As per Gartner, Business intelligence and Analytics remains the top investment priority of CIOs for the fourth consecutive year. This probably justifies and reassures all the buzz on data driven culture among organizations in the recent years. The terms Business Intelligence and Business analytics are certainly not alien to us, they have been doing rounds for more than a decade now with tons of literature available about each of them. The problem is, they are broadly defined and interchangeably used, that even those well entrenched in the industry sometimes fail to define them with clarity! The bulk of this nomenclature problem arises from the vendors who embrace the terminologies that gives them the best marketing mileage. They are not to be blamed, who wants to sell ‘used cars’ when they can sell ‘Certified Pre-owned’ vehicles! This post is thus a primer for anyone who would like some conceptual clarity on these famous industry jargons.
I would try and make it as simple as possible. So here we go. Let us assume that you have a small chain of grocery stores across the state. You maintain the sales & operations data in excel sheets. You generate too much data but too less insights. One fine day you wish to understand your business better and not just overall top line and bottom line. Like for example,
Which of your shops are doing more business on weekends ?
Which locality shops are more profitable ?
Which product categories are selling more?
When are your shops making more business (in a year/month/week )? etc.
So what would you do? You would bring together all the sales & operations data from all of your shops to a common location(Could be as simple as collation of multiple excel files in your laptop), do some necessary aggregations. For example you will sum total sales made by an outlet or sum the sales of a product category(soaps,biscuits etc.) Now you have all your what, when, how much, where questions answered. You may further draw a bar chart or pie chart or a graph for better & easier understanding. You will score top marks if there is a quiz about your business.You wish to automate these tasks. You hope someone would mail you these reports periodically; after all it feels great to be informed! Welcome to the age of Business Intelligence systems.
- Bringing together all the data is called Data-warehousing and this is at the core of any business intelligence system. If you copy your data from several excel sheets into one master sheet then that becomes your data-warehouse. Now replace excel sheets with different databases, flat files, public data the concept remains the same. All kind of routine and ad-hoc reports are generated from data-warehouses and ETL tools usually facilitate the data-warehousing process.
- Displaying the information in a visually appealing way to help understand the data better & faster is Data visualization. Appropriateness in the selection of graph/chart/infographic usually differentiates a good visualization from bad ones. For example, Pie chart communicates relative performance better than bar charts while bar charts are great to see trends!
- A well grouped set of visualizations that indicate the most important information for a specific business context is what we call as Management Dashboards or Executive dashboards. The best of dashboards communicate more and more relevant information in less and less space so all you need is one quick glance!
Everything we discussed till this point used to be known as decision support systems(Used cars) but now we call them as Business Intelligence systems (Pre-owned vehicles) also in a more jazzier version as Descriptive analytics (Certified Pre-owned vehicles)!
To summarize, broad set of tools, technologies and techniques that provide complete visibility into your business operations forms the Business Intelligence layer. Based on the scale of your business and volume of data, the tools differ, the databases differ, hardware platforms may differ, method of data- warehousing may differ but the concepts remain the same. For most small businesses good old Microsoft excel is more than sufficient for all their business intelligence needs!
Now you have understood which days of the week your shops are performing better and which days they are not. Now you want to understand WHY is that the shops in a locality ‘X’ are performing badly on certain days. If you understand that, then you can probably make them perform better or at least you will be able to predict your business outcomes better. Now we enter the Business Analytics space.
Answering the ‘WHY’ always takes you into the realms of data sciences/statistics! All those tools and techniques that helps you statistically answer your business questions falls under data analytics. Business Intelligence ends with the basic mathematical operations (Sum, Average,Percentages) on your dataset. It does not answer much deeper questions on your datasets that require statistical operations like regression analysis, correlation analysis and numerous other advanced techniques.
For example you may want to understand whether the sales performance of those shops are correlated with advertisements & promotions. You run a statistical test and the results suggest a strong correlation. You now realize that the sales dip on those days could be because you don’t advertise on those days like you do on most days. What you just did is diagnostic analytics!
Now you will increase your spending on advertisements, start advertising on those days and monitor through Business Intelligence systems whether or not the sales performance improves on those days. This is just a simple example of data analytics but in reality it helps you investigate your past and predict your future in numerous statistical ways.
WHAT is happening to your business = Business Intelligence (For Visibility)
WHY it is happening, WHAT WILL likely happen in future = Business Analytics (For Investigation, Prediction & Prescription)
BI(What) –> Diagnostic analytics(Why) –> Predictive analytics(What will) –> Prescriptive analytics(Next best action) is the path smarter organizations adopt and rightly so! Do every business needs Diagnostic, Predictive and Prescriptive analytics ? Probably not, but do all of them need a robust BI system ? Probably yes! Business Intelligence system provides the solid base for most of your higher level analytics. ‘Where-you-are‘ is always the right starting point in understanding ‘why-you-are-there‘, ‘where-you-want-to-be‘ and ‘How-you-can-reach-there‘!
In a subsequent post we shall discuss whether or not the latest rock star ‘BigData’ is something from a different planet or just a logical evolution of Enterprise Business intelligence stack!