|Customer: An IT Consulting Firm||Profile: The client offers comprehensive IT management and consultancy services to organizations worldwide, empowering them to grow their businesses, manage risk and compliance and increase their competitive position|
Services: Data Visualization and Predictive Modeling
|Size: 1,000 employees|
|Region: San Diego, CA|
|Industry: Information Technology and Services|
The IT Consulting Firm has been working in partnership with organizations from multiple industries for over two decades and focuses on providing maximum business value to its clients. This strong partnership leads to the IT firm earning a significant part of its revenue from recurring customers.
To further strengthen the bond with the customers, the IT firm decided to undertake a project that helps identify reasons (variables and attributes) that potentially contribute to some of its customers leaving. This understanding had to be further leveraged to identify and address the dissatisfied customers’ concerns and problems in time.
The project started with the study of the historical data of already churned customers. The intent was to read patterns and trends behind these customers’ dissatisfaction.
Solution and Approach
To start, Synoptek’s Business Intelligence (BI) team helped the IT firm perform a preliminary exercise of analyzing data of 15 customers that got churned between 2018-19 from diverse Monthly Recurring Revenue (MRR) ranges, business units, regions, and industries.
This exercise was further extended to 500 Active as well as Churned customers. Customer data across tickets, feedback, and survey scores from the past two years were analyzed. To understand the reasons for these customers’ dissatisfaction and discover patterns and trends, Synoptek used the Power BI and R tools. Power BI tool was used to visualize data, and the R tool was used to develop predictive models. The algorithms and modeling techniques Synoptek used to perform this exercise are Decision Trees, Random-forest, and Neural Network. Through this exercise, Synoptek helped the IT firm create a portfolio of its customers – and understand who they are, filter out reasons for their churn, identify patterns and trends, and give recommendations around corrective measures to improve relationships with these customers. The predictive models helped the IT firm classify active and churned customers using historical data. Depending on the attributes’ pattern, the model provided a probability score to customers’ chances of getting churned, allowing the IT firm to address the customers’ areas of concern in time. The model helped classify churned customers by the following characteristics:
To get a deeper understanding of the variables that contributed to the churn, Synoptek also analyzed customers’ tickets’ data of the IT firm from the following angles:
Through this project, Synoptek helped the IT firm get a more in-depth insight into its customers’ issues. It also enabled the IT firm to make timely and well-informed decisions to keep its customers engaged and satisfied. The timely intervention helped the IT firm in minimizing customer churn and drive value from the investment it had made in its customers.
By using this data, the company’s business development managers, client advisors, and sales teams were able to plan strategies to address the customer grievances in a more meaningful and timely manner.
As this model is in a mature state, the IT firm now runs it against all its active customers periodically to identify the customers who can probably churn in the next cycle/month. This exercise intends to identify and address the concerns of these customers without any further delay. Synoptek, simultaneously, is working on this model to enhance its performance and accuracy – paving the way for better predictions.
Given that acquiring a new customer takes considerable time and effort, companies working towards minimizing customer attrition rates can leverage the same strategy and model.Download Pdf