MARCH 202319· Ensure the usage of analytics and actions are taken appropriately based on insights.· The last step would be to handover the analytics to Business sponsor team to maintain on regular basis.Some Key Aspects While Doing any AnalyticsPathos is very Important: While as finance function, we are very strong on logic with numbers and have good credibility on financials but what we lack is the emotional connect with stakeholders and this is very important while working on any Data Analytics.Establish Emotional Connect through Story Telling:Story telling is an art to explain the numbers starting with a problem and then explaining the peak of problem and how the data can help to have happy ending. It can be corelated with real life example or a theme to emotionally connect with audience fore.g. A firm who is bleeding on bad margins due to bad pricing for few products can explain to COO's their numbers with a story like ­ "What is our Everest on Pricing of XYZ Product and then go on to co relate on how one would summit the margins with steps to summit the Everest".Fail Fast & Fail CheapAs finance guys, we tend to get into 'Analysis -Paralysis', Reconciliations and may at times complicate the analytics. It is very important to take feedback, respect experience of stakeholders and fail fast to improvise the analytics. At times, one may stop doing the analytics further if it doesn't add value to the organization but only a piece of fancy report.Ensure you use the Right Analytics as per ProblemThere are various forms of analytics, and one needs to determine the right one to use to substantiate the problem. (Ref Fig. 1.1)Reinforce AnalyticsFrom time to time, it is very important that you feed the analytics with additional information to improve the sense of data and hence insights. This is called reinforcing analytics with new data. For e.g. We are working on predicting attrition and a new data emerges from HR about people working in shift then the attrition prediction should be reinforced with change in probability for people working in shift/over time.One Story ­ Multiple InterpretationOne analytics may mean different things to different stakeholders. Hence it is very important to tell the right story as per the stakeholder being addressed. For e.g. Once we did analytics on Quality of hires and how we · Machine Learning· Predicting default probabilityFor eg ­ Predicting Employee attrition basis past behaviour· AlgorithmicRecommendationVC Investing decision ORBank loan approval basis score· Summary Statistics ­ Trend of people work from home· Data Visualisation ­ which days people come to office· High Frequency Trading· Driverless car· Fraud Detection· Regression Analysis· Impulse Response· Variancede composition. Eg- dramatic rise in sales during a particular period or seasonFigure 1.1
< Page 9 | Page 11 >