Suggesting an item for gifting to a user of an online concierge system

Patent metadata

Field	Value
Publication number	US-2024112238-A1
Application number	US-202217956217-A
Country	US
Kind code	A1
Filing date	Sep 29, 2022
Priority date	Sep 29, 2022
Publication date	Apr 4, 2024
Grant date	—

Abstract

An online concierge system receives a request to purchase a gift for a user of the system and retrieves a profile associated with the user. Based on the profile and attributes of items included among inventories of one or more retailer locations, the system identifies a set of candidate items for which the user is likely to have an affinity. The system accesses a machine learning model trained to predict a giftability score for an item and applies the model to attributes of each candidate item to predict its giftability score. Based on its giftability score and the profile, the system computes a composite score for each candidate item indicating an appropriateness of gifting the candidate item to the user. The system ranks the set of candidate items based on the composite scores and selects one or more suggested items for gifting to the user based on the ranking.

First claim

What is claimed is:

1 . A method performed at a computer system comprising a processor and a computer-readable medium, the method comprising:
receiving a request to purchase a gift for a user of an online concierge system; retrieving a profile associated with the user, wherein the profile comprises a set of attributes of the user, an order history associated with the user, or a list of items for which the user has expressed a preference; identifying a set of candidate items included among a plurality of items for which the user is likely to have an affinity based at least in part on the profile associated with the user and a plurality of attributes associated with the plurality of items, wherein the plurality of items is included among one or more inventories of one or more retailer locations; accessing a machine learning model that is trained to predict a giftability score for an item, wherein the machine learning model is trained by:
receiving the plurality of attributes associated with the plurality of items,
receiving, for each item of the plurality of items, a label indicating an appropriateness of gifting a corresponding item, and
training the machine learning model based at least in part on the plurality of attributes and the label for each of the plurality of items;
applying the machine learning model to the plurality of attributes associated with each candidate item of the set of candidate items to predict the giftability score for each candidate item; computing a composite score associated with each candidate item of the set of candidate items indicating the appropriateness of gifting a corresponding candidate item to the user, wherein the composite score is based at least in part on the giftability score for the corresponding candidate item and the profile associated with the user; ranking the set of candidate items based at least in part on the composite score associated with each candidate item of the set of candidate items; selecting, from the set of candidate items, one or more suggested items for gifting to the user based at least in part on the ranking; and responsive to receiving the request to purchase a gift for the user, sending a message including a suggestion identifying the selected one or more suggested items.

2 . The method of claim 1 , wherein identifying the set of candidate items included among the plurality of items for which the user is likely to have the affinity comprises:
generating an embedding associated with the user based at least in part on the profile associated with the user; generating a plurality of embeddings associated with the plurality of items based at least in part on the plurality of attributes associated with the plurality of items; and identifying the set of candidate items included among the plurality of items for which the user is likely to have the affinity based at least in part on a proximity between the embedding associated with the user and each embedding of a set of embeddings included among the plurality of embeddings, wherein the set of embeddings is associated with the set of candidate items.

3 . The method of claim 2 , wherein the composite score associated with each candidate item of the set of candidate items is computed as a product of the giftability score for the corresponding candidate item and a dot product of the embedding associated with the user and an additional embedding associated with the corresponding candidate item.

4 . The method of claim 1 , wherein the set of attributes of the user comprises one or more of: a set of preferences of the user, a set of responses to a questionnaire describing an affinity of the user for one or more items included among the plurality of items, and a set of actions associated with one or more items included among the plurality of items performed by the user on a social networking system.

5 . The method of claim 1 , wherein the label indicating the appropriateness of gifting the corresponding item is based at least in part on one or more of: an indication that the corresponding item included in an order received from a user of the online concierge system is a gift, a description from a manufacturer indicating the appropriateness of gifting the corresponding item, a response to suggesting the corresponding item as a gift for a user of the online concierge system, an item category associated with the corresponding item, a packaging for the corresponding item, and a price of the corresponding item.

6 . The method of claim 1 , further comprising:
receiving a set of bid values associated with a subset of the set of candidate items, wherein the set of bid values is received from a retailer associated with a retailer location; and boosting the composite scores associated with the subset of the set of candidate items based at least in part on the set of bid values.

7 . The method of claim 6 , wherein ranking the set of candidate items is further based at least in part on the boosted composite scores.

8 . The method of claim 1 , wherein the order history associated with the user is associated with a set of retailers and the suggested one or more candidate items comprises a gift card for a retailer not included among the set of retailers.

9 . The method of claim 1 , wherein identifying the set of candidate items included among the plurality of items for which the user is likely to have the affinity is further based at least in part on one or more of: a catalog of items and a hierarchical taxonomy associated with the plurality of items, wherein the hierarchical taxonomy is based at least in part on one or more item categories associated with each item of the plurality of items.

10 . A computer program product comprising a non-transitory computer readable storage medium having instructions encoded thereon that, when executed by a processor, cause the processor to:
receive a request to purchase a gift for a user of an online concierge system; retrieve a profile associated with the user, wherein the profile comprises a set of attributes of the user, an order history associated with the user, or a list of items for which the user has expressed a preference; identify a set of candidate items included among a plurality of items for which the user is likely to have an affinity based at least in part on the profile associated with the user and a plurality of attributes associated with the plurality of items, wherein the plurality of items is included among one or more inventories of one or more retailer locations; access a machine learning model that is trained to predict a giftability score for an item, wherein the machine learning model is trained by:
receiving the plurality of attributes associated with the plurality of items,
receiving, for each item of the plurality of items, a label indicating an appropriateness of gifting a corresponding item, and
training the machine learning model based at least in part on the plurality of attributes and the label for each of the plurality of items;
apply the machine learning model to the plurality of attributes associated with each candidate item of the set of candidate items to predict the giftability score for each candidate item; compute a composite score associated with each candidate item of the set of candidate items indicating the appropriateness of gifting a corresponding candidate item to the user, wherein the composite score is based at least in part on the giftability score for the corresponding candidate item and the profile associated with the user; rank the set of candidate items based at least in part on the composite score associated with each candidate item of the set of candidate items; select, from the set of candidate i

Assignees

• Maplebear Inc Dba Instacart

Inventors

• Narlikar Girija
• Ahuja Karuna
• Goel Radhika
• Ahuja Chakshu
• Zhang Xiaoming
• Das Devlina

Classifications

• G06Q10/40

  Business processes related to social networking or social networking services · CPC title

• G06Q10/42

  Determination of affinities or common interests between users · CPC title

• G06Q30/0631 Primary

  Recommending goods or services · CPC title

• G06Q30/0203

  Market surveys; Market polls · CPC title

• G06Q30/0603 Primary

  Catalogue creation or management · CPC title