Dynamically adjusting digital component campaign values for disengaged consumers

1 . A method, comprising:

receiving, by one or more processors from a merchant, digital component campaign information comprising a base conversion value, one or more additional conversion values, a base bid value, one or more additional bid values, and ground truth data indicating one or more disengaged consumers;

receiving, by the one or more processors from a publisher, a request for a digital component;

generating, by executing a first stage model of an artificial intelligence (AI) model trained offline using the ground truth data, one or more consumer embedding representations;

generating, by executing a second stage model of the AI model trained offline using the ground truth data, one or more merchant embedding representations, wherein at least one of the first or second stage model is trained offline;

determining, based on the request for the digital component and the one or more consumer and merchant embedding representations, by the one or more processors executin the AI model online:

(i) a disengaged consumer probability representing a probability a consumer to be presented the digital component is a disengaged consumer, and

(ii) a qualifying disengaged consumer probability representing a probability the consumer to be presented the digital component is a qualifying disengaged consumer; and

dynamically adjusting, by the one or more processors based on the one or more additional bid values or the one or more additional conversion values, and at least one of (i) the disengaged consumer probability or (ii) the qualifying disengaged consumer probability, at least one of a resulting bid value or a resulting conversion value associated with the digital component.

2 . The method of claim 1 , wherein the one or more disengaged consumers are previously existing consumers that meet one or more disengagement criteria.

3 . The method of claim 2 , wherein the one or more disengagement criteria comprise a set of criteria, and the previously existing consumer is a disengaged consumer when the previously existing consumer fulfills a threshold number of disengagement criteria of the set of criteria.

4 . The method of claim 1 , wherein:

the qualifying disengaged consumer is a disengaged high value consumer,

a disengaged consumer is a previously existing consumer that meets one or more disengagement criteria defined by the merchant,

a high value consumer is a consumer that meets one or more value qualifications defined by the merchant, and

the disengaged high value consumer is the consumer that meets the one or more value qualifications and the one or more disengagement criteria.

5 . The method of claim 1 , wherein determining the disengaged consumer probability and the qualifying disengaged consumer probability occurs at a time of bidding to select the digital component in response to the received request.

6 . The method of claim 1 , wherein at least one of the one or more additional conversion values corresponds to a disengaged consumer conversion value or a qualifying disengaged consumer conversion value, and when dynamically adjusting the resulting conversion value the method further comprises multiplying the disengaged consumer conversion value or the qualifying disengaged consumer conversion value with the respective disengaged consumer probability or qualifying disengaged consumer probability.

7 . The method of claim 1 , wherein at least one of the one or more additional bid values corresponds to a disengaged consumer bid value or a qualifying disengaged consumer bid value, and the method further comprises determining a total bid value based on the base bid value, the disengaged consumer bid value, the qualifying disengaged consumer bid value, the disengaged consumer probability, and the qualifying disengaged consumer probability.

8 . The method of claim 1 , further comprising determining, by the one or more processors based on the disengaged consumer probability or the qualifying disengaged consumer probability, whether a conversion is at least one of a disengaged consumer conversion or a qualifying disengaged consumer conversion.

9 . The method of claim 8 , wherein determining whether the conversion is the at least one of the disengaged consumer conversion or the qualifying disengaged consumer conversion comprises comparing, by the one or more processors, a randomized number to the qualifying disengaged consumer probability,

wherein when the randomized number is less than the qualifying disengaged consumer probability, the conversion corresponds to the qualifying disengaged consumer conversion, and

wherein when the randomized number is greater than the qualifying disengaged consumer probability, the conversion corresponds to another type of conversion, wherein the other types of conversions include at least one of a disengaged consumer conversion, an existing consumer conversion, or an unknown consumer conversion.

10 . A system, comprising:

one or more processors, wherein the one or more processors are configured to:

receive, from a merchant, digital component campaign information comprising a base conversion value, one or more additional conversion values, a base bid value, one or more additional bid values, and ground truth data indicating one or more disengaged consumers;

receive, from a publisher, a request for a digital component;

generate, by executing a first stage model of an artificial intelligence (AI) model trained offline using the ground truth data, one or more consumer embedding representations;

generate, by executing a second stage model of the AI model trained offline using the ground truth data, one or more merchant embedding representations, wherein at least one of the first or second stage model is trained offline;

determine, based on the request for the digital component and the one or more consumer and merchant embedding representations, by executing the AI model online:

(i) a disengaged consumer probability representing a probability a consumer to be presented the digital component is a disengaged consumer, and

(ii) a qualifying disengaged consumer probability representing a probability the consumer to be presented the digital component is a qualifying disengaged consumer; and

dynamically adjust, based on the one or more additional bid values or the one or more additional conversion values, and at least one of (i) the disengaged consumer probability or (ii) the qualifying disengaged consumer probability, at least one of a resulting bid value or a resulting conversion value associated with the digital component.

11 . The system of claim 10 , wherein the one or more disengaged consumers are previously existing consumers that meet one or more disengagement criteria.

12 . The system of claim 11 , wherein the one or more disengagement criteria comprise a set of criteria, and the previously existing consumer is a disengaged consumer when a previously existing consumer fulfills a threshold number of disengagement criteria of the set of criteria.

13 . The system of claim 10 , wherein:

the qualifying disengaged consumer is a disengaged high value consumer,

a disengaged consumer is a previously existing consumer that meets one or more disengagement criteria defined by the merchant,

a high value consumer is a consumer that meets one or more value qualifications defined by the merchant, and

the disengaged high value consumer is the consumer that meets the one or more value qualifications and the one or more disengagement criteria.

14 . The system of claim 10 , wherein determining the probability that the consumer being presented with the digital component is the at least one of a disengaged consumer or a qualifying disengaged consumer occurs at a time of bidding to select the digital component in response to the received request.

15 . The system of claim 13 , wherein at least one of the one or more additional conversion values corresponds to a disengaged consumer conversion value or a qualifying disengaged consumer conversion value, and when dynamically adjusting the resulting conversion value the one or more processors are further configured to multiplying the disengaged consumer conversion value or the qualifying disengaged consumer conversion value with the respective disengaged consumer probability or qualifying disengaged consumer probability.

16 . A non-transitory computer-readable medium storing instructions, which when executed by one or more processors, cause the one or more processors to:

receive, from a merchant, digital component campaign information comprising a base conversion value, one or more additional conversion values, a base bid value, one or more additional bid values, and ground truth data indicating one or more disengaged consumers;

receive, from a publisher, a request for a digital component;

generate, by executing a first stage model of an artificial intelligence (AI) model trained offline using the ground truth data, one or more consumer embedding representations;

generate, by executing a second stage model of the AI model trained offline using the ground truth data, one or more merchant embedding representations, wherein at least one of the first or second stage model is trained offline;

determine, based on the request for the digital component and the one or more consumer and merchant embedding representations, by executing the AI model trained online:

(i) a disengaged consumer probability representing a probability a consumer to be presented the digital component is a disengaged consumer, and

(ii) a qualifying disengaged consumer probability representing a probability the consumer to be presented the digital component is a qualifying disengaged consumer; and

dynamically adjust, based on the one or more additional bid values or the one or more additional conversion values and at least one of (i) the disengaged consumer probability or (ii) the qualifying disengaged consumer probability, at least one of a resulting bid value or a resulting conversion value associated with the digital component.

17 . The method of claim 1 , wherein:

the first stage model is trained, based on at least one of consumer data or digital component data for one or more merchants, to identify one or more characteristics of a consumer that is likely to complete a conversion for a given digital component, and

the second stage model is trained to determine:

(i) the disengaged consumer probability, and

(ii) the qualifying disengaged consumer probability.

18 . The method of claim 1 , wherein:

the first stage model is configured to map raw consumer data to a consumer representation,

the second stage model is configured to determine, based on the consumer representation:

(i) the disengaged consumer probability, and

(ii) the qualifying disengaged consumer probability, and

the consumer representation comprises the one or more consumer embedding representations.

19 . The method of claim 1 , further comprising:

associating, by the one or more processors, labels of the ground truth data with the digital component campaign information; and

training, by the one or more processors, the first stage model and the second stage model with at least some of the labels associated with the ground truth data and the digital component campaign information.

20 . The method of claim 1 , wherein:

the one or more consumer embedding representations correspond to an embedded set of numbers to represent one or more consumers; and

the one or more merchant embedding representations correspond to one or more embedded set of numbers to represent respective merchants.

21 . The method of claim 1 , further comprising:

receiving, by the one or more processors, feedback related to the determinations of the AI model; and

updating, by the one or more processors, based on the feedback, the AI model.