Thinkbox’s annual look at the shifting sands and solid ground in the UK video world is now complete. It certainly didn’t get any less complicated in 2017 – especially as we now have a new methodology from comScore which we can build into the analysis.
So, in the absence of a single source of measurement for all video viewing, and with Netflix and Amazon not divulging viewing figures, let’s have a look at what the best available data we have in the UK tells us about our video viewing here in the UK. The methodology is at the end.
TV accounts for 95% of video advertising seen
Important to put this one first, given this should be what matters most to advertisers. And TV ads are seen in full, with the sound (most likely) turned on and with unbeatable effectiveness.
For 16–34s, TV advertising accounts for 90%. In 2017, YouTube accounted for 0.9% of video advertising time for all individuals and 2.9% for 16–34s. Netflix and Amazon accounted for none.
TV accounts for 71% of all video viewing
A caveat: we can confidently say ‘viewing’ when we talk about TV, cinema and SVOD, for example, but not all video activity is strictly speaking viewing (or by humans). For example, a significant chunk of YouTube activity is where the service is being used more like a jukebox.
Nonetheless, in total, the average person in the UK watched 4 hours, 39 minutes a day of video in 2017. TV’s share of total video time in 2017 was slightly down from 2016, a total of 4% points fewer. The astonishing choice of video on offer has unsurprisingly taken a slice of time from watching TV – most likely the TV we watched more as a compromise than a choice.
Live TV viewing continues to be the most popular form of video
But it has moved from 60% of all video to 56%. The shares of the playback of recorded TV and Broadcaster VOD are largely unchanged year on year.
New Comscore data alters the picture
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).
So, following comScore’s change in methodology, YouTube activity has grown from 6.5% of all video in 2016 to 9.1% in 2017.
88% of YouTube use is by 20% of its users
It's important to take into account the fact that, unlike BARB, comScore reports the activity of devices not humans. comScore’s data shows that 88% of YouTube activity is generated by only 20% of devices. There is a hardcore of YouTube super-users that affect the average (for TV 48% of viewing is accounted for by the heaviest 20%).
In addition, a portion of YouTube activity is using the service as a jukebox, so the service is being listened to as opposed to watched. This use of YouTube is on the increase. IPA Touchpoints data for 2017 shows that 71% of 16-34s use YouTube at least once a month for ‘background music’, up from 57% in 2016.
Facebook now accounts for 1.2% of video time
This equates to 3 minutes a day per person, which echoes Facebook’s own claims that 100 million hours of video are watched on Facebook a day. With 2 billion users globally, this averages out at 3 minutes a day per person. In 2016, Facebook accounted for 1.7% of video time, so no growth year on year.
Amongst 16-34s, TV accounts for 49% of all video viewing
This is down from 59% in 2016 following the impact of comScore’s new methodology and the effect it has had on YouTube activity, which increased from 13% of all 16-34s’ video in 2016 to 22% in 2017
SVOD viewing has grown from 4.1% to 6.4%
It is likely that this growth has been at broadcaster TV viewing’s expense, which makes sense as SVOD is TV and more homes now have access to SVOD services. As of Q4 2017, 10.2 million homes in the UK (43%) subscribed to either Amazon Prime or Netflix, according to the BARB establishment survey, up 23% from 8.3 million in Q4 2016. Based on our viewing estimates, this means that in homes with access to Netflix or Amazon Prime, the average viewer spends 50 minutes watching SVOD per day.
For 16-34s, SVOD has grown from 9% to 13% of all video viewing. 54% of 16-34s have access to either Amazon Prime or Netflix according to the BARB establishment survey. In these homes we estimate that 16-34s spend on average 1 hour, 3 minutes watching SVOD per day.
Life stage influences video consumption
No one has a crystal ball, but we do have the IPA’s Touchpoints study. It shows the impact that our life stage has on our viewing habits. Looking within the ‘millennials’ generation, it shows that as millennials get older they watch less YouTube and Netflix, and more TV. If they have kids, TV becomes even more popular. With 91% of the UK having internet access, 73% having VOD services on their TV and 81% carrying smartphones, it is reasonable to think that this may now be an established pattern of behaviour.
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.