CRM Predictive Analytics

Home/CRM Predictive Analytics

CRM Predictive Analytics

Leverage data science in digital marketing

 zigihub leverages machine learning and data science extensively across the platform. Statistical modelling provides a scientific approach to identify and convert high potential leads and provides insights to retain and grow existing customers.

Our data science team follows a comprehensive data discovery and statistical modelling process to ensure the best results.

Micro Segmentation

Customer is getting more demanding than ever. For every marketer, it is essential to understand the transaction behaviour of the customer with your products or services to address different segments of the customers accordingly. This not only ensures proper attention to high value customers but also to address concerns of the customers who may be unhappy and likely to leave (Churn).

There are multiple approaches to segmentation in the platform, but the prominent amongst them are based on the purchasing pattern of the customers  (Usage Segmentation) and other a comprehensive purchase and profile of a customer (Profile Segmentation)

Usage Segmentation

Usage Segmentation

zigihub’s segmentation approach to usage segmentation is to model the Recency (how recent was the last transaction), Frequency (frequency of transactions) and Monetary value (average monetary value) of transactions of the customers. The segmentation algorithms automatically identifies the purchasing pattern of the customers and creates micro-segments like “Valuable”, “At Risk”, “Elite” etc. as shown below:

The segment of each customers gets integrated for all the stakeholders that interact with the customer like sales, support, service etc. These segments are also very helpful in creating and running targeted campaigns to ensure revenue growth from valuable customers and reduce churn of high risk customers. The event driven marketing module of zigihub ensures that each segments gets notified automatically with a proper campaign or response.

Profile Segmentation

Profile Segmentation

Profile segmentation approach leverages the profile as well as the usage behaviour of the customers to identify the similar clusters of the customers.The clustering algorithms creates that relevant groups of customer showing similar behaviour. This is very helpful to identify the parameters that leads to customers in a particular cluster and have marketing strategies too.

Recommendation Engine

Customers love the power to choose from multiple options presented to them. zigihub’s recommendation algorithms recommends products and services by using 2 different methodologies:

User Based : Recommends products based on preferences of a customer and his past purchasing behavior

Collaborative Grouping : Recommends products based on “What other customers have bought” who have similar profile

A proactive approach to present options to customers increased loyalty of a customer and his increased business. The process and algorithms used for these 2 recommendation and their details are as below:

User Based

User Based

User based algorithms takes into account the following parameters to recommend the products

  • Demographic profile: This represents a customer’s age, income, family details, location etc.
  • Purchase behaviour: The transaction history of the customer is analysed to create a preference profile

The algorithms ranks all the products and recommends the top 5 products having highest score

Profile Segmentation

Profile Segmentation

Profile segmentation approach leverages the profile as well as the usage behaviour of the customers to identify the similar clusters of the customers.The clustering algorithms creates that relevant groups of customer showing similar behaviour. This is very helpful to identify the parameters that leads to customers in a particular cluster and have marketing strategies too.

Successful organizations have a very high share of their business coming from existing customers. This not only gives a predictability of revenues but also keeps the sales costs under check. How to sell more to existing customers can include either selling additional products (Cross-sell) or upgrading the customers to a higher version of his current product (Up-sell).

Statistical Lead Scoring

Statistical Lead Scoring

The potential of a lead to convert into a customer is the most important parameter for a sales person to focus his efforts on. This potential of lead depends on numerous parameters and a data science driven lead scoring approach ensures that all the variables are considered while scoring a lead.

zigihub has developed a lead scoring engine based on more than 100 parameters related to demographics, behaviour patterns, interactions, social presence etc. This exhaustive machine learning driven approach ensures that the sales person focuses his efforts on the high potential leads and increased the conversion rate.

Customer Churn Prediction

Customer Churn Prediction

It is 6 times more expensive to acquire a new customer than to retain and grow an existing one. This this era of dwindling customer loyalty, it is important more than ever to reduce the customers leaving you.

Knowing which customers are likely to leave in advance and take appropriate actions to stop them, can be a great way to enhance revenues.

zigihub’s customer churn prediction model follows and exhaustive process to identify the customer who have high probability of churning. This provides an early warning indication to by assigning a churn score to each customer and categorizing then in “Very High”, “High”, “Medium” and “Low” chances of their churning. Sales and marketing team then focus on probability churn customers and device appropriate strategies.

Sales Forecasting

Sales Forecasting

Scenario driven Forecasting

Achievement of sales depends on the health of the pipeline which is measured based on the deals in various stages like Initiation, Proposal, Negotiation, Closure etc. More deals in negotiation or closure stage will lead to higher conversions. In an easy to use GUI, a user can define the % rate of success of deals at each pipeline stage. This gives him the ability to forecast sales for coming months, quarters and the years.

User can modify the % rate of success to create various scenarios and decide where he should divert his efforts to achieve his goal.

Statistical Forecasting

Statistical forecasting leverages machine learning algorithms to project sales. These algorithms take host of parameters and classify them based on their importance. The model then predicts the sales for the defined period.

Sales Success Modeling

Sales Success Modelling

Sales success modelling guide the teams to identify the parameters leading to a sales person being successful and focus on improving those parameters. The various statistical algorithms models various parameters and identifies the top parameters impact performance of sales teams.

The platform enables sales managers to define and measure the sales activity targets like sales pipeline, quality of pipeline, # meetings, distance travelled, # calls etc.