System and method for evaluating the true reach of social media influencers

Abstract

A non-transitory computer readable storage media having computer-executable instructions, when executed by a processor, performs a method for evaluating a reach of a social media influencer. The methods provides for receiving a plurality of influencers at a server, wherein a data set is associated with each of the plurality of influencers; parsing the data set into quantitative data readable by a machine learning algorithm at the server; receiving, inputting, or both, a type of product or service at the server; classifying the type of product or service into at least one class of goods or services; training a node using the machine learning algorithm using the date set an input; and executing the machine learning algorithm to determine a score of each influencer for each class of goods or services. Systems for evaluating the reach of a social media influencer as it relates to advertisers and content is also disclosed herein.

We claim: 
     
         1 . A non-transitory computer readable storage media having computer-executable instructions, when executed by a processor, performs a method for evaluating a reach of a social media influencer, the instructions comprising:
 receiving a plurality of influencers at a server, wherein a data set is associated with each of the plurality of influencers;   parsing the data set into quantitative data readable by a machine learning algorithm at the server;   receiving a request from a third party at the server to place an influencer for its product or service, and to set up a campaign using the influencer, wherein the data from the campaign is stored on the server;   classifying the type of product or service into at least one class of goods or services;   training a node using the machine learning algorithm using the date set as an input;   executing the machine learning algorithm to determine a score of each influencer for each class of goods or services;   
     
     
         2 . The method of  claim 1 , wherein the machine learning algorithm comprises random forest, and the method further comprises:
 constructing a training data set from the data sets associated with each influencer, wherein the training data set comprises results of previous campaigns in at least one of the class of goods or services;   analyzing the training data set input into the server;   generating a forest of decision trees based on the training data set;   receiving additional influencers on to the server, wherein the additional influencers have respective data sets associated with them;   executing the machine learning algorithm to determine a score of each additional influencer as it relates to each class of goods or services;   prioritizing the influencers, additional influencers, or both, for each of the classes of goods or services that most are likely to have the highest influence in that class of goods or services.   
     
     
         3 . The method of  claim 2 , wherein training the node generating a forest of decision trees further comprises:
 inputting a historical data set for the influencers that are known to have successful campaigns in the class based upon known commercial success;   inputting the data set associated with the additional influencer;   generating vector spaces that correspond the input of the historical data set and the data set associated with the additional influencer;   running the data associated with the additional influencer through the forest of decision trees;   comparing the additional influencer data to the previous influencer data, and grouping the additional influencer data in the vector space that relates to a probabilistic level of success the influencer, additional influencer, or both may have with respect to the class of goods or service;   bagging the additional influencer data and continuously running further additional influencers through the node to increase efficacy;   wherein the score for each influencer ranges from 0.0 to 1.0 for each class of goods or services.   
     
     
         4 . The method of  claim 2 , wherein generating a forest of decision trees comprises generating an ensemble of a plurality of randomly trained decision trees, wherein each of the decision trees are split into subsets based on an attribute value test, and each of the additional influencers are given the score based on the value test. 
     
     
         5 . The method of  claim 1 , wherein classifying the goods or services into at least one of a plurality of classes comprises grouping businesses together with common-type products, services, or both, and generating sectors of classes. 
     
     
         6 . The method of  claim 1 , wherein the data set comprises:
 a first quantitative data set that directly corresponds to distribution potential, wherein distribution potential comprises a number of followers and number of times the influencer is mentioned in other posts;   a second date set that corresponds directly to engagement and interaction potential, wherein the engagement and interaction potential comprises likes, comments, social media shares, forwarding, reviews, check-ins, contributors and active contributors, clicks on page views, unique visitors from social media, sessions from social networking sites time spent thereon, and average response times;   a third data set that corresponds directly to influence potential, wherein influence potential comprises shares of the influencer's conversation, historical data based on businesses for which the influencer has acted as an influencer for or has otherwise advocated for, the satisfaction of any of its followers from such other businesses it has acted as an influencer for or has otherwise advocated for;   a fourth data set that corresponds directly to creative information, wherein creative information comprises content posted, social causes, and subjective brand relation;   a fifth data set that corresponds directly to influencer personal information;   wherein the step of scoring the influencer comprises using the quantitative data sets.   
     
     
         7 . The method of  claim 6 , further comprising comparing each of the quantitative data sets with a predetermined range of data gathered from the machine learning algorithm to assess the probability of success of a campaign with the influencer or additional influencer in the class of goods or services. 
     
     
         8 . The method of  claim 1 , further comprising receiving, inputting, or both, a type of product or service at the server; 
     
     
         9 . The method of  claim 1 , wherein the at least one node is a plurality of nodes that are self-tuning. 
     
     
         10 . The method of  claim 8 , wherein once the campaign is complete, the campaign data is run through the machine learning algorithm to increase efficacy. 
     
     
         11 . A system for evaluating a reach of a social media influencer, the system having non-transitory computer-readable media comprising a program of machine executable instructions for a programmable computer system that, when executed by the programmable computer system, will cause the programmable computer system to execute instruction, the system comprising:
 a media processing module in communication with an influencer data base, the media processing module configured to receive a plurality of influencers at a server, wherein a data set is associated with each influencer;   a social media footprint estimator configured to parse the data set into quantitative data readable by a machine learning algorithm at the server;   an influencer prioritization module configured to receive, input, or both, a type of product or service at the server, and to further classify the product or service into at least one of a plurality of classes;   an influencer real footprint module comprising a machine learning module configured to train a node using the data set to identify the influencer's influence associated with in selling the type of product or service input and to determine a score of each influencer for each product input and further configured to receive a request from a third party at the server to place an influencer for its product or service, and to set up a campaign using the influencer, wherein the data from the campaign is stored on the server;   
     
     
         12 . The system of  claim 11 , wherein the machine learning algorithm comprises random forest, and the machine learning module is further configured to:
 construct a training data set from the data sets associated with each influencer;   analyze the training data set input into the server;   generate a forest of decision trees based on the training data set;   receive additional influencers on to the server, wherein the additional influencer has a data set associated with then, the additional influencers being the influencers to be scored, ranked, or both;   prioritize the influencers from each of the classes of products or services that most likely to have the highest influence in that class.   
     
     
         13 . The system of  claim 12 , wherein the machine learning module is further configured to:
 input historical influencer data for influencers that are known to have successful campaigns in the class based upon commercial success;   input the data set associated with the additional influencer;   generate vector spaces that correspond the inputs;   run the data associated with the additional influencer through the random forest;   compare the additional influencer data to the previous influencer data, and grouping the additional influencer data in the vector space that relates to a probabilistic level of success the influencer may have with respect to the class of goods or service;   bag the additional influencer data that and continuously running additional influencers through the node to increase efficacy;   generate the score ranging from 0.0 to 1.0 for the additional influencers for each class.   
     
     
         14 . The system of  claim 12 , wherein generating a forest of decision trees comprises generating an ensemble of a plurality of randomly trained decision trees, wherein each of the decision trees are split into subsets based on an attribute value test, and each of the additional influencers are given the score. 
     
     
         15 . The system of  claim 11 , wherein classifying the product or service into at least one of a plurality of classes comprises grouping businesses together with common-type products, services, or both, and generating sectors of classes of business. 
     
     
         16 . The system of  claim 11 , wherein the data set comprises:
 a first quantitative data set that directly corresponds to distribution potential, distribution potential comprising number of followers, and number of times the influencer is mentioned in other posts;   a second date set that corresponds directly to engagement and interaction potential, the engagement and interaction potential comprising likes, comments, social media shares, forwarding, reviews, check-ins, contributors and active contributors, clicks on page views, unique visitors from social media, sessions from social networking sites time spent thereon, and average response times;   a third data set that corresponds directly to influence potential, influence potential comprising shares of the influencer's conversation, historical data based on businesses for which the influencer has acted as an influencer for or has otherwise advocated for, the satisfaction of any of its followers from such other businesses it has acted as an influencer for or has otherwise advocated for;   a fourth data set that corresponds directly to creative information, creative information comprising of content posted, social causes, and subjective brand relation;   a fifth data set that corresponds directly influencer persona information;   wherein the step of scoring the influencer comprises using the quantitative data sets.   
     
     
         17 . The system of  claim 16 , wherein the influencer real footprint module is further configured to receive a request from a third party to place a influencer for its product or service, and to set up a campaign using the influencer, wherein the data from the campaign is stored on the server. 
     
     
         18 . The system of  claim 11 , wherein the at least one node is a plurality of nodes that are self-tuning. 
     
     
         19 . The system of  claim 17 , wherein once the campaign is complete, the campaign data is run through the machine learning algorithm. 
     
     
         20 . A computer implemented influencer evaluation system, the system comprising:
 a server configured to receive a request for an influencer based on a product or service;   a random forest tree generator that receives input data associated with an influencer and the product or service and generates a forest of decision trees based on the data set for the class of goods;   an influencer real footprint module configured to determine a score of each influencer for the product or service.