Using Machine Learning for predicting subscription churn and identifying high value customers
How a company leading in virtualization products utilized Machine Learning to:
- Improve productivity in its subscription renewals by predictively identifying customers at risk for churn
- Detect potential high value customers by analyzing e-commerce data, and create a targeted and proactive customer growth strategy for the same
The client is a leader in virtualization products- specializing in corporate IT systems and cloud service providers. Their products are used in many enterprises across industries and are recognized for driving innovation. This is echoed in the client’s own company culture, which is keenly focused on fostering innovation and change and prompted their Renewals and E-commerce Teams to look to Machine Learning and Artificial Intelligence in the cloud as a means of showcasing innovation to help drive business benefits.
This client looked to Google Cloud and Pluto7 to help them identify the right business use cases for driving transformational change. Pluto7, using their deep Machine Learning expertise, ran workshops with the customer to establish the root of their needs. It was decided that the client would focus on the following two use cases to achieve the transformation they needed:
- Use service contract and subscription renewal data to categorize subscription customers by their risk for churn, and then use the same data to have a focused process for handling their renewals based on their risk rating.
- Use e-commerce and external data to identify purchases made by individuals, and from this, identify potential high value customers for the future (known as ‘diamonds in the rough’). Then, develop a customized customer relationship and marketing strategy for these customers.
The subscription renewal process was based on simple rules, and the process could not scale alongside the need to handle multiple low value renewals. The then-current methods could not adequately assign risk to renewals, resulting in loss of revenue.
The client had detected many now well-known companies (such as Uber and Airbnb) that had placed e-commerce orders. The rapid growth of these companies surprised the client, who had no way to proactively detect these companies until they were household names.
Based on Pluto7’s expert knowledge of Google Cloud Platform, advanced analytics including Machine Learning and Artificial Intelligence, and applications development, the following approach was taken:
- Utilize Google Cloud Platform to collect all related internal and external data required to solve both use cases.
- Leverage Machine Learning and Artificial Intelligence to predict the classification for subscription renewal contracts and detect potential ‘diamonds in the rough’ examples.
- Use the results of the subscription churn classification to design the process for renewal management.
- Use the results of the ‘diamond in the rough’ detection to drive the appropriate customer relationship management for high-profile companies.
The subscription classification churn model showcased an accuracy within 75- 92% which was promising enough to drive the subscription renewal process changes based on churn risk.
The detection of ‘diamond in the rough’ customers identified 2-3 potential high value customers which was considered a successful use case to deploy into production.
“Today, our renewal process can only go by value and the rest of the lower value renewals are done manually by outsourcing which leaves revenue on the table. If I can get the ability to assign renewal churn risk we can scale our renewal processes regardless of their value and increase renewal revenue.”
- Sales Leader, Renewals/Customer Growth
The client has experienced rapid growth, with its products being used by many enterprises and cloud providers. This growth meant that certain processes- such as renewals- needed to be looked at from the perspective of achieving breakthrough productivity.
The Sales Team also wanted an innovative approach for identifying high value customers (before they actually became high value customers). This would help build a core competency to help define competitive edge for this type of customer, compared to their peers.
What does it take to predict customer churn in a subscription environment?
The subscription based offering market is complex, and the techniques used to manage renewals needs to be different than those normally applied to traditional offerings. The client’s sales and operations teams realized this and decided to take a more innovative approach to improve their subscription rates, increase revenue and scale with rapid growth.
The variety of information that needs to be analyzed to come up with a decent prediction model for customer churn risk is not something traditional solutions can handle. This is because a 360-degree view of the customer needs to be created before any insights related to the customer’s churn risk can be assessed. The 360-degree view includes data such as the historical behavior of renewing subscription contracts, the most recent activities around product usage, sentiments around product usage and more. Machine Learning based classification model is the best method to derive customer churn risk from this 360-degree view.
What does it take to proactively predict high value customers in an e-commerce environment?
Similarly, in order to identify potential high value customers, the 360-degree view of the customer mentioned above must be enhanced (known as Digital Marketing Platform or DMP). This enhanced view is merged with external insights such as employee growth, web traffic, revenue growth and more. This information, combined with internal e-commerce transactional information and customer behavior data, can then be used to classify potential high value customers. This is only achievable through Machine Learning and Artificial Intelligence methods, and is nearly impossible to enact using traditional analytics.
“It took a lot of automated and manual effort to process our e-commerce data to identify potential high value customers, and even when we did, it was only evident once the customers had already become big .”
- Sales Leader, Renewals/Customer Growth
Proactively identify accounts at high risk for renewals, and potential high value customers
With these objectives in mind, this large virtualization products leader partnered with Pluto7 to transform and consolidate their data. Leveraging Google Cloud Platform, they used Google BigQuery to create a Digital Marketing Platform. Utilizing advanced analytics, Machine Learning and Artificial intelligence on top of their DMP, Pluto7 and GCP helped create predictions for customer churn risk for renewal subscriptions and also identify potential high value customers, per expectations.
This laid the foundation for the customer to initiate the change management and planning activities needed to deploy such an innovative process change in production. These are the logical next steps, and will of course take time.
“Google Cloud Platform’s immense processing power, its ability to enable quick ramp up of a Digital Marketing Platform based on BigQuery, with its extreme flexibility, scalability and speed, and its state of the art Machine Learning capabilities resulted in achieving the expected results within the expected 6-8 week timeframe. The results that were seen by the client validated the trust they put in our team and in the Google Cloud Platform.”
- Salil Amonkar, Global Practice Lead Ai and ML and Project Lead, Pluto7
Stronger through innovation
Working with Pluto7, this client was able to take their innovative mindset and apply it to their business, impacting use cases with the leverage of Google Cloud Platform and its components (BigQuery, Dataflow/DataPrep, Tensorflow and Machine Learning Engine). They were able to drive innovative process improvement and saw tangible business benefits.
This client chose Google Cloud Platform because it:
- Offers a flexible, scalable, ready-made infrastructure in the cloud.
- Delivers powerful data processing, data warehousing, and state-of-the-art Machine Learning and Artificial intelligence capabilities.
- Provides a cost-effective platform that’s easy for business team members to use.
- Delivers speed, security, reliability and flexible pricing.
Google Cloud Platform
Google Cloud Storage
Google Cloud Dataprep
Google Cloud Dataflow
Google Machine Learning