High volume infringers analysis report

Published 11|09|13

Executive summary

In early 2012 the Intellectual Property Office funded Ofcom to conduct research intended to improve understanding of consumers' behaviour and attitudes towards lawful and unlawful access of copyright material using the internet. The primary objective was to gather data and generate insights that could be used to assist policy making related to online copyright enforcement. This followed the adoption by Government of a recommendation made in the Hargreaves Review of Intellectual Property and Growth.

We have now completed this research programme, with data and findings from the fourth quarterly wave published alongside this report http://stakeholders.ofcom.org.uk/market-data-research/other/telecoms-research/oci-wave4/ . Kantar Media, who conducted the research, processed a total of 21,475 survey responses across the four waves, during the period May 2012 - May 2013. This has provided a robust and substantial base which has been used in this report to conduct a more in-depth analysis on how attitudes, behaviours and spend differ between higher and lower volume infringers, and those who don't infringe at all.

This work expands upon the previous high-volume infringer analysis, which used data from the first two waves of the study, and can be found at: http://stakeholders.ofcom.org.uk/binaries/research/telecoms-research/online-copyright/deep-dive.pdf

The research demonstrated the clear demand for online access to copyright material, with well over half (58%) of internet users downloading or streaming at least one item of content during the year. However, infringement was a minority activity; we estimate that 17% of internet users consumed at least one item of infringing content, which equates to around a third (29%) of all consumers of online content.

When considering individual content types in isolation, infringement levels varied, but in absolute terms were low. The highest incidence of infringement occurred in music, but even here fewer than one in ten (9%) internet users infringed; for software and video games the figure dropped to just 2% of internet users. However, in relative terms infringement appeared more prevalent; a third (33%) of all consumers of online film infringed film at least once, while a quarter (26%) of music consumers infringed music.

On a volume basis, almost a quarter (22%) of all content consumed online during the year was infringing that's over 1.5bn files. However, this differed by content type; infringing films accounted for 35% of all films consumed online, infringing music for 22%, while infringing TV programmes for just 18%.

Looking more closely at the distribution of infringement, it becomes clear that infringement is heavily skewed; a small proportion of infringers account for the large majority of infringements. The data show that just 2% of internet users make up the top decile of infringers by volume, and these people accounted for three-quarters (74%) of all infringements. There is, therefore, a long tail of casual, low level or infrequent infringers.

However, high volume infringement tends to be quite category-specific. For example, there was only limited overlap (a maximum of 36%, and typically much less) between the top 20% of infringers of any two different types of content. This tends to support the approach we have taken in this report of examining content types in isolation. It further suggests that the highest volume infringers are those who have a passion for a particular content type.

We also found that some people appeared to be accessing vast amounts of infringing content. We estimate that on average the top 10% of film infringers consumed 80 infringing films over a three month period. This equates to almost one per day, a figure that is likely to be explained by a combination of interest in film as a content type, and hoarding or stacking films for viewing at a later date.

This distribution of infringers also provides some interesting insights into claimed consumer spend on legitimate content and its relationship with infringement. During an average three-month period, infringers tend to spend more than non-infringers on legal digital content (�26 vs. �16). A similar pattern, although less pronounced, is seen when we include wider spend on offline content-related purchases (�110 vs. �83). Furthermore, the top 10% and top 20% of infringers tend to spend the most, in contrast to the bottom 80%, whose lower spend is more in keeping with non-infringers. But this hides significant variation between content types. For example, among film, software and video game consumers, non-infringers reported a higher average quarterly total spend than infringers.

Infringers were generally skewed towards being male, 16-34, and C2DE compared to non-infringers. Music, film and TV programme infringers were also more likely to be unemployed. These trends were amplified among the top 10% and top 20% infringers. The bottom 80% infringers were typically older, and, for some content types, more likely to be female, than high-volume infringers.

High volume infringers demonstrated a higher level of familiarity with online technology through the means and location they choose in order to access online content. The top 20% of infringers were significantly more likely than the bottom 80% of infringers to have streamed or accessed content either out of their home (37% vs. 25%) or using a mobile network (39% vs. 26%). Given the complexity involved in detecting infringing activity on mobile and out-of-home networks, this may have significant implications for enforcement, particularly if these trends spread to the long tail.

Across all types of infringer, the most commonly given reasons for infringing copyright online were because 'it's free', 'it's easy/convenient' and 'it's quick'. However, there were significant differences between infringer groups; the highest volume infringers were more likely to say they already spend enough on content' (19% for top 10% infringers vs. 7% among the bottom 80%); that 'legal content is too expensive' (38% vs. 13%), that they didn't want to wait for content to become available on legal services (19% vs. 8%); and that 'the industry makes too much money' (19% vs. 8%).

There were also differences in the attitudes towards potential measures designed to encourage people to stop infringing. The top 10% were more likely than the bottom 80% to claim they would stop if they thought they might be sued (25% vs. 17%); if they received a letter from their ISP informing them that their account had been used to infringe (19% vs. 11%); and if there were articles in the media about people being caught (13% vs. 7%). Although these figures are based on claimed rather than actual behaviour, they may suggest that enforcement measures are more likely to be effective with heavier infringers. They also suggest that no single enforcement solution is likely to address online copyright infringement in isolation; a complementary mix of measures including better lawful alternatives, more education about copyright matters, and targeted enforcement is more likely to be successful.

Please note that due to their size, we are not publishing the tables and CSV files. To request a copy of the files, please e-mail Monika Kochanowska-Tym at monika.kochanowska@ofcom.org.uk