This site uses different types of cookies, including analytics and functional cookies (its own and from other sites). To change your cookie settings or find out more, click here. If you continue browsing our website, you accept these cookies.
Alteryx used Alteryx to predict how long a support case will take to resolve from the moment it arrives in the Customer Support queue. By generating immediate, actionable predictions on this critical Time to Resolution metric, Alteryx Support is able to measure the workload of each support engineer and assign the case appropriately to the engineer with the bandwidth to solve the case in the shortest amount of time. This Alteryx solution will help Customer Support reduce overall resolution times and ultimately provide quicker and higher quality support to its customers.
Describe the business challenge or problem you needed to solve
The Customer Support department at Alteryx is the first point of contact for many customers when they run into obstacles that impede them from using the Alteryx platform optimally. Responsive and helpful support from Alteryx Customer Support helps create great customer experiences and reinforces the power of the platform.
The rapid growth of the Alteryx platform as the data industry continues to expand has created a need to scale Customer Support to meet customer needs and expectations. In the past, it was feasible to manage support cases with anecdotal data and “quick and dirty” spreadsheets. As the customer base has expanded, these traditional methods became too slow, unreliable, and prone to mistake, creating a risk to the customer experience.
Every day Customer Support faces a deluge of e-mail, ranging from the typical spam to sales questions to important issues that our customers are running into with the platform. Sorting through this pile is typically the responsibility of one or a small group of individuals, whose goal is to clean up the mess and get customer support questions in front of an engineer quickly.
There are, of course, some reports that help with this and ensure the workload is distributed appropriately so that each case can be responded to in a reasonably short period of time, but these only help so much. With four major products and a rapidly growing customer base, the case could be a simple question resolved with a 30-second e-mail response, or it could be a critical Server down issue that would require several meetings and advanced troubleshooting to diagnose and resolve.
Describe your working solution
To reduce overall case resolution time, we developed a predictive model using Alteryx to harness the vast amount of historical case data to make predictions on how long new cases would take to resolve. The predictions could then be used to measure the availability of each support engineer and find the appropriate engineer to resolve the issue the fastest for the customer.
During the proof of concept testing, the workflow was able to make predictions for 86% of cases. The overall error rate was about 90%. This proof of concept shows the ability of the Alteryx platform to leverage its own internal data to develop predictive models that can have a direct, actionable impact on its customers’ experience. This solution has a great deal of promise for Customer Support and with some fine-tuning can deliver on its goal of reducing overall resolution time.
Process, workflow and technical components:
Initial Data Prep
Generate Additional Variables
Append Additional Variables
Describe the benefits you have achieved
By leveraging the Alteryx platform to provide better customer experiences, Alteryx is proving the ability of its platform to directly benefit its customers. Without requiring an extensive background in data science, Alteryx used Alteryx to develop complex predictive models, and test-and-measure Linear Regression, Boosted Models, Forest Models, and Spline Models in a matter of days, not months. As a result of Alteryx using Alteryx to enable itself internally, it has inspired others across the organization to continue to innovate and improve upon existing business processes.