Business analytics has become quite popular among all kinds of industries over the past few years. Its all about performing analysis over the data set which enables organizations’ stakeholders to answer certain business questions which ultimately help them in decision making to develop their business.
A company running any sort of business collects a huge amount of data over time like customer demographic data, sales data, sales-force performance data, etc. These data in raw format are of minimal use for the stakeholders unless certain tools are used to process them and bring valuable insights.
Editor’s note: Join our Big Data Analytics and learn more about how and where predictive analytics is being used.
Who is responsible for bringing value out of raw data?
Generally companies have an existing business intelligence team which collate all the data generated by various activities of the organization i.e marketing, sales, etc. and perform an analysis to see interesting trends.
Business intelligence team perform descriptive analysis, which includes applying certain analytics tool to bring the raw data to a meaningful and valuable data set, and study important patterns in the same. Descriptive analytics provides a summarized report of their current business to give insights on what is happening currently in the organization. It may be used to track their progress in sales target of month or a quarter or can be helpful in analyzing performance of their sales representatives. It also helps in analyzing which area of your organization is under-performing and which is over-performing.
Descriptive analytics studies various dimensions of an organization like customers, geographical locations, time, products, partners, campaigns, etc.
What is the requirement of predictive analysis?
Predictive analysis is about future forecasting. It doesn’t mean that it will tell you what is going to happen in future with your company. But it can predict what might happen in the future as predictive analytics is based on probabilistic forecasting based on the historical data. Suppose you are planning to introduce a new product in your company or developing some new features in your existing product. How will you decide that what price your customers will be willing to pay for these added features or your new product. Recently, I came across an article which says Airbnb used predictive pricing to improve user experience. Similarly, a lot of other companies are using predictive analytics in various aspects of their business. It was a news that the e-commerce giant Amazon will predict what its customers need and will deliver the product before they place an order.
So how descriptive analytics is different from business analytics?
Descriptive analytics tells you how many customers are using the product a of your company while predictive analytics tells you how many of these customers may stop using product a after a month. Also remember that descriptive analytics outcomes are always true unless there is some error in the raw data or you make some error while processing it. But in case of predictive analytics the results may not be completely true even provided all the datasets to be errorless. Its just a probabilistic forecasting and the final outcome may be affected by some external factor that you might not have considered.
For example, you may predict a 20% increase in sales by next quarter which may be substantially affected by several factors depending on the product you are selling. If you are an air conditioner dealer, your sales may decrease for a month due to weather change in your city inspite of all your marketing campaigns and sales efforts. Similarly, an immediate recession can affect the sales of your car selling company.
Let’s consider a situation to explain the requirement of descriptive and predictive analysis in an organization. Suppose you have an electric car selling company which has 8 distribution outlets in four different cities.
Your business intelligence team is responsible to collate all the raw data generated by your sales and marketing team and provide you a report on their performance in each city. They may also create a report based on customer demographic details which will help you understand your customers better. You may also track if any particular outlet is under-performing and you can take required actions for the same. While, the predictive analytics team will give you a forecast of what is probable to happen in terms of sales, revenue/profit generation or performance of your sales team in future.
Questions answered by descriptive analysis:
- Which of the 8 outlets sold maximum numbers of car last month?
- Which outlet generated maximum revenue/profit last month?
- How many sales managers could not reach their month targets?
- What was the average number of car sold in a day over the past 2 months?
- How many customers visited the stores last week but did not purchase any car?
- How many leads were generated by our marketing team last month?
Questions answered by predictive analysis:
- Which outlet is likely to sell maximum number of cars next month?
- Which outlets are likely to generate more revenue as next month as compared to the last month?
- Hiring how many sales agents can lead to a 50% growth in revenue in the next quarter?
- What is the projected number of sales lead our marketing team will generate in the coming week?
- What percentage of the leads generated last week are likely to convert?
- Which sales manager has the potential to become top performer next month?
For an organization to run well and develop its business, it is necessary to have a business analytics team to keep a track of what happened in past, but it also very important to have a predictive analytics team to forecast what is likely to happen in future. To perform both it is necessary to have a strong structured thinking and business understanding while you need to be expert in using certain analytics tool depending on the role you play in your business analytics team.
Credits: This blog post is written by Rajesh Ranjan from IIT, Kharagpur.