Performing interactive updates to a precalculated cross-channel predictive model

Abstract

A computer-implemented method, simulation and prediction system, and computer program product for advertising portfolio management are disclosed. An example method includes simulating, by executing a first simulated model utilizing a weight matrix, a first user provided scenario of an advertising campaign and collecting time series stimulus data and time series response data for marketing channels of the advertising campaign. The example method also includes reducing computational resource consumption associated with generating an updated version of the first simulated model by applying the time series stimulus data and the time series response data to the first simulated model to generate a second simulated model and simulating, by executing the second simulated model, a second user provided scenario of the advertising campaign to determine an effect of a change in a marketing media spend value of a first marketing channel on other ones of the marketing channels of the advertising campaign.

What is claimed is: 
     
         1 . An apparatus comprising:
 memory to store a weight matrix, the weight matrix including weights corresponding to cross-channel effects across marketing channels of an advertising campaign;   instructions to implement at least a first simulated model; and   processor circuitry to execute the instructions to at least:
 simulate, based on the first simulated model and the weight matrix, a first user provided scenario of the advertising campaign; 
 collect time series stimulus data and time series response data for the marketing channels of the advertising campaign; 
 reduce computational resource consumption associated with generating an updated version of the first simulated model by applying the time series stimulus data and the time series response data to the first simulated model to generate a second simulated model; and 
 simulate, based on the second simulated model, a second user provided scenario of the advertising campaign to determine an effect of a change in a marketing media spend value of a first marketing channel on other ones of the marketing channels of the advertising campaign. 
   
     
     
         2 . The apparatus of  claim 1 , wherein the effect is quantified as a percentage. 
     
     
         3 . The apparatus of  claim 1 , wherein the time series stimulus data includes at least one of an advertising spend, a number of direct mail pieces, a number of TV spots, a number of radio spots, a number of web impressions and a number of coupons printed, and the time series response data includes at least one of a number of calls into a call center, a number of clicks on an impression, or a number of coupon redemptions. 
     
     
         4 . The apparatus of  claim 1 , wherein the processor circuitry is to:
 compare respective ones of the weights to a threshold; and   designate ones of the weights that satisfy the threshold as true scores.   
     
     
         5 . The apparatus of  claim 1 , wherein the time series stimulus data is first time series stimulus data, the time series response data is first time series response data, the first simulated model includes second time series stimulus data and second time series response data, and the processor circuitry is to:
 update the second time series stimulus data with the first time series stimulus data; and   update the second time series response data with the first time series response data.   
     
     
         6 . The apparatus of  claim 5 , wherein:
 the second time series stimulus data and the second time series response data correspond to first known stimuli and first known responses collected at a first time; and   the first time series stimulus data and the first time series response data correspond to second known stimuli and second known responses collected at a second time, the first time earlier than the second time.   
     
     
         7 . The apparatus of  claim 1 , wherein the processor circuitry is to generate a report based on the effect, the report to quantify a return on investment for at least one of the marketing channels that accounts for the cross-channel effects across the marketing channels. 
     
     
         8 . A non-transitory computer readable medium comprising instructions that, when executed, cause processor circuitry to:
 simulate, based on a first simulated model and a weight matrix, a first user provided scenario of an advertising campaign, the weight matrix including weights corresponding to cross-channel effects across marketing channels of the advertising campaign;   collect time series stimulus data and time series response data for the marketing channels of the advertising campaign;   reduce computational resource consumption associated with generating an updated version of the first simulated model by applying the time series stimulus data and the time series response data to the first simulated model to generate a second simulated model; and   simulate, based on the second simulated model, a second user provided scenario of the advertising campaign to determine an effect of a change in a marketing media spend value of a first marketing channel on other ones of the marketing channels of the advertising campaign.   
     
     
         9 . The non-transitory computer readable medium of  claim 8 , wherein the effect is quantified as a percentage. 
     
     
         10 . The non-transitory computer readable medium of  claim 8 , wherein the time series stimulus data includes at least one of an advertising spend, a number of direct mail pieces, a number of TV spots, a number of radio spots, a number of web impressions and a number of coupons printed, and the time series response data includes at least one of a number of calls into a call center, a number of clicks on an impression, or a number of coupon redemptions. 
     
     
         11 . The non-transitory computer readable medium of  claim 8 , wherein the processor circuitry is to:
 compare respective ones of the weights to a threshold; and   designate ones of the weights that satisfy the threshold as true scores.   
     
     
         12 . The non-transitory computer readable medium of  claim 8 , wherein the time series stimulus data is first time series stimulus data, the time series response data is first time series response data, the first simulated model includes second time series stimulus data and second time series response data, and the processor circuitry is to:
 update the second time series stimulus data with the first time series stimulus data; and   update the second time series response data with the first time series response data.   
     
     
         13 . The non-transitory computer readable medium of  claim 12 , wherein:
 the second time series stimulus data and the second time series response data correspond to first known stimuli and first known responses collected at a first time; and   the first time series stimulus data and the first time series response data correspond to second known stimuli and second known responses collected at a second time, the first time earlier than the second time.   
     
     
         14 . The non-transitory computer readable medium of  claim 8 , wherein the processor circuitry is to generate a report based on the effect, the report to quantify a return on investment for at least one of the marketing channels that accounts for the cross-channel effects across the marketing channels. 
     
     
         15 . A method comprising:
 simulating, by executing a first simulated model utilizing a weight matrix, a first user provided scenario of an advertising campaign, the weight matrix including weights corresponding to cross-channel effects across marketing channels of the advertising campaign;   collecting time series stimulus data and time series response data for the marketing channels of the advertising campaign;   reducing computational resource consumption associated with generating an updated version of the first simulated model by applying the time series stimulus data and the time series response data to the first simulated model to generate a second simulated model; and   simulating, by executing the second simulated model, a second user provided scenario of the advertising campaign to determine an effect of a change in a marketing media spend value of a first marketing channel on other ones of the marketing channels of the advertising campaign.   
     
     
         16 . The method of  claim 15 , wherein the effect is quantified as a percentage. 
     
     
         17 . The method of  claim 15 , wherein the time series stimulus data includes at least one of an advertising spend, a number of direct mail pieces, a number of TV spots, a number of radio spots, a number of web impressions and a number of coupons printed, and the time series response data includes at least one of a number of calls into a call center, a number of clicks on an impression, or a number of coupon redemptions. 
     
     
         18 . The method of  claim 15 , further including removing at least one noisy weight from the weights by:
 comparing respective ones of the weights to a threshold; and   designating ones of the weights that satisfy the threshold as true scores.   
     
     
         19 . The method of  claim 15 , wherein the time series stimulus data is first time series stimulus data, the time series response data is first time series response data, the first simulated model includes second time series stimulus data and second time series response data, and applying the first time series stimulus data and the first time series response data to the first simulated model includes:
 updating the second time series stimulus data with the first time series stimulus data; and   updating the second time series response data with the first time series response data.   
     
     
         20 . The method of  claim 19 , wherein:
 the second time series stimulus data and the second time series response data correspond to first known stimuli and first known responses collected at a first time; and   the first time series stimulus data and the first time series response data correspond to second known stimuli and second known responses collected at a second time, the first time earlier than the second time.