The SVOD Report 2020 12 May 2020 The 2010s saw a seismic change in how British viewers could watch television, with an explosion in the number of platforms, devices and catch-up services available. The only constant was that the British love affair with the TV set remained undiminished; the total daily time spent with a TV set in 2019 was 242 minutes, the same as in 2010. However, TV set use changed significantly over the decade. One change is the ability to access subscription video on-demand (SVOD) services, and with the BARB Establishment Survey tracking the take-up of SVOD subscriptions in UK homes (chart 1), the trend is clear. In Q4 2019, the proportion of homes that had a subscription to at least one of Netflix, Amazon Prime Video or Now TV was 50.5%, a majority of homes for the first time. Overall, 14.3 million UK households had one of these SVOD services, an increase of 2.0 million homes (16%), compared with Q4 2018. Netflix and Amazon saw similar levels of growth in the number of households that subscribe to their services. Amazon’s increase represents a larger year-on-year percentage growth of 35% to 7.1 million homes, versus Netflix’s 20% increase to 12.4 million homes, due to Amazon’s lower number of subscribers. The number of homes with access to more than one SVOD service has also gone up. As of Q4 2019, 6.0 million homes had access to two or more services. We can also look at the progression of these overlapping subscriptions. Chart 2 compares the Q4 2019 data with Q4 2017 and Q4 2015.Compared with Q4 2015, the number of UK homes with a subscription to at least one of the three SVOD services has more than doubled. Of these homes, just over 42% take more than one service – with the proportion that takes only one continuing to decline. In Q4 2015, homes that had Netflix as their only service represented almost two-thirds of SVOD households. That figure is now at 45%, but we can see that Netflix has not lost those customers, rather these Netflix homes have added other SVOD services. Netflix and Amazon is the most popular combination and has increased significantly, from 9% of SVOD homes in Q4 2015 to almost one in three now. This may be because Amazon offers other benefits as well as television content. The number of households that can’t miss out on any service is also growing; homes with a subscription to all three services now represent 6% of total SVOD households. The UK SVOD market is continuing to fragment with the launch of new services like Disney+ and Apple TV+, so it will be interesting to monitor how the various overlaps develop. Establishment Survey data also provide an insight into the age profile of SVOD users. We can compare Q4 2014 data with Q4 2019 data to see how the proportion of subscribers from different age groups has changed for each of the three SVOD services. Chart 3 illustrates that in Q4 2014, younger age groups made up the highest proportion of users for the SVOD services, as you might expect from early adopters of new technology. For example, 62% of Netflix users were then aged under 35, with similar proportions for the other two services. A younger skew to these services remains in Q4 2019, but there has been a larger proportional growth among older audiences. Users aged 55-74 have increased the most of any group, particularly for Amazon, where the proportion of these users doubled from 2014 to 2019 to 16% of the total. Although older SVOD audiences have grown over the last five years, these groups are still under-represented when compared to their proportion of the general population (see final bar), with younger audiences over-represented. Those aged 55-74 were 22% of the UK population in Q4 2019. While it may be unlikely that the age profile of SVOD service users will reach parity with the general population, it suggests that older audiences could be a subscriber growth opportunity for SVOD services. As SVOD subscription levels rise, we have seen an increase in unidentified viewing, which includes all TV set use that BARB cannot yet identify. This includes viewing to programmes more than 28 days after broadcast, gaming, watching DVDs and watching SVOD services or online video platforms. Unidentified viewing now accounts for 52 minutes of TV set viewing per day on average, or a 21% share of TV set viewing for all individuals, an 8.5% growth since 2018. Over the same period, viewing to BARB-reported channels fell by 4% to 191 minutes per day. It is possible to break down unidentified viewing by ethnic group (chart 4). It’s clear that the increase in unidentified viewing is universal – every ethnic group does more unidentified viewing now than five years ago. There is a wide spread in the levels of unidentified viewing. For example, in black Caribbean (including mixed white/black Caribbean) homes, unidentified viewing is now 66 minutes a day on average, higher than for all individuals. This is more than 15 minutes a day greater than the 49 daily average minutes in Bangladeshi/Indian/Pakistani homes. However, these two figures are both around 30% of the total TV set viewing time in these homes – so there is also a higher amount of TV set usage in black Caribbean homes, compared to other ethnic groups. In fact, for all ethnic groups apart from white homes, the proportion of total TV set viewing time that is unidentified is greater than the average for all individuals. Whether these high levels of unidentified viewing are driven by SVOD services, gaming, or by viewing to overseas-based channels not measured by BARB (or any combination of these), it is clear that viewer behaviour is not one-size-fits-all. BARB data reveal that unidentified use of the TV set is increasingly becoming a primetime activity. Chart 5 looks specifically at the 9-10pm slot for every day of the year to see whether it was won by unidentified viewing as a whole, or a BARB-reported channel. There were 110 days in 2019 when aggregate unidentified viewing was ahead of any single BARB-reported channel in the 9pm slot, almost double the 2018 figure. In a timeslot where broadcasters air their best programming, they are increasingly competing with unidentified use of the TV set as well as with one another. It is important to note that by taking unidentified viewing as a whole, we are not comparing BARB-reported channels with SVOD viewing alone, let alone viewing to individual SVOD services, as unidentified viewing includes anything that isn’t viewing to a BARB-reported channel. We have excluded unidentified viewing via a games console from this analysis, to reduce the likelihood that the use of the TV set is for gaming. That said, the pattern is clear. We know that there is a significant correlation between the increase in unidentified viewing and the rise in the number of SVOD subscriptions. Chart 6 demonstrates the strength of this correlation since 2014, with an almost-perfect R2 of 0.97. This illustrates that SVOD services are the major factor leading to the growth in unidentified viewing. The SVOD services have been the catalyst for significant transformation over the past decade; that these three services have got into half of UK homes in this time is a major change to the television landscape. We know that finding out how much TV set viewing is accounted for by SVOD services is high on the industry’s wish list. BARB began rolling out router meters to panel homes in October 2019. One of the major benefits is that router meters will enable us to measure SVOD viewing at an aggregate level. Alternatively, should these services decide to be BARB-reported, we can also measure audiences to their individual programmes with the same technology we use to measure viewing to broadcasters’ non-linear content. Either way, by 2021, we expect to have a great deal more insight into SVOD viewing.