The Dark Side Strategy

I recently watched a Netflix documentary called ‘The Great Hack’. Some of you will probably have seen it or at least will have been presented it in your feeds – thanks algorithms. For those that haven’t, it’s an excellent and gripping film about Cambridge Analytica and their nefarious dealings in social manipulation for Trump and Brexit.

The doco takes you behind the scenes and into the lives of ex-employees and whistleblowers, as well as revealing the extent of their data enrichment programs and behavioral change capabilities. It’s nothing short of fascinating and terrifying all at once.

The Great Hack: Netflix
The Great Hack: Netflix

Beyond the data though, there were the far more disturbing strategies that sat behind the technology; the manipulation and behavioral tactics these agents use to swing elections and influence voter mindsets.

The documentary, for example, features a case where social media was used to create an anti-vote movement amongst Trinidanian Youth – this powerful uprising tapped into an apathetic generation and quickly swelled.

In the film, Cambridge Analytica says it worked for “the Indians” – implying they worked on behalf of the majority-Indian United National Congress (UNC) party. According to the film, the inaction of this voter segment meant a 6% swing was achieved in what was considered to be a neck and neck race. This weakness was exploited through an anonymous Dark Side strategy that exploited fake news, privacy data and misdirected public sentiment.

These, of course, are the tools of politics, the Dark Side strategies that political strategists employ to not only activate advocacy amongst supportive bases but to also disrupt and nullify oppositions. These are the same forces that influenced Brexit, and Trump’s win.

Which got me thinking about advertising. Because from where I sit, it seems that the majority of advertising strategy is what I would call ‘Light Side’ or tactics and messages designed to persuade apathetic or casual buyers to buy a brand over another. There is very little by way of Dark Side strategy, actively discrediting another brand or rendering their audience impotent.

There are of course some exceptions and even some famous public stoushes, but for the most part, I think advertisers play a pretty fair and above board game. There are few instances I could readily think of where Dark Side strategy is central to a brand’s ongoing strategy or if they do exist, there’s a very good reason they are invisible.

I would suggest there are some examples of ‘Grey side strategies’ about; the famous Mac v. PC comparisons, for example, which threw shade at a competitor in a funny way. Or much more seriously, the infamous research and medical propaganda of the tobacco industry, which for years waged war against the medical community and its warnings about the dangers of smoking.

So why don’t more brands consider Dark Side approaches?

There could be a number of reasons. Perhaps brands don’t want to be seen as manipulative or risk brand damage. Maybe marketing leaders are inherently good and not Dark Side inclined or perhaps budgets don’t allow the exploration of concurrent strategies. And there’s one more possibility, maybe those responsible for strategy have just not yet really considered it.

I’m open to the Dark-Side (at least as a thought experiment).

Andrew Reeves, Communications Director, The Royals

What I learnt about recommendation engines when I built my own

Streaming platforms like Netflix seem to know what you should be watching before you do. To better understand how and why, film buff and The Royals’ data scientist Dr Paul Vella built his own recommendation engine.

Netflix devotes a staggering amount of time, money and computational power to keep me happy, content and watching. But why do they think they know me so well? Every time their algorithm makes a recommendation, there’s a risk I might not like it and will consider switching to Stan (psych!).

But according to a 2017 article published in Wired, more than 80 per cent of the TV shows people watch on Netflix are discovered through the platform’s recommendation system. And Netflix are definitely not alone in being a recommendation-obsessed content provider.

Formulas have been implemented across Spotify, Amazon, YouTube and other platforms to recommend anything and everything. You could say they’re as common as Game of Thrones spoiler alerts on social media.

So why do these companies think I would like the songs, books and films they recommend? How did they reach those conclusions about me?

Man sitting in gallery viewing blank art work.

To satisfy my curiosity, I decided to try my hand at building a film recommendation system and see for myself how content providers arrive at their conclusions. The point was not to build a proper model per se, but to understand the inner logic of these systems and their potential use.

There are many different techniques for building recommendation systems. And approaches involve NLP (natural language processing), vector factorisation, nearest neighbour clustering and similarity indices.

Stay with me.

Because if you take a step back from the ‘technique’ and think about the methodology (or purpose) of these approaches, all of them are trying to do one of two things:

  1. Recommend items that people who are similar to you like (called collaborative filtering)
  2. Recommend items that have similar attributes to others you like (called content-based filtering)

A third, hybrid filtering approach combines these two, then applies weighting to reach a recommendation. And the logic behind each can be set out in a relatively straightforward way:

Collaborative filtering

  1. Aaron and Bob both like Jurassic Park (1993)
  2. Aaron also likes Ready Player One (2018)
  3. Bob hasn’t seen Ready Player One
  4. Recommend Ready Player One to Bob

Content-based filtering

  1. Aaron likes Jurassic Park (1993)
  2. Jurassic Park is an action movie and so is The Meg (2018)
  3. Bob hasn’t seen The Meg
  4. Recommend The Meg to Bob

Hybrid filtering

  1. Aaron and Bob both like Jurassic Park (1993)
  2. Aaron also likes Ready Player One (2018)
  3. Aaron 14 and Bob is 36
  4. 30-somethings aren’t into Ready Player One (2018), they like The Commuter (2018)
  5. The Commuter and Jurassic Park are both action movies
  6. Recommend The Commuter to Bob

To give an example of how this works on a larger scale, let’s look at Spotify. Their algorithm is pretty complex, and takes in data about what you’ve listened to and how long for, what you’ve liked or added to playlists, and more granular elements of the songs themselves like genre, tempo and duration. It also pays attention to what others who have similar preferences to you have listened to or liked.

The model I built in a Google Sheet is based on a much simpler collection of information. It recommends films from a list and tracks just two variables: when I last watched a film, and how much I like the genres the film fits into.

Viewership

The logic behind tracking the date I last watched a film is pretty simple:

  • Films that I’ve watched most recently shouldn’t be highly recommended.
  • Films that I haven’t watched should be highly recommended.
  • The longer it has been since I last watched a film, the more highly it should be recommended.

The viewership score is therefore just a count of the number of days since I’ve last seen the film. This puts less importance on films I’ve seen recently and more on those I haven’t seen for a while.

To get a viewership score for films I haven’t seen, I simply take the maximum number of days from the films I have seen. This means films I haven’t seen in a long time and films I haven’t seen at all are equally weighted.

Genre Preference

I also kept the logic behind the genre preference score simple:

  • Films can be classified in many categories. Avatar (2009), for example, contains elements of science fiction, futuristic, fantasy and adventure films.
  • Giving a film a rating (out of five stars) counts equally across all genres (attributes) of the film.
  • The genre preference score is therefore the sum of ratings given to all films in that genre.

Table of genre preference scores

This simple calculation reveals I prefer sci-fi and action films over drama, which is true.

Getting to a Recommendation Score

Since both variables are integers and there’s no logically necessary reason to place more importance on one or the other, I simply add the scores together to arrive at a recommendation rating (the higher the score, the higher the recommendation):

Table of recommendation ratings data

The Data

Now you know the mechanics behind a relatively simple content recommendation system, let’s see how good it’s been at improving my movie nights.

I have around 1,016 films in my database. And I’ve given a rating to 712 of these. I’ve watched 165. Given I can watch one film a night – well, two, if the first one was terrible – it took roughly six months of data collection before the system was recommending films I’d actually consider watching. This is evidenced by how strongly it kept recommending Eat, Pray, Love (2010). Ugh.

If I arrange my film ratings by date from Jan 1, 2018 to Apr 1, 2019, a simple linear regression reveals a slight positive trend in my ratings (it is a five-point scale after all, so any positive trend has to be small). So, there’s some evidence the films I’m watching more recently are getting better ratings – and therefore my movie nights are more enjoyable.

So what did I learn about recommendation engines?

  • I can trust my spreadsheet’s recommendation more than a friend’s opinion

Anyone can build a recommendation model, and it will probably improve your choices. The system I designed doesn’t include any Python code or API calls, just a few fancy spreadsheet formulas and some stats know-how.

An element of DIY is probably better, anyway, because I can classify films the way I like. For example, I can break down ‘sci-fi’ into 10 micro-classifications (futuristic, time travel, zombies, etc) I am interested in, giving more accurate recommendations than just using ‘sci-fi’ on its own.

The more you can describe the elements in a set of choices, the better the model can be at recommending things you might like. Harvard’s cognitive psychologist George Miller famously published research back in the 50s that showed we can only hold about seven items in our short-term memory (or in this case, make a choice from around seven films).

And how many elements of those can we compare? Because a recommendation model can make suggestions based on hundreds, thousands or millions of elements.

  • You can uncover patterns in your decision making you didn’t even know you make

Since I was tracking the order I watch films and their genres, it was possible to build a database of which genres I would tend to watch next by finding patterns in my preferences.

For example, if I watch a crime film, there’s a moderate association (0.29) that the next film will be a fantasy film. And if I watch an action film, there is a negative association (-0.15) that the next film will be a superhero film. That’s probably because my wife will want to watch something else!

  • My feelings still play a part, they’re just quantified

It may come as a surprise, but recommendation engines are entirely reliant on the way a person feels. All the data and analysis in my film recommendation engine comes from two variables: my ratings of the films (how much I liked them), and when I last watched the films (was interested enough to act).

Netflix does the same thing, just in a more complex way. Its recommendation algorithm considers what you’ve watched, when and how long for, the order you watch films or series, your ratings, and the ratings given by other members who are similar to you.

The more descriptive these algorithms get, the better their recommendations are – to the point of factoring in ‘hyper-specific micro genres‘ I’ve proved to be at least curious about. Even the artwork of their content is displayed based on what I’ve engaged with in the past.

  • You can flip the system to make predictions

Probably the most interesting take-away from building a recommendation engine is the possibility of extracting the importance scores or average ratings to make a prediction of how much I might like movies that aren’t yet released.

There are 13 films in my database that fall into the space, action and adventure genres, and they have an average rating of 3.15 stars. Does this mean I’d give Star Wars: The Rise of Skywalker three stars when it comes out at the end of the year? Will I be disappointed?

I’d probably have to use something a lot more advanced to get a more accurate prediction, something that might work out the part-worth (choice-based conjoint analysis) or standardised beta coefficients (stepwise linear regression) of individual aspects of films (actors, directors, release year, genres, etc) which could be used in as inputs in a model of my film ratings.

I could then use this model on a list of films being released over the next year or so, to filter them down those I’m most likely to give five stars to, all without the need to rely on other people’s opinions.

But first, I’m off to watch Hot Fuzz (2007), because an algorithm told me I’d like it.

– Dr Paul Vella

eSports: it’s game time

 

UPDATE: recently, Royals partner-at-large, Andrew Siwka, was invited onto Paul Gardner’s Radio 2GB show to chat about the eSports phenomenon. Have a listen here.


The following post was originally seen in  Everybody Knows, our new weekly newsletter publication exploring the realms of creativity, popular culture, media, art and technology. – one topic at a time. Consider subscribing now.


Picture this: Millions of teenagers glued to screens across the world, watching their favourite athletes battle it out in leagues, competitions and matches to the death – but here’s the odd thing, the athletes are hardly moving. At least in the physical world. In the near future, it’s what your kids, nephews and nieces will be following instead of an AFL or NRL team.

Why? Because eSports is live competitive video gaming. A sports industry that builds tournaments and competitions for professional gamers to compete against each other for huge prizes and the glory and adoration of million of fans. It’s exciting, widely distributed and can make kids rich.

The setups are epic. Think World Wrestling Federation with cheerleaders, spectacular stages, lasers, booming music, screaming fans and cosplay. Young, (often) male, energy drink-fuelled players vie for shares in prize pools of up to $10M. Huge pressure, epic rewards, celebrity status ensured for the victors.

Here the players are the rock stars, tethered to their high performance PCs by Beats headphones, their stages are the worlds and levels within games like League of Legends and Starcraft.

Newzoo estimates that there 205 million people who watch eSports globally, and interestingly 40% of that audience don’t even play the games they watch. Last year’s League of Legends championship, for example, drew nearly 30 million viewers, putting it in line with the combined viewership of the 2014 MLB and NBA finals in the US.

Originating in Asia, the eSports industry is now gaining popularity in the US and Europe, and that means prizes, sponsorships and endorsements are starting to reflect global impact. Global revenue is expected to reach $US465M by 2017, tripling in 3 years, the main benefactors being the publishers, event organisers and organised teams. The money itself comes from brands that want to tap into a highly attractive group of young, educated, and wealthy people. These fanbases are deeply engaged, in both a media and community sense, and are increasingly hard to reach via traditional channels.

And of course, where money goes, scandal often follows. In July this year, there were revelations that a Counterstrike-winning team had admitted to widespread use of the ADHD medication, Adderall. Normally performance enhancing drugs that are associated with physical sports build an advantage in strength or stamina. But in eSports, it’s all about concentration levels and reaction times.

Then in August, Valve’s own DOTA 2 championship was brought to a screaming halt by a denial of service attack (DDos). In physical sporting terms, this would be like hundreds of thousands of people defying security and running onto the field. Except it’s millions. And really, really hard to stop.

eSports matters because it’s an example of an industry that is built for today, one that has truly embraced a global opportunity by putting players and fans at the centre of its brand of entertainment. By using live streaming, interactivity and fan involvement, eSports delivers an experience for audiences used to being involved in their entertainment – something traditional sports will always find hard to match.
Don’t take my word for it: ask a kid.

Andrew Reeves

The little machine that commits one hundred crimes a second.

The following post is Edition 9 of Everybody Knows, our new weekly newsletter publication exploring the realms of creativity, popular culture, media, art and technology. – one topic at a time. Consider subscribing now.


 

What have you been watching this Summer? Did you pay for it all? I only ask because Aussies love to download illegally. Yep, we are right up there ranking in the top five when it comes to accessing film, TV and music on the sly. The premiere of season 5 of Game of Thrones triggered a record rate of “more than a million and a half downloads in a day”. Torrent Freak has previously revealed Australia as the leader (11.6%) in illegal sharing of Game of Thrones episodes, followed by the US (9.3%) and UK (5.8%). Aussie, Aussie, Aussie..!

Despite the laws, and sometime lawsuits, we seem to happily ignore the reality that copyright infringement is indeed a crime. And despite 30% of Australians readily admitting to committing that crime, we never expect anyone to come knocking with a court order seeking damages to the tune of hundreds of thousands of dollars.

But that does happen from time to time – and those damage claims sometimes seem nuts. In the US in 2009 a Minnesota woman was famously fined $1.9M for downloading 24 songs. Locally, there was also the recent case of the Dallas Buyers Club where some 5,000 iiNet account holders were accused of torrenting the file. While they initially faced ludicrous fines in the millions, a judge effectively delivered a verdict in December of 2015 reducing those claims to a $20 single use license fee.

It’s these incredible fines that have got the back up of Peter Sunde the co-founder of the now infamous Pirate Bay. And it’s what he has done about it that has primed this edition of Everybody Knows which features an art project simply know as “The Kopimashin”.

The Kopimashin (copy machine) is a project created by Sunde under the banner of art collective Konsthack, a playful jibe on the Swedish art institute “Konstfack” (the faculty of art). The collection launched in 2015 ‘to play with hacking as an art form for political and artistic reasons and to question the status quo’.

The Kopimashin is what it says on the tin – it’s the ultimate copying machine. Using some pretty simple tech: a Raspberry Pi, an LCD display and some Python code the “Kopimashin” makes 100 copies of the Gnarls Barkley track “Crazy” every second. Based on a current run rate of eight million copies per day this translates to theoretical ‘losses’ to the label of more than $10 million – more than enough to put them to of business. But the copies aren’t stored permanently making this a great political statement but not one that can be easily taken further in the courts.

kopimashin

So why is Sunde doing this? Here’s a guy who spent time in jail last year for his role in facilitating copyright infringement with Pirate Bay.

The Kopimashin is firstly a manifestation of his commitment to ‘Kopimi’ (‘Copy me’) a belief system, symbol and set of values ‘that may be considered an anti-copyright notice’. This is an ideology that he and his fellow Pirate Bay founder are truly passionate about. And its other fans are so dedicated to it that ‘Kopism’ is now a recognised religion in Sweden. Its central belief is that everyone has the right to and should freely copy and manipulate anything they want.

And of course, it’s a statement about the ridiculous damages license owners attempt to dish out. They know regular people with a copies of Game of Thrones or Dallas Buyer club cannot and never will pay these fines – they’re simply tactics used to intimidate, silence and scare people into obedience.

Kopimashin ultimately queries our own ethics as copiers and copyright infringers. But it also interrogates the methods of a content industry that continues to employ seemingly outdated and hamfisted ways of policing behaviour around original content.

What will you be watching tonight?

Andrew Reeves

Back with another one of those ad blocking bleats

The following post is Edition 1 of Everybody Knows, our new weekly newsletter publication exploring the realms of creativity, popular culture, media, art and technology. – one topic at a time. Consider subscribing now.


So let’s get this straight: online ad blocking is not new. Since the dawn of the web, people have been trying and buying a variety of ways to stop the blinking, flashing and irritation. And as more and more options like Chrome extensions and mobile plugins have become available, in recent years ad blocking has become more widespread amongst a non-tech audience. But over the last couple of months, ad blocking has begun to pop up in more and more industry forums, vendor pitches and marketing meetings. There was the widely distributed study suggesting far wider use of ad blocking than previously thought. Then there was the tale of the celebrated app developer who famously announced it ‘just doesn’t feel good’ to make ad blocking software. And of course, when Apple moves, people take note. By incorporating ad blocking features into its Safari browser in iOS9, the industry conversation has recently moved from a constant mumble into a noisey bar room public debate.

From a behavioural consumption perspective, this feels like it might just be a natural evolution from TV’s ad-skipping and time-shifting. Alternatively, ad blocking could be seen as part of the progression from listening to commercial radio to paying Spotify and others to help us avoid ads completely. But when it comes to online advertising, it’s hard not to consider that we may be going into a cycle of ‘Creative destruction’. This is a term, also know as ‘Schupmter’s Gale’, coined by economist Joseph Schumpter, that describes the “process of industrial mutation that incessantly revolutionizes the economic structure from within, destroying the old one and incessantly creating a new one”. In other words, in this case the online ad industry might well be eating itself from within, a process hastened by its own dodgy practices, like bloated code and ad assets, and its appetite for ultimate revenue at the expense of usability and privacy. Are ad blocking people the real problem?

unnamed

Then there’s the ethics of ad blocking. Publishers, and all sorts of makers of online content, could have today’s types of revenue streams decimated in a world that embraces turning off ads. It will be the long tail of publishers hit hardest, those that don’t have close relationships with brands or agencies thus are not able to collaborate on bespoke content or arrangements. As consumers, many of us claim that if ad platforms behaved themselves, we wouldn’t object to ads at all. Or if we could simply pay publishers and creators for the content that we end up consuming every month, we’d be fine with that. But what we say and how we act are often two very different things. And many smaller publishers may not have the time to find out.

With fat ad creative and intrusive tracking code, it’s mobile users that get hit hardest. Getting rid of mobile ads can help reduce unwanted data collection and privacy infractions – and massively save on mobile data bills. It’s conceivable that with the move to mobile for the majority of our daily online activities, combined with the increasingly available mobile ad avoidance options, we’ll see a huge spike in ad blocking on our favourite devices. Mobiles put all issues of usability right up in your face – and there’s no bigger impediment to content consumption than bad ads.

Much of the industry is looking to native advertising to retain an embedded marketing presence in front of the ad blockers. But there are concerns that many of these native ad units may also fall foul of blocking software. The report from Adobe and Pagefair recently estimated ad blockers will cost publishers nearly $22 billion in revenue this year, and that nearly 200 million people worldwide already regularly block ads. These numbers seem on the high side compared to other estimates, but they surely suggest that the issue isn’t unsubstantial.

The IAB has now admitted we all messed up. A tidal wave of ad blocking might still only be on the horizon, but in the meantime let’s try and make better ads, use empathetic formats, let’s temper our greed, and for god’s sake, think of the children!

 

Everybody Knows

Now and then at the Royals, we look to engage with our friends and treasured community with publications, perspectives and ideas from the periphery. Everybody Knows is our, singular, compact electronic newsletter to help you explore the realms of creativity, popular culture, media, art and technology. – one topic at a time. We don’t want to bombard you (this time!) with a bunch of links, we only want you to pause for four minutes, read, digest and consider. And we’d love you to do it once a week. We want to feed your insatiable need for the intriguing, the unorthodox and the curious. Our only ambition is too completely remix your understanding of the world around you. One email at a time. Is that so much to ask?

Pop over here and subscribe if you’re keen:
http://www.theroyals.com.au/everybodyknows

Dave
@daveking

Subscribe to Everybody Knows

Everybody Knows is our singular, compact electronic newsletter to help you explore the realms of creativity, popular culture, media, art and technology – one topic at a time. We won’t bombard you (this time!) with a bunch of links, we only want you to pause for four minutes, read, digest and consider. So signup below and if you know anyone else who thinks about how things are, how they used to be and how they might turn out, forward this on to them too.