# US 10,872,377 B2 · Dynamic origination of capital pricing based on historical point-of-sale data

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 10,872,377 B2
- Document ID: US10872377B2
- Status: Granted
- Kind: B2
- Application number: 16/406,550
- Filed: 2019-05-08
- Published: 2020-11-12
- Granted: 2020-12-22
- Earliest claimed: 2019-05-08
- Assignees: Toast, Inc.
- Applicants: Toast, Inc.
- Inventors: Benjamin Gordon; Alexander H. Hails; Clement Masson
- Web page: https://patentgrep.com/patent/US10872377B2

## CPC

- G06Q 40/03: Credit; Loans; Processing thereof
- G06Q 10/04: Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" (market predictions or forecasting for commercial activities G06Q30/0202)
- G06Q 20/20: Point-of-sale [POS] network systems
- G06Q 20/201: {Price look-up processing, e.g. updating}
- G06Q 20/202: {Interconnection or interaction of plural electronic cash registers [ECR] or to host computer, e.g. network details, transfer of information from host to ECR or from ECR to ECR}
- G06Q 30/0202: Market predictions or forecasting for commercial activities
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

A system for dynamic origination of capital is provided. The system includes a rate processor, a revenue forecaster, and an offer processor. The rate processor is configured to perform a binary logistic regression on subscriber data and metrics derived from historical POS data corresponding to subscribers of a point-of-sale (POS) subscription service to generate logistic regression coefficients that are used to calculate probability of default (PD) values for each of the subscribers, and configured to employ new subscriber data and metrics to generate updated PD values for the each of the subscribers. The revenue forecaster is coupled to the rate processor and is configured to employ the historical POS data to predict future POS data for establishments corresponding to the each of the subscribers, and to employ the future POS data to generate predicted total revenues corresponding to the each of the subscribers over a payback period, and configured to employ new POS data to generate updated predicted total revenues corresponding to the each of the subscribers. The offer processor is configured to generate and transmit capital product offers corresponding to the each of the subscribers, where the capital product offers comprise a payback period, prices, and maximum dollar amounts that are a percentage of the predicted total revenues, and configured to compare the updated predicted total revenues with the predicted total revenues and the updated PD values with the PD values, and configured to dynamically adjust the prices and the maximum dollar amounts according to the updated PDs and the updated predicted total revenues.

## Claims

### Claim 1

A system for dynamic origination of capital, the system comprising:
a logistic regression engine, configured to perform a binary logistic regression on subscriber data and metrics derived from historical POS data corresponding to subscribers of a point-of-sale (POS) subscription service to generate logistic regression coefficients that are used to calculate probability of default (PD) values for each of said subscribers, and configured to employ new subscriber data and metrics to generate updated PD values for said each of said subscribers;
a revenue forecaster, coupled to said rate processor, configured to employ said historical POS data to predict future POS data for establishments corresponding to said each of said subscribers, and to employ said future POS data to generate predicted total revenues corresponding to said each of said subscribers over a payback period, and configured to employ new POS data to generate updated predicted total revenues corresponding to said each of said subscribers; and
an offer processor, configured to generate and transmit capital product offers corresponding to said each of said subscribers, wherein said capital product offers comprise a payback period, prices, and maximum dollar amounts that are a percentage of said predicted total revenues, and configured to compare said updated predicted total revenues with said predicted total revenues and said updated PD values with said PD values, and configured to dynamically adjust said prices and said maximum dollar amounts according to said updated PDs and said updated predicted total revenues.

### Claim 2 (depends on 1)

The system as recited in claim 1, wherein said logistic regression engine, said revenue forecaster, and said offer processor are disposed within a backend server that is coupled via the internet to a plurality of fixed and mobile terminals in establishments corresponding to said each of said subscribers, and wherein said subscribers employ said plurality of fixed and mobile terminals in conjunction with said backend server to initiate and process orders within said establishments, and to take payments for said orders, wherein said payments comprise processing of credit card transactions.

### Claim 3 (depends on 2)

The system as recited in claim 2, wherein said backend server takes said payments for accepted capital product offers by automatically deducting holdback amounts from processing of said credit card sales transactions.

### Claim 4 (depends on 3)

The system as recited in claim 3, wherein said holdback amounts are deducted daily.

### Claim 5 (depends on 1)

The system as recited in claim 1, wherein said establishments comprise restaurants, said capital product offers comprise merchant cash advance offers, and said prices comprise interest rates.

### Claim 6 (depends on 1)

The system as recited in claim 1, wherein said offer processor withdraws said capital product offers for those establishments whose PD has increased such that calculation of their corresponding prices rises above an upper rate limit.

### Claim 7 (depends on 1)

The system as recited in claim 1, wherein said revenue forecaster comprises a recurrent neural network that is trained on said historical POS data corresponding to said each of said subscribers, and that processes a subset of said historical POS data to predict said future POS data, wherein a length of said subset corresponds to said payback period.

### Claim 8

A system for dynamic origination of capital, the system comprising:
a logistic regression engine, configured to perform a binary logistic regression on subscriber data and metrics derived from historical POS data corresponding to subscribers of a restaurant point-of-sale (POS) subscription service to generate logistic regression coefficients that are used to calculate probability of default (PD) values for each of said subscribers, and configured to employ new subscriber data and metrics to generate updated PD values for said each of said subscribers;
a revenue forecaster, coupled to said rate processor, configured to employ said historical POS data to predict future POS data for restaurants corresponding to said each of said subscribers, and to employ said future POS data to generate predicted total revenues corresponding to said each of said subscribers over a payback period, and configured to employ new POS data to generate updated predicted total revenues corresponding to said each of said subscribers; and
an offer processor, configured to generate and transmit capital product offers corresponding to said each of said subscribers, wherein said capital product offers comprise a payback period, prices, and maximum dollar amounts that are a percentage of said predicted total revenues, and configured to compare said updated predicted total revenues with said predicted total revenues and said updated PD values with said PD values, and configured to dynamically adjust prices and said maximum dollar amounts according to said updated PDs and said updated predicted total revenues.

### Claim 9 (depends on 8)

The system as recited in claim 8, wherein said logistic regression engine, said revenue forecaster, and said offer processor are disposed within a backend server that is coupled via the internet to a plurality of fixed and mobile terminals in restaurants corresponding to said each of said subscribers, and wherein said subscribers employ said plurality of fixed and mobile terminals in conjunction with said backend server to initiate and process orders within said restaurants, and to take payments for said orders, wherein said payments comprise processing of credit card transactions.

### Claim 10 (depends on 9)

The system as recited in claim 9, wherein said backend server takes said payments for accepted capital product offers by automatically deducting holdback amounts from processing of said credit card sales transactions.

### Claim 11 (depends on 10)

The system as recited in claim 10, wherein said holdback amounts are deducted daily.

### Claim 12 (depends on 8)

The system as recited in claim 8, wherein said offer processor withdraws said capital product offers for those restaurants whose PD has increased such that calculation of their corresponding prices rises above an upper rate limit.

### Claim 13 (depends on 12)

The system as recited in claim 12, wherein said revenue forecaster comprises a recurrent neural network that is trained on said historical POS data corresponding to said each of said subscribers, and that processes a subset of said historical POS data to predict said future POS data, wherein a length of said subset corresponds to said payback period.

### Claim 14

A method for dynamic origination of capital, the method comprising:
performing a binary logistic regression on subscriber data and metrics derived from historical POS data corresponding to subscribers of a point-of-sale (POS) subscription service to generate logistic regression coefficients that are used to calculate probability of default (PD) values for each of the subscribers, and employing new subscriber data and metrics to generate updated PD values for the each of the subscribers;
employing the historical POS data to predict future POS data for establishments corresponding to the each of the subscribers, and using the future POS data to generate predicted total revenues corresponding to the each of the subscribers over a payback period, and employing new POS data to generate updated predicted total revenues corresponding to the each of the subscribers; and
generating and transmitting capital product offers corresponding to the each of the subscribers, wherein the MCA offers comprise the payback period, prices, and maximum dollar amounts that are a percentage of the predicted total revenues, and comparing the updated predicted total revenues with the predicted total revenues and the updated PD values with the PD values, and dynamically adjusting the prices and the maximum dollar amounts according to the updated PDs and the updated predicted total revenues.

### Claim 15 (depends on 14)

The method as recited in claim 14, wherein the method is performed by one or more special-purpose central processing units disposed within a backend server that is coupled via the internet to a plurality of fixed and mobile terminals in establishments corresponding to the each of the subscribers, and wherein the subscribers employ the plurality of fixed and mobile terminals in conjunction with the backend server to initiate and process orders within the establishments, and to take payments for the orders, wherein the payments comprise processing of credit card transactions.

### Claim 16 (depends on 15)

The method as recited in claim 15, wherein the backend server takes said payments for accepted capital product offers by automatically deducting holdback amounts from processing of the credit card sales transactions.

### Claim 17 (depends on 16)

The method as recited in claim 16, wherein the holdback amounts are deducted daily.

### Claim 18 (depends on 14)

The method as recited in claim 14, wherein the establishments comprise restaurants, the capital product offers comprise merchant cash advance offers, and the prices comprise interest rates.

### Claim 19 (depends on 18)

The method as recited in claim 18, further comprising:
withdrawing the capital product offers for those establishments whose PD has increased such that calculation of their corresponding offer interest rates rises above an upper rate limit.

### Claim 20 (depends on 14)

The method as recited in claim 14, wherein said employing the historical POS data to predict future POS data corresponding to the each of the subscribers comprises utilizing a recurrent neural network that is trained on the historical POS data corresponding to the each of the subscribers, and that processes a subset of the historical POS data to predict the future POS data, wherein a length of the subset corresponds to the payback period.