CONDUCTING MULTIVARIATE EXPERIMENTS
Systems and methods for organizing and controlling the display of content, then measuring the effectiveness of that content in modifying behavior, within a particular temporal and special dimension, so as to minimize or eliminate confounding effects.
1 . A computer-implemented method for conducting a multivariate experiment using a plurality of content delivery channels, comprising:
receiving, for the plurality of content delivery channels, temporal factor data and spatial factor data;
for a particular content delivery channel of the plurality of content delivery channels, computing a plurality of experimental units, the experimental units comprising a response area based on the spatial factor data for the particular content delivery channel, and a response duration based on the temporal factor data for the particular content delivery channel;
arranging the plurality of experimental units into hierarchies based on overlap of the response areas of each experimental unit; and,
computing a content playlist for each content delivery channel based on the experimental units, wherein the content playlist at least specifies content to be played on each content delivery channel.
2 . The computer-implemented method of claim 1 , wherein computing the content playlist additionally comprises the order of the content to be presented and/or the relative or absolute frequency of play within the experimental unit.
3 . The computer-implemented method of claim 2 , further comprising: causing the content delivery channels to play content consistent with the content playlist.
4 . The computer-implemented method of claim 1 , wherein the response area is derived by doubling the spatial factor for channel for content delivery.
5 . The computer-implemented method of claim 1 , wherein the response duration is derived by doubling the temporal factor.
6 . The computer-implemented method of claim 1 , wherein the randomization constraint is derived using functions comprising an explore utility function and an exploit utility function.
7 . The computer-implemented method of claim 1 , further comprising collecting data indicative of behavior in the experimental units.
8 - 14 . (canceled)
15 . The computer-implemented method of claim 1 , wherein the channel for content delivery is selected from a group of channels of content delivery, comprises fixed-location digital displays, digital billboards, mobile devices, and web pages.
16 . (canceled)
17 . The computer-implemented method of claim 1 , wherein the content playlist is determined based on varying the combination of content randomly, with experimental controls.
18 . A computer-based system for conducting multivariate experiments on content effectiveness, comprising
a plurality of content delivery channels, each comprising one or more displays;
a processor configured to execute program instructions that:
receive special factor data and temporal factor data, and based thereon, compute a plurality of experimental units associated with one content delivery channel of the plurality of content delivery channels, comprising a response area that is based on the spatial factor data, and a response duration based on the temporal factor data;
arranging the plurality of experimental units into a hierarchy based on the geographic overlap of the response areas; and,
from a content database having information defining a plurality of different content pieces, assigning content pieces associated with some of the experimental units to a playlist, utilizing a randomization constraint.
19 . The computer-based system of claim 18 , further comprising:
causing the content pieces to be displayed on content delivery channels in a manner consistent with the playlist.
20 . The computer-based system of claim 18 , further comprising:
a plurality of data collection devices situated within the response areas of the experimental units.
21 . The computer-based system of claim 18 , wherein the response area is derived by doubling the spatial factor for channel for content delivery.
22 . The computer-based system of claim 18 , wherein the response duration is derived by doubling the temporal factor for the channel for content delivery.
23 . (canceled)
24 . The computer-based system of claim 18 , further comprising a content effectiveness database comprising data regarding known content effectiveness of various of content pieces in the content database.
25 . The computer-based system of claim 18 , further comprising: receiving response data from response areas.
26 . The computer-based system of claim 25 , further comprising: parsing the response data based on the applicable experimental condition.
27 . The computer-based system of claim 26 , wherein the experimental condition comprises the content displayed during the experimental unit, the location of the experimental unit, and the time that the experimental unit started and ended.
28 . (canceled)
29 . The system of claim 18 , wherein the channel for content delivery comprises fixed-location digital displays, digital billboards, mobile devices, or web pages.
30 . The system of claim 18 , further comprising a plurality of data collection devices, and wherein the data collection devices comprise intermediate metric sensors associated with each channel for content communication and overall metric sensors associated with a cross-channel metric.
31 .- 32 . (canceled)