Demystifying KPIs in Analytics
Working in the Data Analytics domain one thing which I have realized is the first and foremost thing a business demands is to recognize and create meaningful indicators and analyze them before diving into any sort of intelligence.
Let's understand what a KPI is and why it is so important with simple use cases.
1 Suppose you are a captain of a cricket team, and tomorrow you have a test match at the MCG stadium. So in the dressing room, you are discussing strategies for winning the game. Your analytics team comes up with the following stats to help you out.
a) Average first inning score at MCG is 450 runs in the last 5 years.
b) Average Second inning score at MCG is 200 runs in the last 5 years.
c) Out of a total of 180 tests played till now 80 tests are won by the team batting first.
So now your analysis team has given you these numbers to understand whats the trend and take better decisions.
2 Now suppose you are an e-commerce seller selling stuff on major platforms. So how and which numbers will you be interested in
a) Total Sales Value say for the month is 13K $ as compared to 15K $ last year.
b) Total canceled orders for this week is 309 which is the highest of all weeks.
c) 30% of my revenue is contributed by the hottest selling product in my inventory for this season.
But let me ask how are these number essential, oh common it's your business and with the help of these you can grow and expand it, let me say if you recognize your hot selling product then obviously you will never fell short in making sure it's always available in stock.
The list of use cases is humongous as with every action in any field there is a data point generated and the collection of these data point if seen later can turn into an important metric.
Let me make you reailze while going through use cases you have goan through every single bit a KPI comprises of and in terms knowingly or unknowningly you have seen 6 KPIs being used in real world.
Now its just time to consolidate your understanding and devise KPI.
A KPI stands for Key Performance Indicator, and indicates a measure or data point aggregated/formulated on given raw data over a period of time/granularity inorder to showcase important/key performance of the use case.
With use case 1 (Cricket match) :-
Measures : 450 runs, 200 runs, 80 Tests
Aggregration/Formula : Sum(Total), Average
Time Period : Last 5 years,Till now.
With use case 2 (Seller) :-
Measures : 13K $, 15k $, 30% revenue
Aggregration/Formula : Sum(Total), Max(Hottest), total sales by hottest selling/total sales
Time Period : Current Month, Current Week, Last Year, This Season.
Point to note if you would have realised while reading use cases the formulas where in bold and time periods where in itallic, that was done intentionally to catch your attention.
So KPI has a formula, a granularity (time period), a unit and on application to raw data gives a value which a none another then a measurable data point.
Now I hope, if you recieve data sets and have business use cases you will be able to create lots of important KPIs and make buisness understand its performance over past.
Stay tuned as will be bringing more articles demystifying jargons of analytics industry and make you learn and grow everytime you read.
