The quantitative analysis of total video consumption in the UK combined 2017 data from BARB, comScore, the IPA’s Touchpoints and Rentrak box office data.
BARB data shows how much time is spent viewing broadcaster content, live or time-shifted (DTR and VOD) on the TV set. It also records how much time the TV set is being used for other activities such as SVOD, DVDs and gaming. However, BARB is unable to determine what the splits are between each of these different activities.
To determine these, we used Touchpoints 2017 to calibrate the BARB data. This survey provides estimates for the time spent split by the different TV set activities and provides estimates for how much time is spent viewing this content on other devices such as tablets, smartphones, and laptops.
The combination of these sources creates a solid estimate as it utilises the robustness of BARB data (12,000 panel members, representative of the UK TV population, metered actual consumption data, analysis across a whole year) alongside the detailed splits of viewing activity provided by Touchpoints diary data.
In 2017, for the first time, there was a full year of data available from comScore’s multiplatform Video Metrix analysis software. This has had an impact on the YouTube data. In the absence of mobile data, previous Thinkbox analyses of the video world have employed IPA Touchpoints data to estimate the proportion of YouTube activity on mobiles. The new comScore data – provided by YouTube – suggests that mobile accounts for a significantly higher proportion than the Touchpoints study found (39% for Touchpoints vs. 66% comScore). As comScore multi-platform data doesn’t include viewing on TV sets from connected TVs or games consoles, we add in the volume believed to be missing based on Touchpoints data.
As comScore multi-platform data is not available across all sites, the missing portion of mobile viewing for Facebook, other online and Adult XXX is still calculated from Touchpoints data.
The analysis also used census level broadcaster stream data to estimate the time spent watching Broadcaster VOD on both the TV set and other devices. OFCOM’s Digital Day study was used to account for the viewing to BVOD generated by under 15s, whilst Touchpoints is used to estimate the relative size for 16-34s.
Time spent viewing video at cinemas was based on box office sales from Rentrak. 16-34 estimates were generated from the Touchpoints profile data of Cinema goers.
All data is weighted to the entire population based on ONS UK population estimates.
Video advertising viewing:
YouTube stopped supplying data to comScore on the number of ads served as of January 2017. As a result, we have had to model the amount of time spent viewing advertising on YouTube based on the past relationship between time spent using YouTube and time spent viewing ads on YouTube using 2016 data.
This data suggests that 10 minutes of YouTube activity on average converts to 5 seconds of time spent with advertising. In contrast, live TV converts 64 seconds of time spent viewing ads for every 10 minutes of time spent watching shows.
There are several reasons that impact YouTube’s ability to convert activity into viewing of ads:
- YouTube doesn’t put advertising within content that hasn’t reached a certain volume of views and the vast majority of activity on YouTube is with the long tail.
- Many ads on YouTube are skippable.
- 90% of activity on YouTube is generated by 20% of users, frequency capping will restrict ads from playing out excessively to the same users.
- YouTube generally only runs one per-roll per video.
Other online video advertising includes everything else other than YouTube, including Facebook and all auto-play advertising across all publishers as reported by comScore. This is also up-weighted to estimate mobile consumption using Touchpoints 2017 data.
Cinema data is based on Rentrak box office sales and an estimate of 10 minutes ad viewing per film. Broadcaster VOD data is based on impression delivery provided by all broadcasters from their player data alongside average ad impression view-through rates to estimate total time. Playback and Live TV is based on BARB data.