Methods and apparatus to correct attribution errors and coverage bias for digital audio ratings
View Patent ↗Methods, apparatus, systems and articles of manufacture to correct attribution errors and coverage bias for digital audio ratings are disclosed. An example method includes applying, by executing an instruction with a processor, a duration-based count matrix to first demographic data associated with collected impressions of digital audio to determine second demographic data, the duration-based count matrix being based on a misattribution adjustment matrix, applying, by executing an instruction with the processor, a coverage adjustment vector to the second demographic data to determine third demographic data, applying, by executing an instruction with the processor, a scaling factor to the third demographic data to determine fourth demographic data, and generating, by executing an instruction with the processor, ratings data at a daypart-level using the fourth demographic data, the ratings data associated with the digital audio.
1. A method to generate digital audio ratings, the method comprising:
providing media monitoring instructions to instantiate with digital audio, the media monitoring instructions to cause the client devices to send messages to a server of the database proprietor in response to the client devices accessing the digital audio;
collecting, by executing an instruction with a processor, impressions of the digital audio accessed by the client devices;
obtaining, by executing an instruction with the processor, first demographic data from the server of the database proprietor based on the collected impressions, the first demographic data including misattribution errors generated by the server of the database proprietor based on respective cookies included in the messages sent by the client devices to the server of the database proprietor, at least one of the cookies to inaccurately identify a user of a corresponding one of the client devices;
applying, by executing an instruction with the processor, a duration-based count matrix to the first demographic data associated with the collected impressions of digital audio to determine second demographic data, the duration-based count matrix based on a misattribution adjustment matrix, the misattribution adjustment matrix including a first probability that a person is attributable to a first demographic group when the first demographic data associates the person with a second demographic group different from the first demographic group;
converting, by executing an instruction with the processor, the first probability associated with the first demographic group and the second demographic group to a duration-based probability included in the duration-based count matrix, the duration-based probability corresponding to a duration unit count when (1) a streaming station distributing the digital audio is a digital counterpart of an over-the-air station, and (2) the streaming station and the over-the-air station have a same ad-load;
applying, by executing an instruction with the processor, a coverage adjustment vector to the second demographic data to determine third demographic data;
applying, by executing an instruction with the processor, a scaling factor to the third demographic data to determine fourth demographic data; and
generating, by executing an instruction with the processor, ratings data at a daypart-level using the fourth demographic data to correct the misattribution errors generated by the server of the database proprietor, the ratings data associated with the digital audio.
2. The method of claim 1 , further including using survey calibration data to determine the probability.
3. The method of claim 1 , wherein the duration-based probability corresponds to a quarter-hour unit count when (3) a streaming station distributing the digital audio is not associated with a counterpart over-the-air station or (4) the over-the-station and the streaming station do not have a same ad-load.
4. The method of claim 1 , wherein the first demographic data includes at least one of a duration unit count, a session starts count, a quarter-hour unit count and a unique audience count.
5. An apparatus to generate digital audio ratings, the apparatus comprising:
an audience measurement entity impressions collector to:
provide media monitoring instructions to instantiate with digital audio, the media monitoring instructions to cause the client devices to send messages to a server of the database proprietor in response to the client devices accessing the digital audio; and
collect impressions of the digital audio accessed by the client devices;
a database proprietor data manager to obtain first demographic data from the server of the database proprietor based on the collected impressions, the first demographic data including misattribution errors generated by the server of the database proprietor based on respective cookies included in the messages sent by the client devices to the server of the database proprietor, at least one of the cookies to inaccurately identify a user of a corresponding one of the client devices;
a misattribution adjuster to apply a duration-based count matrix to the first demographic data associated with the collected impressions of digital audio to determine second demographic data, the duration-based count matrix based on a misattribution adjustment matrix, the misattribution adjustment matrix including a probability that a person is attributable to a first demographic group when the first demographic data associates the person with a second demographic group different from the first demographic group;
a matrix converter to convert the probability associated with the first demographic group and the second demographic group to a duration-based probability included in the duration-based count matrix, the duration-based probability corresponding to a duration unit count when (1) a streaming station distributing the digital audio is a digital counterpart of an over-the-air station, and (2) the streaming station and the over-the-air station have a same ad-load;
a coverage adjuster to apply a coverage adjustment vector to the second demographic data to determine third demographic data;
a data scaler to apply a scaling factor to the third demographic data to determine fourth demographic data; and
a ratings data generator to generate ratings data at a daypart-level using the fourth demographic data to correct the misattribution errors generated by the server of the database proprietor, the ratings data associated with the digital audio.
6. The apparatus of claim 5 , further including a matrix generator to use survey calibration data to determine the probability.
7. The apparatus of claim 5 , wherein the duration-based probability corresponds to a quarter-hour unit count when (3) a streaming station distributing the digital audio is not associated with a counterpart over-the-air station or (4) the over-the-station and the streaming station do not have a same ad-load.
8. The apparatus of claim 5 , wherein the first demographic data includes at least one of a duration unit count, a session starts count, a quarter-hour unit count and a unique audience count.
9. A tangible machine-readable storage medium comprising instructions that, when executed, cause a processor to at least:
provide media monitoring instructions to instantiate with digital audio, the media monitoring instructions to cause the client devices to send messages to a server of the database proprietor in response to the client devices accessing the digital audio;
collect impressions of the digital audio accessed by the client devices;
obtain first demographic data from the server of the database proprietor based on the collected impressions, the first demographic data including misattribution errors generated by the server of the database proprietor based on respective cookies included in the messages sent by the client devices to the server of the database proprietor, at least one of the cookies to inaccurately identify a user of a corresponding one of the client devices;
apply a duration-based count matrix to the first demographic data associated with the collected impressions of digital audio to determine second demographic data, the duration-based count matrix based on a misattribution adjustment matrix, the misattribution adjustment matrix including a probability that a person is attributable to a first demographic group when the first demographic data associates the person with a second demographic group different from the first demographic group;
convert the probability associated with the first demographic group and the second demographic group to a duration-based probability included in the duration-based count matrix, the duration-based probability corresponding to a duration unit count when (1) a streaming station distributing the digital audio is a digital counterpart of an over-the-air station, and (2) the streaming station and the over-the-air station have a same ad-load;
apply a coverage adjustment vector to the second demographic data to determine third demographic data;
apply a scaling factor to the third demographic data to determine fourth demographic data; and
generate ratings data at a daypart-level using the fourth demographic data to correct the misattribution errors from the server of the database proprietor, the ratings data associated with the digital audio.
10. The tangible machine-readable storage medium as defined in claim 9 , wherein the instructions are to further cause the processor to utilize survey calibration data to determine the probability.
11. The tangible machine-readable storage medium as defined in claim 9 , wherein the duration-based probability corresponds to a quarter-hour unit count when (3) a streaming station distributing the digital audio is not associated with a counterpart over-the-air station or (4) the over-the-station and the streaming station do not have a same ad-load.