanced skill sets. Of course, it would be assumed that a general BI skill set is available. One that understands the terminology of metrics, measures, dimensions, reporting, ETL, etc. One that understands the concepts of BI and why it is important for a business.

At a minimum, for all user types, I would recommend taking the appropriate Pentaho on-site or web-based training. It is a great way to get familiar with the technology and understand the tools.

What is very compelling about the Pentaho Open Source BI suite and its modern architecture is that it is built with open standards such as Java, XML, AJAX, SQL, XHTML, CSS, JavaScript, etc . It incorporates all the data integration and content delivery elements that you would see in a proprietary BI tool, without the monolithic stack of mess acquired from BI vendor consolidation. This allows even those with little or no web authoring skills to create flexible and extensible types of BI applications quickly and easily without any migration issues, from one software stack.
For those with more advanced skill sets, all or just pieces of the software can be embedded into your own web, desktop and even mobile based applications. The application programming interfaces (APIs) available are those that are familiar to developers already. There isn't any proprietary no-standard 4GL to learn and struggle with. (I speak from experience) Depending on the business application need, no matter what software you use, one way or another you will hit a wall with the provided GUI and will have to drop into the code. Cognos scripting language or even the older Powerhouse (Quasar Corporation) Axiant 4GL (IDE that helps "minimize" 4GL coding)- , SAP's ABAP, Oracle and PL/SQL, IBM and the Informix 4GL, SAS and Base SAS, SQR, and Information Builders FOCUS, all are some sort of 4GL to perform tasks that a GUI/IDE cannot do alone. So depending on the need of the BI initiative, would you rather learn another language, in addition to what you already know? Why not use a skill set that is based on an open standard and is widely available in the job market as opposed to some that may have experts (that are few and far between) that are approaching retiring age?

I think the bigger question here is not "what skill set is needed" but rather "what skill sets are 'in demand' in today's market, as well as, which are the skills to have for tomorrow"?

Let's not forget to point out that being there isn't any proprietary BI vendor software license and rising maintenance fees with Open Source software, more of your dollars can go towards the expertise and proper skill sets in the form of services, to ensure the success of your BI implementation.

As you can see, Business Intelligence will always be in the heads of people, not the software.