PrimeRevenue Inc., founded in 2003, is a Supply Chain Finance (SCF) company and a current world leader in working capital financial technology solutions. PrimeRevenue is based in Atlanta and has more than 100 funding partners. PrimeRevenue has a single cloud-based, multi-lingual, cross-border network which supports more than 30 different currencies. These unique features are why PrimeRevenue facilitates more than $250 billion in annual payment transactions.
There are three stakeholders in a SCF transaction: Buyer, Supplier, and Funder.
PrimeRevenue focuses its marketing towards Buyers who need SCF to unlock more working capital and who would like to extend payment terms with their Suppliers. Once a Buyer is onboarded, the Buyer begins payment term extension negotiations with their Suppliers. This is known as the commercial phase of the Supplier onboarding process. Once term extensions are agreed, the Supplier goes through the onboarding process detailed in the system description.
PrimeRevenue has faced issues onboarding clients onto their SCF platform, resulting in a relatively long backlog of over 800 waiting clients. Currently, their onboarding process is not data-driven and their decisions are guided by experience and internal biases.
PrimeRevenue uses a FIFO queueing system to onboard its Suppliers. This system has proven to be relatively inefficient, hence causing a consistent backlog. Since Buyer's approach PrimeRevenue and provide a list of Suppliers, not all Suppliers have the same levels of need for PrimeRevenue's services. This causes some Suppliers to be less motivated to use SCF. A less motivated Supplier takes longer to onboard and thus creates a backlog in the system.
PrimeRevenue has multiple employees called Account Executives who assist Suppliers to onboard onto the platform. They essentially act as servers in a multi-server queuing system who are able to serve multiple Suppliers at one time. Right now, Account Executives don't have a standardized method to track their Suppliers they are serving. This causes a large amount of variability within the system.
Our approach to improve PrimeRevenue's queuing system was to change their First-In-First-Out Queuing System to a Priority Queuing System. This would be done by arranging the queue based on a 'Supplier Score'. This score is unique for all Suppliers and is calculated based on three factors: The Probability a Supplier Onboards and trades on PrimeRevenue's system, their predicted Revenue to PrimeRevenue, and their predicted Time Taken to Onboard. These three factors are predicted using historical Supplier Working Capital Data using the Random Forest model, which is a predictive Machine Learning method. We were able to validate are model by backtesting and by using Python in-built methods. We were able to get an overall accuracy of 86%.
Our approach towards the opportunity created by a lack of standardization was to create an Excel Onboarding Tool for Account Executives to keep a track of their Suppliers in the onboarding process. The tracking tool will also have expected dates for tasks to be completed. This tool will be linked directly to the Queue and will give each Account Executive a unique view of Suppliers under their domain. It also gives their managers the ability to monitor progress. This tool will be linked to a Tableau Visualization Dashboard as well which would increase transparency for Senior Management. Since the average time taken to onboard a Supplier was around 150 days, given the length of our Capstone project we couldn't have a dry-run of the tool. However we conducted multiple Account Executive interviews to get feedback on the tool and cater it directly to their needs.
We provided PrimeRevenue with 3 inter-linked deliverables which work together.
Our first deliverable is the Supplier Scoring Random Forest model. This deliverable acts as the basis of the Priority Queuing System.
Our second deliverable is the Excel Tracking Tool. This deliverable actually displays the queue to the Account Executives and helps each of them track the Suppliers under their domain. It also provides monitoring ability for Management, increasing accountability within the company.
Our third deliverable is a Tableau Statistical Dashboard. This deliverable provides feature specific summaries of the Onboarding process which increases transparency for Senior Management, enabling them to make more informed decisions.
Value and Impact
These deliverables will enable the client to continuously improve their onboarding process by introducing data and standardization. PrimeRevenue has informed us that they plan to integrate most of our solutions, if not all, into their current internal software architecture. On implementing the solutions, PrimeRevenue can reduce the size of their queue by 5%, leading to $100k in yearly savings. The Onboarding Dashboard will reduce the time spent by Account Executive’s by 7.5% resulting in $150k in yearly savings.
Furthermore, PrimeRevenue will benefit from the ability to forecast Revenue and Onboarding Times more accurately, enabling them to plan for the future in an informed manner. Senior Mangement will also have a clearer insight into the Onboarding Process through our Statistical Dashboard. This insight will not only help them understand the Onboarding process better, but will also aid their decisions in other divisions of the company. Most importantly, our solutions will decrease the variability in the Onboarding System leading to system stability which enables PrimeRevenue to make more confident business decisions.