The ABC's of RPD (Return Path Data)
First there were diaries, then set-meters, and later people meters. In the television world, Nielsen Media was the main provider of this data measuring 210 individual markets as well as the whole country. They sampled the audience through diaries and/or meters. Then comScore, originally known as Rentrak, came along with another idea: Utilize the data stored in the cable and satellite boxes to provide viewing information. Rather than collect data from just hundreds or thousands of households, comScore would collect data from millions of households. The data in these boxes would be returned to the cable or satellite provider in what is known as RPD.
RPD means Return Path Data
Cable and satellite providers collect data from their subscribers set-top boxes. Just like computers and smart phones, the set-top boxes track and store their use. Return Path capable set-top boxes can store data of what was viewed and then return that information to the MVPD (Multi-channel Video Programming Distributor) provider. The MVPD may also provide the collected data to a third party such as Nielsen or comScore. It's important to understand that not all set-top boxes are return path capable. Older boxes may store limited information or even no information at all. Newer boxes are much more sophisticated.
Quite simply, set-top boxes record tuning to the box. Demographic (persons viewing) is not recorded. When it comes to set-top boxes, Nielsen and comScore can only determine household viewing.
RPD is a Passive Measurement System
Passive means that the viewer does not have to provide any information. He/she only must watch TV thru the set-top box in order to be measured. The resulting viewership can be considered stable data in the sense that the same households are measured over an extended period.
Since it began measuring local television in 2010, ComScore has used RPD data to provide viewing measurement. Nielsen began to utilize RPD in 2018 as a component to replace its Diary measurement. Nielsen and ComScore receive their RPD data from some of the same providers but there are some differences in their partnerships. The chart below shows the breakout of providers.
RPD Boxes are Not Sampling
Sampling would consist of a random number of viewers being measured in a market where that sample represents the market. Nielsen’s use of Local People Meters (LPM) in 25 of the top 26 markets utilize sampling. RPD’s, on the other hand, consist of a variety of set-top boxes that can differ by market depending on the Nielsen or comScore relationship with the provider. Each market is not supplied data equally. Some markets have more cable households supplying data while other markets may have less. There are also markets where there are no cable MVPDs supplying viewing information. Therefore, HH viewing data from these boxes are not a representative sample of the population.
Reliance on RPD Varies
Currently, Nielsen utilizes RPD information in 137 markets which represent 25.5 million TV households or 23.1% of the country’s TV HHs. Nielsen receives return path data from 4.4 million homes in these 137 markets. Each RPD set-top box represents an average of 5.8 homes. Nielsen plans to include RPD within LPM, set-meter, and code-reader markets sometime in the future.
ComScore utilizes RPD information in 210 markets which represent 110 million TV households or 100% of the country’s TV HHs. ComScore receives return path data from 30 million homes in these 210 markets. Each RPD set-top box represents an average of 3.7 homes.
Missing Pieces of the Puzzle
Return Path Data provides for a larger panel of homes than people meters, set meters and code readers. However, the sample can be skewed, because return path data excludes non-measured cable homes and over-the-air homes - a growing segment of viewers (cord-cutters/cord-shavers/cord-nevers) which can sometimes be a significant portion of the local market audience. In markets where Nielsen or comScore do not receive cable return path data, cable is completely excluded from the panel. Both Nielsen and comScore adjust the data in each market to compensate for missing cable and over-the-air households. Nielsen utilizes code reader meters to measure over-the-air homes in their RPD markets.
Collection and reporting of demographics differs greatly from the collection of household data. ComScore and Nielsen cannot determine viewing demographics from the set-top box alone. ComScore utilizes Experian to confirm the gender, age, and other characteristics of the occupants of the household and reports these viewer statistics as a HH "with the demographic". Simply put, they credit the household with a certain characteristic but cannot credit a specific person.
Nielsen, on the other hand, leverages its Viewer Assignment methodology to determine demographics in its RPD markets. With Viewer Assignment, Nielsen uses algorithms to find a “compatible” household that is being measured by a people meter to provide a sample demographic for the RPD household.
Some Households May Be Under-Represented
One major caveat in using set-top box data alone is that some types of households may be under-reported. In a recent Mediapost article, Nielsen shared that “set-top-box/return-path data coming from pay-tv providers and other sources can under-represent certain viewer groups, in particular Hispanic and African-American homes, compared to other household types.” Nielsen also reported that about 14% of RPD households do not have any demographic data assigned to them. Which leaves us scratching our head a bit about that "census" label everyone gives to RPD data.
Let's take a minute to summarize the advantages and disadvantages of RPD data:
If Return Path Data is going to continue to play a larger role in local television measurement, then Nielsen and comScore need to expand their relationships with additional MVPD partners such as Comcast, Mediacom, Suddenlink, etc. to fill in the gaps where return path data is missing. A more representative sample of all providers in the marketplace will also move the needle toward more accurate measurement of the market.
Quantity goes a long way in helping to establish data stability, but the most stable data can still not provide an accurate picture of audience behavior if it isn't representative. There is still a need for representative panels so that over-the-air viewers and those using MVPDs that do not supply return path data are also measured. Nielsen could accomplish this through the addition of more code-readers in the marketplace, especially in the top 125 markets that are not already set-meter or code reader markets. These panels would complement return path data to provide a more precise representation of viewing in the market.
ComScore may also need to consider panels or some additional type of measurement to compensate for over-the-air homes, OTT homes, and MVPDs that are not supplying data to fill in for their coverage gap since they do not fully sample the entire market.
Return path data has become the main source of viewing information for most to all television markets. While it is utilized to measure many homes passively and continuously, it is far from the perfect type of measurement. When it comes to acquiring big data, a balance between the amount of data and the accuracy of the data must always be considered.