Driving efficiency for the Yext billing team
Scaling capability to keep pace with rapid company growth
Role: Product Designer
Background
Yext is a public marketing technology company that helps businesses manage their online presence across platforms. During a period of rapid growth, the billing team needed to manage invoices and collections across thousands of customers – as well as increasing business complexity – without adding head count.
With a small team and so many invoices to manage, the Yext team needed to ensure their time and energy was deployed efficiently in order to support the company as it continued to grow.
How might we…
Help the Yext billing team keep up with the company’s growth?
Research and iteration
Through our regular check-ins with the Yext team, we learned they had tools to automate outgoing invoices but they were on their own when it came to following up with overdue customers. This might not be a big problem for a company with a small customer base, but for Yext, it meant they were often spending just as much time chasing down a $10,000 payment as they were a $50,000 payment. There was clearly an opportunity to help the team by helping them prioritize customers for follow-up.
We learned from the Yext team that it isn’t as simple as “highest amount first” though. There was another big factor - the amount of time an invoice was outstanding or overdue (its “age”). But when asked whether they would prioritize a larger-but-more-recent invoice vs. a smaller-but-older one, the answer was “it depends.”
Given that answer, we oriented toward supporting their decision-making by presenting the key important information in the most consumable manner possible. The goal at Rivet was always to do more than present historical data, so we looked to our machine learning capabilities for ways to add additional insight.
In doing so, we found that we could highlight trends in customer payment patterns that are important, but difficult to notice - i.e. a customer’s days to pay might be gradually increasing over time. We also learned from the Yext team that it’s surprisingly difficult to know when a customer has filed for bankruptcy.
Solution
We designed Rivet’s “Notable Customers” feature to help billing teams prioritize efficiently by using the criteria validated by Yext to automatically highlight customers for follow-up: Overdue amount, age of overdue balance, trends in payment behavior, and bankruptcy.
For easy access, we made Notable Customers a dedicated section within Rivet. The default page lists the top customers suggested for followup in a table that includes their rationale for inclusion described in natural language for easy consumption. Overdue amount and age of overdue balance are simple computations, but the payment behavior descriptions leverage Rivet’s connection to customer data and machine learning techniques. To stay on top of bankruptcies, we created a web crawler that cross-references public bankruptcy filing records with customer lists.
We also opted to include second tab focusing specifically on overdue balance relative to age. By using a bubble chart to visualize relative overdue balances in addition to their ages, we gave the billing team a better way to understand the relationship between age and balance and prioritize accordingly.
“Rivet has let us scale the billing team and collect cash faster even when customer volume is growing rapidly and and business complexity has increased.”
— CFO, Yext (NYSE: YEXT)
Results and reflections
Feedback from Yext was overwhelmingly positive - not only from the billing team, but from the CFO as well. By automating the process of evaluating priority based on certain key criteria, the Yext billing team was able to scale along with the company’s growth without adding staff. Yext renewed their annual contract with Rivet and expanded their team’s engagement.
There’s a lot more I’d have liked to do with Notable Customers. First, it would have been great to explore how predictive techniques could have been incorporated in order to give the billing team a way to prevent outstanding balances from actually becoming overdue. There’s also room for improvement in the user experience—the natural-language descriptions in the table often get truncated by screen width, and some of the text could have been either replaced with or supported by visuals in order to improve digestibility. It would also be easy to make the feature more flexible for different teams by giving users more controls, like how many customers to include in the feature. Given more time, I’d also have liked to explore how the notion of “notability” could have been expressed throughout the rest of the app.
