Designing a predictive model to improve customer satisfaction
About this project:
Predicting the likelihood that our customer may be unhappy creates a better user experience and saves us money.
"It's so rewarding when our partners see real business value from their Nurture projects such as identifying upset customers and proactively trying to put things right... before it escalates!"
Marie May
Head of Nurture
- Reach out to Marie to find out how Nurture can help your business to achieve goals
Executive Summary
Predictive Model | Customer Satisfaction
Objective:
At Onward, we aim to deliver excellent service, but occasionally fall short. When this happens, customers may escalate issues through a formal complaints process, which is costly and signals service failure. We set out to build a model to predict when customers might become dissatisfied and consider making a complaint.
Solution:
Our Nurture student partnered with the customer resolutions team and a data analyst to develop the model. She navigated complex processes and datasets across the business, quickly gaining the knowledge needed to ensure the model was comprehensive, accurate, and useful for predicting potential complaints.
Results:
- Improved service provided for the customers
- Using data to target resource in an informed way
- Encourages innovation - colleagues across Onward can see what is possible with robust data!
Challenge
Solution
Outcome
At Onward we seek to offer our customers the best possible service; sometimes, we don’t always get things 100% right.
Onward customers can request issues to be escalated via a formal complaints process, something that is both costly to the business and a sign that we’ve not delivered the level of service our customers should expect.
The objective was to work on a model that would predict the likelihood that customers might be unhappy and look to make a complaint.
Improving user experience whilst reducing costs
To address this we asked our Nurture student to work on a model that would predict the likelihood that customers might be unhappy and look to make a complaint.
She worked closely with members of our customer resolutions team to deliver the project, scoping alongside one of our data analysts and using data from across the business.
This was no small task, having to quickly familiarise herself with the broad range of processes and datasets available.
Her commitment during the analysis and modelling phases ensured that the model was as robust as possible.
Broad range of processes and data sets
Our resolutions team were really impressed with outputs from the initial work, so much so that we were able to commit to deploying a production version of the model.
Early feedback of this in action has been positive, with our resolutions team highlighting that the model allows them to target and resolve customer issues at the earliest possible opportunity; customers have commented that they’ve been impressed with how proactive Onward has been in trying to put things right.
