The Journey of Business Intelligence
Business intelligence has many names. The most used is Analytics and as an IT guy, my perspective on Analytics.
Analytics for me is presenting information in a way which aids decision making, which aids in finding out what is wrong and what is right about the business, which aids in in finding out what can go wrong in my business.
The first and foremost is to ascertain the objective of taking a step in the Analytics direction. Ask yourself, Is this a business need or just another IT project. If the answer is “Business Need” then there are 90 percent chances that the project will succeed (though the product chosen, quality of data sources, business inputs, timelines, costs, partner skills do matter in the overall success).
One goal has to be set and then broken down into phases in order of business priority. Let’s say, I’m a manufacturing as well as a trading organization, which stream do I take first? What is my principal worry at this time? My problem could be quality of my products, it could be my after sales support, it could be sagging sales and it could be productivity on the manufacturing lines.
Let’s focus on Sales. It could be further broken down, do we need to focus on my supply chain? Do we have stock outs in the market? Do we have unsold inventory in the market? There are umpteen questions for which one needs an answer, which we may not be getting today. We might be getting these answers in pieces and are not be able to put together the complex jigsaw puzzle. Hence ‘Start Small But Scalable’.
One of the most important underlying layer of the journey is the Project management. This will decide the way forward will lead to chaos or timely delivery. Today, Analytics is the buzz word and as soon the word spreads in the organization, everybody starts talking about it, it becomes a fad hence everybody wants their analytics first. The word of caution is to divide the project in phases and stick to it. Take the Top into confidence and understand their view of the business and challenges.
Create a Core team: Business and IT, however, the reality is, business has their own priorities, operational priorities. Get the Top to intervene. This project is like an implementation of an ERP. Pick up the people who have a dire need of analytics and sufficient analytical skills. Do not commit mistake of roping in MIS executives (who bring data from many sources and then collate in excel).
Organize a KPI (Key performance Indicators) Workshop: With peers at an outside location. Ask business guys to bring in the current state of analytics (mostly excel). Deep dive from a perspective of intertwined KPIs. Focus on the said subject ONLY. Prioritize and then present the output to the Top. This exercise will help you to make SoW for the project.
Now, you need to get the Product vendors. There are countless of them today. Standardized products by big organizations who find their way into Gartner’s quadrant and many of them using open source platforms. The decision lies in evaluating them according to your business needs and your budget. Remember not to make this project a “White Elephant”.
Choose an Implementation partner who is well versed with the product you chose and look at their previous experience in a similar industry.
Caution: References provided by the Product vendors as well as by the implementation partners are usually of the people with whom they have good relationships. They will never provide references where it went wrong. Hence do your own research. Good way would be to create the parameters of evaluation of Product as well as Implementation Partner, assign weightage to each parameter and then ask the CORE team to give points individually then collate and decide.
This time around, one needs to ascertain what data the organizations has. The data is usually spread across the organization in pockets. The activity of identification of data sources needs to be undertaken as a first priority.
Once, the data sources have been identified, the most pertinent question comes to mind whether the organization needs Data Warehouse or an Operational Data store. This is a very tricky question from both time and cost perspective. There are products which claim that they do not require any DW and can get the data directly from the data sources but my view is that you should create one. Again, vendors offer big products, choose the best available alternative which fits your budget.
Once you kick-off the project, monitor it very closely supported by steering committee meeting every fortnight and that is the key to success.