METHODS, SYSTEMS, APPARATUS AND ARTICLES OF MANUFACTURE TO MODEL eCOMMERCE SALES

Abstract

Methods, apparatus, systems and articles of manufacture methods, systems, apparatus and articles of manufacture to model ecommerce sales are disclosed. A system to model to eCommerce sales includes a trend identifier to compute commerce metric differences corresponding to products, the commerce metric differences based on first commerce metrics scraped at a first time and second commerce metrics scraped at a second time, a splitter to split the commerce metric differences into a first portion of the commerce metric differences corresponding to a first dataset of eCommerce cooperators, and into a second portion of the commerce metric differences corresponding to a second dataset of eCommerce non-cooperators, a machine learning engine to infer sales data by estimating eCommerce non-cooperators sales based on the second portion of the commerce metric differences, and a sales allocator to estimate sales missing from collected sales data based on the estimate eCommerce non-cooperators sales.

1 . (canceled) 
     
     
         2 . A system to model eCommerce sales comprising:
 sales modeler circuitry to: 
 split panel sales data into first panel sales data corresponding to eCommerce cooperators and second panel sales data corresponding to eCommerce non-cooperators; 
 determine an expansion factor for a first one of a plurality of strata; and 
   bias reducer circuitry to: 
 determine a first set of weights based on a set of consumer profiles and a set of panel profiles, the first set of weights corresponding to the first one of the plurality of strata, ones of the first set of weights to correspond to respective ones of attributes; and 
 determine a total unbiased eCommerce non-cooperator sales for a first product and the first one of the plurality of strata based on first unbiased sales values corresponding to the first product and the first one of the plurality of strata, ones of the first unbiased sales values to correspond to respective ones of the attributes, wherein the ones of the first unbiased sales values are based on (a) the expansion factor for the first one of the plurality of strata, (b) respective ones of the first set of weights, and (c) corresponding portions of the second panel sales data. 
   
     
     
         3 . The system of  claim 2 , wherein the corresponding portions of the second panel sales data correspond to the first product, the first one of the plurality of strata, and respective ones of the attributes. 
     
     
         4 . The system of  claim 2 , wherein the first one of the plurality of strata includes one of city, region, socio-economic class, age, and income. 
     
     
         5 . The system of  claim 2 , wherein, prior to determining the expansion factor, the sales modeler circuitry is to:
 sort the first panel sales data, the second panel sales data, and eCommerce cooperators sales data based on at least one of (a) product category or (b) product; and   arrange each of the sorted first panel sales data, the sorted second panel sales data, and the sorted eCommerce cooperators sales data into the plurality of strata.   
     
     
         6 . The system of  claim 5 , wherein the sales modeler circuitry is to determine the expansion factor for the first one of the plurality of strata based on a ratio of a portion of the eCommerce cooperators sales data and a portion of the first panel sales data, the portions corresponding to the first one of the plurality of strata. 
     
     
         7 . The system of  claim 5 , wherein the bias reducer circuity is to arrange each of (a) the set of consumer profiles and (b) the set of panel profiles into the plurality of strata and the attributes. 
     
     
         8 . The system of  claim 7 , wherein the set of consumer profiles represents unbiased consumer profiles and the set of panel profiles represents biased consumer profiles, the ones of the first set of weights to represent a difference between the unbiased consumer profiles for a respective attribute and the biased consumer profiles for the respective attribute. 
     
     
         9 . The system of  claim 5 , wherein the bias reducer circuity is to determine total unbiased eCommerce sales for the first product and the first one of the plurality of strata across the eCommerce cooperators and the eCommerce non-cooperators based on (a) the total unbiased eCommerce non-cooperators sales for the first product and the first one of the plurality of strata and (b) a portion of the eCommerce sales data corresponding to the first product and the first one of the plurality of strata. 
     
     
         10 . The system of  claim 9 , wherein the sales modeler circuitry is to determine a plurality of expansion factors, ones of the plurality of expansion factors to correspond to respective ones of the plurality of strata; and 
 wherein the bias reducer circuity is to:
 determine a plurality of sets of weights, ones of the plurality of sets of weights to correspond to respective ones of the plurality of strata; 
 determine a plurality of total unbiased eCommerce non-cooperator sales for the first product, ones of the plurality of total unbiased eCommerce non-cooperator sales to correspond to respective ones of the plurality of strata; and 
 determine total unbiased eCommerce sales for the first product across (a) the plurality of strata, (b) the eCommerce cooperators and (c) the eCommerce non-cooperators. 
   
     
     
         11 . The system of  claim 10 , wherein the bias reducer circuitry is to determine total eCommerce sales across products by determining total eCommerce sales for each of the products, the products including the first product. 
     
     
         12 . At least one non-transitory computer readable storage medium comprising instructions that, when executed, cause processor circuitry to at least:
 split panel sales data into first panel sales data and second panel sales data, the first panel sales data corresponding to eCommerce cooperators, the second panel sales data corresponding to eCommerce non-cooperators;   determine an expansion factor for a stratum;   determine a group of weights based on a set of consumer profiles and a set of panel profiles, the group of weights to correspond to the stratum, ones of the group of weights to correspond to respective ones of attributes;   determine a group of unbiased sales values corresponding to eCommerce non-cooperators, the group of unbiased sales values to correspond to a product and the stratum, ones of the group of unbiased sales values determined based on (a) the expansion factor for the stratum, (b) respective ones of the group of weights, and (c) corresponding portions of the second panel sales data; and   determine total unbiased eCommerce non-cooperator sales for the product and the stratum, the total unbiased eCommerce non-cooperators sales for the product and the stratum based on the group of unbiased sales values.   
     
     
         13 . The at least one non-transitory computer readable storage medium of  claim 12 , wherein the corresponding portions of the second panel sales data correspond to the product, the stratum, and respective ones of the attributes. 
     
     
         14 . The at least one non-transitory computer readable storage medium of  claim 12 , wherein the stratum is one of a set of strata, and wherein, prior to determining the expansion factor, the instructions, when executed, cause the processor circuitry to:
 sort the first panel sales data, the second panel sales data, and eCommerce cooperators sales data based on at least one of (a) product category or (b) by product; and   arrange each of the sorted first panel sales data, the sorted second panel sales data, and the sorted eCommerce cooperators sales data into the set of strata.   
     
     
         15 . The at least one non-transitory computer readable storage medium of  claim 14 , wherein the instructions, when executed, cause the processor circuitry to determine the expansion factor for the stratum based on a ratio of a portion of the eCommerce cooperators sales data and a portion of the first panel sales data, the portions corresponding to the stratum. 
     
     
         16 . The at least one non-transitory computer readable storage medium of  claim 14 , wherein the instructions, when executed, cause the processor circuitry to arrange each of (a) the set of consumer profiles and (b) the set of panel profiles into the set of strata and the attributes. 
     
     
         17 . The at least one non-transitory computer readable storage medium of  claim 16 , wherein the set of consumer profiles represents unbiased consumer profiles and the set of panel profiles represents biased consumer profiles, the ones of the group of weights to represent a difference between the unbiased consumer profiles for a respective attribute and the biased consumer profiles for the respective attribute. 
     
     
         18 . The at least one non-transitory computer readable storage medium of  claim 14 , wherein the instructions, when executed, cause the processor circuitry to determine total unbiased eCommerce sales for the product and the stratum based on the total unbiased eCommerce non-cooperators sales for the product and the stratum and a portion of the eCommerce sales data corresponding to the product and the stratum. 
     
     
         19 . The at least one non-transitory computer readable storage medium of  claim 18 , wherein the instructions, when executed, cause the processor circuitry to:
 determine a plurality of expansion factors corresponding to the set of strata, ones of the expansion factors to be based on respective ones of the set of strata;   determine groups of weights, ones of the groups of weights to correspond to respective ones of the set of strata;   determine groups of unbiased sales values, ones of the groups of unbiased sales values to correspond to the product and respective ones of the set of strata, the ones of the groups of unbiased sales values determined based on (a) respective ones of the expansion factors, (b) respective weights of respective ones of the groups of weights, and (c) corresponding portions of the second panel sales data; and   determine a plurality of total unbiased eCommerce non-cooperator sales for the product, ones of the plurality of total unbiased eCommerce non-cooperator sales to correspond to respective ones of the set of strata.   
     
     
         20 . The at least one non-transitory computer readable storage medium of  claim 19 , wherein the instructions, when executed, cause the processor circuitry to determine total unbiased eCommerce sales for the product across the set of strata, across the eCommerce cooperators, and across the eCommerce non-cooperators, wherein the total unbiased eCommerce sales for the product is based on (a) the ones of the plurality of total unbiased eCommerce non-cooperator sales and (b) respective portions of the eCommerce sales data. 
     
     
         21 . A method comprising:
 splitting, by executing machine readable instructions with at least one processor, panel sales data into first panel sales data corresponding to eCommerce cooperators and second panel sales data corresponding to eCommerce non-cooperators;   determining, by executing the machine readable instructions with the at least one processor, expansion factors for strata of interest, ones of the expansion factors corresponding to respective ones of the strata;   determining, by executing the machine readable instructions with the at least one processor, weights for the ones of the strata based on a set of consumer profiles and a set of panel profiles, wherein first weights of the weights correspond to a first stratum, and wherein ones of the first weights correspond to respective ones of attributes;   determining, by executing the machine readable instructions with the at least one processor, unbiased sales values for a product of interest, the unbiased sales values corresponding to eCommerce non-cooperators, ones of the unbiased sales values to correspond to respective ones of the strata, wherein a first one of the unbiased sales values for the first stratum is based on (a) a respective one of the expansion factors, (b) the first weights, and (c) corresponding portions of the second panel sales data;   determining, by executing the machine readable instructions with the at least one processor, eCommerce sales values for the product of interest, ones of the eCommerce sales values to correspond to respective ones of the strata, wherein a first one of the eCommerce sales values corresponding to the first stratum is based on the unbiased sales values and a respective portion of eCommerce cooperators sales data; and   determine a total unbiased eCommerce sales for the product of interest across the strata of interest based on the eCommerce sales values for the product of interest.   
     
     
         22 . The method of  claim 21 , further including determining total eCommerce sales for a plurality of products, the total eCommerce sales for the plurality of products based on a respective plurality of total eCommerce sales, ones of the plurality of total eCommerce sales corresponding to respective ones of the plurality of products.