Analyzing and converting unstructured networking system communications

Patent metadata

Field	Value
Publication number	US-10628851-B2
Application number	US-201615394482-A
Country	US
Kind code	B2
Filing date	Dec 29, 2016
Priority date	Dec 29, 2016
Publication date	Apr 21, 2020
Grant date	Apr 21, 2020

Abstract

The present disclosure is directed toward systems and methods for identifying offers in networking system post. For example, systems and methods described herein identify one or more offer indicators in a networking system post and calculate a confidence score representing a level of confidence that the unstructured networking system post includes a merchant offer. In response to calculating a confidence score above a threshold value, systems and methods described herein prompt the composer of the unstructured post to convert the post into a structured offer. Upon converting the unstructured post into a structured offer, systems and methods described herein intelligently distribute the structured offer for use by networking system users.

First claim

What is claimed is:

1. A method comprising:
detecting, by one or more computing devices of a networking system, a composition of an unstructured networking system post;
analyzing, by the one or more computing devices, the unstructured networking system post to generate a confidence score indicating a confidence that the unstructured networking system post includes an offer;
determining a dynamic threshold amount representing a likelihood that an author of the unstructured networking system post will convert the unstructured networking system post into a structured offer;
generating, based on how close the generated confidence score is to the dynamic threshold amount, a customized prompt to convert the unstructured networking system post into a structured offer by:
determining a distance between the generated confidence score and the dynamic threshold amount, and
generating the customized prompt such that a display size of the customized prompt is based on the determined distance; and

providing the customized prompt.

2. The method as recited in claim 1 , further comprising:
training a machine learning model for calculating confidence scores;
wherein analyzing the unstructured social networking post to generate the confidence score comprises utilizing the trained machine learning model to generate the confidence score for the unstructured networking system post.

3. The method as recited in claim 2 , wherein determining the dynamic threshold amount is based on a history of a page administrator or a merchant associated with the unstructured networking system post.

4. The method as recited in claim 3 , wherein determining the dynamic threshold amount based on the history of the page administrator or the merchant associated with the unstructured networking system post comprises:
determining whether the page administrator has declined a prompt to convert an unstructured networking system post into a structured offer post in the past;
based on a determination that the page administrator has declined a prompt to convert an unstructured networking system post into a structured offer post in the past, selecting a relatively higher dynamic threshold amount; and
based on a determination that the page administrator has not declined a prompt to convert an unstructured networking system post into a structured offer post in the past, selecting a relatively lower dynamic threshold amount.

5. The method as recited in claim 3 , wherein determining the dynamic threshold amount based on the history of the page administrator or the merchant associated with the unstructured networking system post comprises:
determining whether the page administrator has created a structured offer in the past;
based on a determination that the page administrator has created a structured offer in the past, selecting a relatively higher dynamic threshold amount; and
based on a determination that the page administrator has not created a structured offer in the past, selecting a relatively lower dynamic threshold amount.

6. The method as recited in claim 5 , wherein generating the customized prompt further comprises at least one of customizing contents of the prompt, customizing an appearance of the prompt, customizing a location of the prompt, or customizing a timing of the prompt.

7. The method as recited in claim 1 , further comprising:
receiving a request to convert the unstructured networking system post into a structured offer;
extracting information from the unstructured networking system post; and
pre-populating one or more fields of the structured offer with the extracted information.

8. The method as recited in claim 1 , further comprising:
receiving a request to convert the unstructured networking system post into a structured offer;
generating a structured offer based on the unstructured networking system post in response to the received request; and
replacing, within the networking system, the unstructured networking system post with the generated structured offer.

9. The method as recited in claim 8 , further comprising providing insights to a page administrator associated with the structured offer, the insights comprising information regarding at least one of a number of users that received the structured offer, a number of times the structured offer has been redeemed, or a conversion rate associated with the structured offer.

10. The method as recited in claim 1 , wherein detecting the composition of the unstructured social networking post comprises detecting the composition of the unstructured networking system post by a page administrator associated with a merchant's profile page on the networking system.

11. A system comprising:
at least one processor; and
at least one non-transitory computer-readable storage medium storing instructions thereon that, when executed by the at least one processor, cause the system to:
detect a composition of an unstructured networking system communication;
analyze the unstructured networking system communication to generate a confidence score indicating a confidence that the unstructured networking system communication includes an offer;
determine a dynamic threshold amount representing a likelihood that an author of the unstructured networking system communication will convert the unstructured networking system communication into a structured offer;
generate, based on how close the generated confidence score is to the dynamic threshold amount, a customized prompt to convert the unstructured networking system communication into a structured offer by:
determining a distance between the generated confidence score and the dynamic threshold amount, and
generating the customized prompt such that a display size of the customized prompt is based on the determined distance; and

provide the customized prompt.

12. The system as recited in claim 11 , further comprising instructions thereon that, when executed by the at least one processor, cause the system to:
train a machine learning model for calculating confidence scores based on a training data set comprising a plurality of unstructured networking system communications that include offers,
wherein analyzing the unstructured networking system communication to generate the confidence score comprises utilizing the trained machine learning model to generate the confidence score for the unstructured networking system communication.

13. The system as recited in claim 12 , wherein determining the dynamic threshold amount is based on a history of a page administrator or a merchant associated with the unstructured networking system communication.

14. The system as recited in claim 13 , wherein determining the dynamic threshold amount based on the history of the page administrator or the merchant associated with the unstructured networking system communication comprises:
determining whether the page administrator has created a structured offer in the past;
based on a determination that the page administrator has created a structured offer in the past, selecting a relatively higher dynamic threshold amount; and
based on a determination that the page administrator has not created a structured offer in the past, selecting a relatively lower dynamic threshold amount.

15. The system as recited in claim 14 , further comprising instructions thereon that, when executed by the at least one processor, cause the system to provide insights to a page administrator associated with the structured offer, the insights comprising information regarding at least one of a number of users that received the structured offer, a number of times the struct

Assignees

• Facebook Inc

Inventors

• Ketchpaw John Stephen
• Barak Dan
• Jackson Dean
• Geist Jr James F

Classifications

• G06N3/084

  Backpropagation, e.g. using gradient descent · CPC title

• G06Q30/0271

  Personalized advertisement · CPC title

• G06Q30/0277

  Online advertisement · CPC title

• G06Q30/0254 Primary

  based on statistics · CPC title

• G06N20/00

  Machine learning · CPC title