Impression-trend techniques for providing a display advertising supply forecast

Abstract

Techniques are described herein for using an impression-trend technique to provide a display advertising supply forecast. A display advertising supply forecast is an estimate of a number of impressions, which are to occur in a future time period, that have specified attribute values. For example, the specified attribute values may be descriptive of impressions with respect to which an advertiser wishes to place ads. An impression-trend technique is a forecasting technique that uses trends regarding past impressions to forecast a number of future impressions that have specified attribute values. The past impressions include attribute values that are related to the specified attribute values.

What is claimed is: 
     
         1 . A method of providing a display advertising supply forecast, comprising:
 receiving a query to estimate a number of first impressions, which are to occur in a future time period, that have a combination of specified attribute values, the combination including a first subset of the specified attribute values and a second subset of the specified attribute values;   combining an interim trend and a number of second impressions that are linked to the interim trend to provide a revised trend, the second impressions occurring in a past time period and having the first subset of the specified attribute values and the second subset of the specified attribute values, the interim trend being based on a plurality of aggregates of impressions that have a third combination of attribute values, the third combination of attribute values being related to the second subset of the specified attribute values, the plurality of aggregates corresponding to a plurality of respective time periods; and   estimating the number of the first impressions that have the combination of specified attribute values based on the revised trend to provide the display advertising supply forecast.   
     
     
         2 . The method of  claim 1 , further comprising:
 combining a second interim trend and a number of the second impressions that are linked to the second interim trend to provide a second revised trend, the second interim trend being based on a second plurality of aggregates of impressions that have a fourth combination of attribute values, the fourth combination of attribute values being related to the second subset of the specified attribute values, the second plurality of aggregates corresponding to the plurality of respective time periods;   wherein estimating the number of the first impressions that have the combination of specified attribute values comprises:   estimating the number of the first impressions that have the combination of specified attribute values based on the first revised trend and the second revised trend to provide the display advertising supply forecast.   
     
     
         3 . The method of  claim 2 , wherein estimating the number of the first impressions that have the combination of specified attribute values comprises:
 estimating the number of the first impressions that have the combination of specified attribute values based on the first revised trend to provide a first intermediate forecast;   estimating the number of the first impressions that have the combination of specified attribute values based on the second revised trend to provide a second intermediate forecast;   combining the first intermediate forecast and the second intermediate forecast   to provide the display advertising supply forecast.   
     
     
         4 . The method of  claim 1 , further comprising:
 determining the number of the second impressions that are linked to the first interim trend based on a random sample of a plurality of impressions that includes the second impressions; and   determining the number of the second impressions that are linked to the second interim trend based on a random sample of the plurality of impressions that includes the second impressions.   
     
     
         5 . The method of  claim 1 , wherein a duration of the past time period is approximately one week. 
     
     
         6 . The method of  claim 1 , wherein combining the interim trend and the number of the second impressions that are linked to the interim trend comprises:
 combining the interim trend and a proportion of the second impressions that are linked to the interim trend to provide the revised trend.   
     
     
         7 . The method of  claim 1 , wherein combining the interim trend and the number of the second impressions that are linked to the interim trend comprises:
 combining a periodically normalized interim trend and the number of the second impressions that are linked to the periodically normalized interim trend to provide the revised trend, the periodically normalized interim trend being based on a plurality of normalized aggregates of impressions that have the third combination of attribute values, the plurality of normalized aggregates corresponding to the plurality of respective time periods.   
     
     
         8 . A system to provide a display advertising supply forecast, comprising:
 a query module configured to receive a query to estimate a number of first impressions which are to occur in a future time period, that have a combination of specified attribute values, the combination including a first subset of the specified attribute values and a second subset of the specified attribute values;   a combination module configured to combine an interim trend and a number of second impressions that are linked to the interim trend to provide a revised trend, the second impressions occurring in a past time period and having the first subset of the specified attribute values and the second subset of the specified attribute values, the interim trend being based on a plurality of aggregates of impressions that have a third combination of attribute values, the third combination of attribute values being related to the second subset of the specified attribute values, the plurality of aggregates corresponds to a plurality of respective time periods; and   a forecast module configured to estimate the number of the first impressions that have the combination of specified attribute values based on the revised trend to provide the display advertising supply forecast.   
     
     
         9 . The system of  claim 8  wherein the combination module is further configured to combine a second interim trend and a number of the second impressions that are linked to the second interim trend to provide a second revised trend, the second interim trend being based on a second plurality of aggregates of impressions that have a fourth combination of attribute values, the fourth combination of attribute values being related to the second subset of the specified attribute values, the second plurality of aggregates corresponding to the plurality of respective time periods and wherein the forecast module is further configured to estimate the number of the first impressions that have the combination of specified attribute values based on the first revised trend and the second revised trend to provide the display advertising supply forecast. 
     
     
         10 . The system of  claim 9  wherein the forecast module is further configured to estimate the number of the first impressions that have the combination of specified attribute values based on the first revised trend and the second revised trend to provide the display advertising supply forecast by estimating the number of the first impressions that have the combination of specified attribute values based on the first revised trend to provide a first intermediate forecast, estimating the number of the first impressions that have the combination of specified attribute values based on the second revised trend to provide a second intermediate forecast; and combining the first intermediate forecast and the second intermediate forecast to provide the display advertising supply forecast. 
     
     
         11 . The system of  claim 8  further comprising a link module configured to determine the number of the second impressions that are linked to the first interim trend based on a random sample of a plurality of impressions that includes the second impressions and to determine the number of the second impressions that are linked to the second interim trend based on a random sample of the plurality of impressions that includes the second impressions. 
     
     
         12 . The system of  claim 8 , wherein a duration of the past time period is one week. 
     
     
         13 . The system of  claim 8  wherein the combination module is configured to combine the interim trend and the number of the second impressions that are linked to the interim trend by combining the interim trend and a proportion of the second impressions that are lined to the interim trend to provide the revised trend. 
     
     
         14 . The system of  claim 8  wherein the combination module is configured to combine the interim trend and the number of the second impressions that are linked to the interim trend by combining a periodically normalized interim trend and the number of the second impressions that are linked to the periodically normalized interim trend to provide the revised trend, the periodically normalized interim trend being based on a plurality of normalized aggregates of impressions that have the third combination of attribute values, the plurality of normalized aggregates corresponding to the plurality of respective time periods. 
     
     
         15 . A system for providing a display ad supply forecast, the system comprising:
 a query module configured to receive a query for estimation of a number of first impressions, which are to occur in a future time period, that have a combination of attribute values specified in the query including a first subset of the specified attribute values and a second subset of the specified attribute values, and to provide the attribute values specified in the query to a trend module;   a trend module configured to receive link indicators and the attribute values specified in the query and determine second impressions that have occurred in a past time period and that have the first subset of the specified attribute and the second subset of the specified attribute values and that are linked to an interim trend based on the link indicators, and output an indication of the second impressions that are linked among the interim trend;   a link module configured to receive the indication of the second impressions that are linked among the interim trend, determine a number of the second impressions that are linked among the interim trend, and output the number of the second impressions that are linked among the interim trend and the interim trend;   a combination module configured receive the number of the second impression that are linked to the interim trend and the interim trend, combine the interim trend and the number of the second impressions linked to the interim trend to provide a revised trend, and output the revised trend; and   a forecast module configured to receive the revised trend, estimate the number of first impressions that have the combination of specified attribute values based on the revised trend, and output a display advertising supply forecast based on the estimated number of first impression.   
     
     
         16 . The system of  claim 15  wherein the wherein the link module is further configured to determine a plurality of interim trends based on the link indicators and to reiteratively operate on each of the interim trends to output a plurality of indications of the second impressions that are linked among the plurality of the interim trends. 
     
     
         17 . The system of  claim 16  wherein the combination module is further configured to operate reiteratively on each of the indications of the second impressions and output a plurality of revised trends. 
     
     
         18 . They system of  claim 17  wherein the forecasting module is further configured to estimate the number of first impressions based on the plurality of revised trends. 
     
     
         19 . The system of  claim 16  wherein each of the plurality of interim trends is a periodically normalized interim trend. 
     
     
         20 . The system of  claim 16  further comprising a counter module configured to output an identification of the interim trend to be operated on to the link module.