by Anna Meadows
Australian pirates announced their intention to change their behaviour once Netflix arrived in 2015. Three years on Australians have embraced not only Netflix but also the many similar offers such as Fetch and Stan. Yet piracy persists. More than 20% of Australian adults still admit to pirating content, and that is likely a conservative figure.
Content creators never expected subscription video on demand (SVOD) to swoop in and solve so complex a consumer behaviour as piracy. Environmental, social, technological and economic factors all influence piracy behaviours. It’s naïve to believe there is a single intervention that will provide a complete solution.
Different strategies to limit piracy behaviour have been tried and tested. For example, studies have shown that increasing the ‘cost’ of piracy though blocking1 or decreasing the quality2 of pirated content can reduce piracy.
SVOD also has a role to play by making it easier for consumers to access what they want when they want it. But its true impact can only be inferred from available data. In Australia claimed piracy has plateaued, SVOD subscriptions have increased but, as any statistician will tell you, correlation does not imply causation. So can we be certain of its influence?
In early 2015 De Matos, Ferreira and Smith conducted a study to understand the true effect of SVOD on piracy3. When the paper was published, journalists jumped to comment on the findings. “Netflix should cut prices to beat piracy” one headline announced, “Pirates are willing to pay up to US$3.25 per month to stream files and TV shows legally, researchers say” stated another.
While journalists may not have wanted to wade through the 20 page, statistically heavy paper, plus appendices, it might provide some explanation for their somewhat misleading headlines.
Let’s look in a bit more detail at what the study proved.
The study itself is high quality. The paper reports the findings from a randomised controlled experiment – the gold standard when it comes to academic research as they establish an unbiased estimate of causal effect. By randomly assigning an intervention to a treatment and control group (of sufficient size) one can be statistically certain that the only difference in behaviour that emerges is as a result of exposure to the treatment.
The study was conducted in partnership, not with Netflix as the headlines may suggest but, with an unnamed multinational telco. It had a very robust sample size of more than 10,000 households. The treatment group were ‘gifted’ a free cinema pack of movies and TV shows for 45 days which resembled a subscription video on demand service. For those 45 days, data on time spent watching TV channels, aggregate Internet usage and identifiable illegal Bit Torrent streams was collected from all households in the sample.
Despite being the gold standard, all studies of this type have limitations and the authors have acknowledged these. For example their method of measuring willingness to pay for SVOD is achieved by secondary analysis of the experimental data. A “better approach” they say would be to “offer SVOD at random prices to a sample of pirates and track their behaviour”. In addition, there are limitations on collecting data at a household level (we don’t know who in the household is doing what) and also, only one form of piracy (likelihood of using Bit Torrent in the study period) is being measured.
So what did they discover? Firstly, during the experiment, treated households (who received the free cinema pack) watched more TV and decreased their internet traffic – evidence that the free cinema pack caused a substitution away from the internet and towards the TV.
However, there was no evidence that the free cinema pack reduced the use of BitTorrent to pirate screen content. To quote the authors: “pirate households in our sample kept pirating”.
Secondly, to explore this further, the authors built a recommender system from Bit Torrent logs. Using this data they inferred pirates’ viewing preferences by modeling what such a system would recommend. They then calculated the level of ‘fit’ between these preferences and what was available on the cinema pack. Here they essentially found that a pirate household was more likely to use the cinema pack if there was a fit between what they wanted to watch and what was available.
To put this in perspective, a catalogue the size of Netflix US (5,600 titles cited in this study) would at best achieve a fit of 50%. This study showed that even when there was a hypothetical 100% predicted fit between a household’s preferences and the content available from the cinema pack, there would be a mere 18% reduction in the likelihood of using BitTorrent. “This suggests that for the vast majority of pirate households, the marginal costs of consuming pirated content is essentially zero” the authors state.
If we let that sink in – when offered content that matches their preferences, for free, when they want to watch it, through a legal channel, the majority of pirates will still choose to pirate.
For a minority of pirates, the authors hypothesise that the marginal cost of piracy cannot be zero since some did choose to turn to SVOD if they had preferences that were aligned with the content offered.
Finally, using a multinomial choice model and setting indicative prices for SVOD, TVOD and piracy options, they estimated the dollar value a pirate places on piracy and the amount they would hypothetically be willing to pay for SVOD instead. This analysis estimates that households whose preferences aligned 50% with the content offered in the cinema pack would be prepared to pay $US3.25 per month for it.
What the analysis doesn’t say is that “Netflix should cut prices to beat piracy”
This study shows that unless the ‘cost’ of piracy is increased then SVOD’s role in reducing piracy will be limited. To have a major effect, content creators would need to provide significantly larger catalogues, make content available earlier and cut the price of SVOD to a fraction of what it is currently, something that the authors of this study conclude “may also affect the production of content and the pace of innovation in the entertainment industry which may ultimately also reduce consumer surplus”.
The study also shows that SVOD’s impact on illegal Bit Torrent use is likely small and possibly would be more effective when used in conjunction with other proven measures such as site blocking which increase the ‘cost’ of piracy.
Well-executed studies such as this support evidence-based decision making for those wanting to limit piracy.
Media over-simplification of the findings and the suggestion that there is a single, simple economic solution to the issue is not helpful.
In the words of one of the authors: “All headlines about Netflix have nothing to do with our paper and were easy titles that don’t reflect our analysis.”4
———–
Anna Meadows is a research and communications professional with over 20 years of experience in advertising, market research, marketing and behavioural science consultancy. Anna founded Sycamore Research in Sydney in 2007 having worked for multinational advertising and research agencies in London, Sydney & Singapore. Anna has worked with some of the world’s largest brands including MasterCard, BBC, Nestle and HSBC. She has been researching copyright infringement since 2008 and has been involved in 17 studies on the topic across Asia Pacific.
Anna is an accredited Qualified Practitioner of Market Research with the Australian Market and Social Research Society and she holds a Master’s degree in Behavioural Science from the London School of Economics.
———–
1Danaher, B., Smith, M. D., & Telang, R. (2015). The Effect of Piracy Website Blocking on Consumer Behavior.
2Geng, X., & Lee, Y. J. (2013). Competing with piracy: A multichannel sequential search approach. Journal of Management Information Systems, 30(2), 159-184.
3Godinho de Matos, M., Ferreira, P., & Smith, M. D. (2017). The Effect of Subscription Video-on-Demand on Piracy: Evidence from a Household-Level Randomized Experiment. Management Science.
4Private correspondence January 2018.