Predictive Analysis to Improve CRM Experience
A company's economic success is largely dependent on customer satisfaction. To have loyal,happy customers and to maintain those customers is by no means a simple task. Hence improving customer relations,maintaining customer retention, formulating feedback systems where the customer concerns are reviewed, solved and updated, is what caters to the overall performance efficiency.
It is true that Customer Relationship Management applications, collect information from customers to know about customer preferences, desires and purchase pattern. But predictive analytics provide deeper insights.
Forrester analyst James Kobielus, says that the existing data will improve one on one interactions with customers, also customers analyse the incoming data,and contact the companies which results in additional sales.
For example, considering factors like buying history will be ideal in such cases. if a customer buys only branded, expensive items, the company should know enough not to offer cheap items to the customer, at the same time it shouldn't offer expensive items to someone whose salary can't cover the expense.
The idea is to identify how customers interact with online resources such as social media channels and websites.Focusing on potential customers will eventually increase the profit and prevent the wastage of time, hence marketing strategies are redesigned based on customer interactions.
There are three predictive analytics:
If a person performs action A, then continues to perform action B, what is the probability that they perform action C, where A, B,C are downloading white paper, clicking on pricing page and buying the product respectively. This is found by studying historical patterns of A and B.
If there isn't sufficient historical data for the analysis, we make use of Markov Assumption,which states that "If you consider only the last successful event to predict future behavior, you can assume the probability between that single event and the desired outcome is about the same as it would be otherwise.''
Imagine someone, say 'x' downloaded the white page, clicked on the pricing page and they end up buying the product, according to the assumption any person 'y' who comes after x, has every chance to buy the product, this possibility is reached without even considering if y has downloaded white page, or clicked the pricing page.
2) Cross-Selling :
The idea of cross selling comes from the need of a companion product along with the product purchased.The suggestions often seen while online shopping is an example of the same. Sites like Amazon, Flip kart etc display suggestions like 'people who bought this product also bought _'.
It makes use of analytic data, transaction history over a period of time, to formulate the predictive analysis such that companion products are identified and suggested.
3) Lack of Action :
When the first two predictive analytics dealt with customer actions, this particular analytic deals with customer inaction. The customer may be dissatisfied or unhappy with a service and may decide to break ties, the idea is to reach out to the customers instead of waiting for the last minute.
This is done by creating a fall off model which considers factors like customer responses, response time, purchase pattern, sequencing data etc.
Hence behaviour patterns can be studied and used to improve customer relations. Companies like Microsoft have successfully implemented such systems and enhanced the CRM experience.