Onramp Funds Launches Machine-Learning Algorithm After Two Years of Development and 6 Months Test

Onramp, an innovative funding provider that supports the growth of eCommerce businesses, has launched its first-ever machine-learning algorithm designed to identify eCommerce merchants who will best utilize funds in a productive way to grow their business and can successfully repay the loan.  


The algorithm evaluates nearly 240 characteristics, with over 100 additional features currently being developed. The algorithm is trained on data collected from a number of sources, including data from the merchant’s shopping cart and marketplace sales channels, their bank account(s), and their interactions with Onramp’s lending platform.  


“Last year alone, our customers sold and processed over $2 billion in eCommerce orders, so our dataset is vast,” said Dr. John Coaster, Onramp’s Head of Credit and Analytics. “Training an algorithm on such a large data set results in a number of features that are useful in predicting positive and negative outcomes for each loan decision. This prediction is made in fractions of a second within the lending platform to return a score for the underwriter to evaluate a specific lending decision.”


Onramp has been able to improve risk detection via the machine-learning algorithm to the point where hurdles have been removed from the underwriting process.  For example, Onramp has no minimum requirement for time in business.  Some lenders require at least six to 12 months in business to be eligible to borrow. Onramp is able to approve more loans than others while maintaining target risk tolerance. Onramp’s approval rate through underwriting is nearly 50%, whereas other funding providers approve 10-30% of borrowers.


“By removing many of the traditional obstacles in the underwriting process, getting a loan from Onramp is much easier compared to traditional lines of credit and other financing products,” said Eric S. Youngstrom, Onramp’s CEO.  “Onramp does not generally pull a personal credit report, as our algorithm is better at assessing risk. By not using off-the-shelf software, we customize our lending capabilities via a more diverse range of data feeds that specifically addresses the needs of eCommerce businesses. Onramp’s entire lending platform was developed in-house, with data and risk as key priorities.”


Another unique feature of Onramp’s machine-learning algorithm is that its data set was collected during challenging economic times, so the company is much better at assessing risk in today’s economic climate.  Onramp re-trains the algorithm much more frequently and adapts to changing market conditions. This occurs because Onramp’s data is collected and aggregated within its core lending platform, as opposed to having to cleanse and integrate data from multiple sources.


Onramp’s next-generation algorithm, which will utilize true AI technology, is already in development, and is scheduled for launch later in 2024.  

About Onramp FundsOnramp is redefining the way eCommerce companies are funding their businesses. Based in Austin, TX, Onramp’s team is steeped in eCommerce, building tools for online businesses and helping sellers thrive with a unique repayment model that aligns with their sales. Onramp’s mission is to democratize finance and empower online sellers to scale their business, capture their demand, and meet their dreams.

Technology News