The Evolution of Video Streaming

Surveying the landscape: 

As of writing, the leading players in video streaming are Youtube, Netflix, Amazon, and because this is not 2005, Apple is lagging behind in the media space.

Although these outlets are different, they fall under the same general umbrella. However, the video streaming market could be further subdivided into two camps- platforms and services.

Platforms vs services: 

Websites, like Youtube, are platforms and others like Netflix, are services

Netflix, Disney, and HBO are services, in the sense that you provide them with money, and they will provide you with pre-selected content. Content that they decide is worthy of your attention.

On the other side of the spectrum we have platforms. A platform, in the general sense, (roughly) is when the total of all added value of the users of that platform surpasses the value add of the platform itself. Something like Microsoft is considered a platform. Where the total that humanity was able to achieve with Windows XP from electronic oil drilling, to aeroplanes to the pure distribution of knowledge is greater than the value of the Windows operating system itself. 

Youtube is a platform, as instead of pre-selecting content, they aggregate their content from the crowd. Youtube gives the average person the ability to broadcast themselves via broadband. And outside of their terms and conditions, Youtube does not curate the content. 

Before Youtube, the closest thing you would be able to get to this would be broadcasting your own radio show. And let’s face it, you were definitely a weirdo if you were broadcasting your voice via radio waves. 

This means that one company (Netflix) has to pay billions to get a single click while the other (Youtube) just needs to wait for a viral cat video. And although per piece of content whatever Netflix produces is probably better than YouTube (even the Adam Sandler movies). Youtube has the ultimate advantage. 

As Napoleon said: 

Quantity has a quality all its own.

It is worth remembering that the internet remains undefeated. You cannot out-create it no matter your budget. People are migrating to the internet not only because of the change from satellite to bandwidth. But also the content on the internet is not overproduced drivel. As on TV, there are 9 producers who will ruin anything to justify their existence. Additionally, the idea of a platform absent of gatekeepers is not specific to Youtube. But is one of the underlying business models of the internet. 

At this point, you must be thinking Netflix provides different video content than Youtube. So you could say that there will always be a demand for a product like Netflix. 

But then I would say, yes there will always be a demand for high budget content. But what makes you think high budget content won’t be found in the same space as aggregated crowdsource content in the next 5 to 10 years?

Fractal competition: 

This is because tech firms are trying to attract as much as your attention as possible. Because their currency of interest is no longer the US dollar but your attention span as well. Said another way, the attention economy underpins all other economies.

This more intertwined competition in tech allows for a more of an isomorphism to occur in the field. Isomorphism in a business sense is (roughly) the phenomena where companies which compete with each other become more like each other.

This is taken to the nth level in tech. Where you have a video game company (Fortnite), competing with a media company (Netflix), competing with an online store (Amazon), who is then competing with a search engine (Google). 

As everyone is competing for the same output, each individual internet firm is competing with each other firm through some sort of fractal competition structure. Fractal in the sense that companies are competing at different scales/products. And is, in turn, becoming more like their competitor in terms of product offerings (and a variety of other factors).

That is why everyone, their mother and even Facebook (given their privacy history) are competing in the smart speaker market (still funny.) Because if you control the home, you control most of the attention of said home.

So although Google, Amazon, Twitter want to do different things with your attention, they all want it equally and desperately. 

How will the future of video streaming look? 

In the future, you are going to get a bifurcation in the types of video streaming outlets. Some services will survive and keep producing the kind of content they’ve always been producing (think HBO). And another group of platforms which combine both aggregated content and pre-selected content in the same space. 

In this model, the aggregated content would be free, and you would need to pay a subscription fee for the dramatic content. 

Formally, the best way to think about this would be a two-part tariff in economics where you are still getting the benefits of the network effect by having people use your website for free. But you are also monetizing those who would be interested in paying to watch your dramatic content. So, in essence, you are able to monetize two different subscriber bases. 

This has the added benefit of in-housing your advertisement. Where instead of paying other people to advertise the next season of Westworld (like is traditionally done). You can just run ads on the people who use your platform for free. This has the effect of virtually integrating your advertising cost. 

Who is in the best position to win the streaming wars? Short answer- Amazon:

Amazon has the best chance of winning the streaming wars because it already has a video streaming service (Amazon Video), a live streaming service (Twitch) that are still currently separate. They now only need to come up with a real rival to Youtube (no offence Vimeo) before they have a slice of all the streaming pie. I would even go so far as to argue that Amazon has some of the best types of content on the internet.

It’s also a great business move because Amazon vertically integrates their costs through AWS. But also manage a horizontal integration as they are taking up more of the attention pie, from their main rival no less. 

Amazon’s potential advantage over Google:

I think it is fair to say at this point, Amazon and Google are playing by different rules. Amazon, unlike Google, doesn’t have to generate profits for a quarterly earnings call. The principal metric is being judged on is its rate of growth.

If you want to map this formally, Amazon in a game theory sense is playing an infinite game while Google is playing a finite game. Because of this, Amazon is able to think in a longer time frame, this vision for the future, allows them to focus on the long term while their competitors are increasingly obsessed with their next quota. 

So how does this advantage translate into a rival against Youtube? 

Like I mentioned previously, for now, Youtube is predominantly a crowdsourced platform. 

So to attract the consumers, you must first focus on the producers. 

Content creators are fickle. Especially nowadays, due to the lack of competition, they are severly underpriced by the market. 

Amazon could give its platform a structural edge by increasing its advertisement share (ad-share) with content creators. This is the start of a dynamic process. The more you pay the independent content creators on your platform, the higher the quality of content they produce. 

Ben Thompson (who is very good) makes a good point of explaining that the Apple App store is stifled because Apple insists on taxing developers. Not allowing them to spend their additional profits back into actually making, you know, apps. This feedback loop would probably result in them earning more in the end, but alas, Tim Cook has quarterly earnings to prepare for. 

As a side note, as we see Apple slowly drift away from the iPhone, expect some of this sort of rent-seeking behaviour to creep up more and more as Apple tries to keep up their share price.

What type of data analysis could Amazon use: 

Amazon could use Network Analysis to find out who the main drivers of flows are from Youtube (or whatever proxy service they want to use). And give them an exclusivity deal. The old media equivalent would be what Howard Stern did for Sirius XM Radio As he is the one who brought most of the attention that they could later build on.

As of right now, Youtube has made the bet that they bring viewers to content creators and not the other way around anymore. They are wrong. The internet (and video streaming by extension) is highly Pareto. Gobble up as many Jake Paul’s as you can (while you still can). 

And if you want to maximize profits, just take a page out of Moneyball. Go after the influencers with the highest ROIs based on whatever criteria you are trying to attract. 

Why Netflix was never going to win the streaming wars:

The narrative at least early on was that Netflix (by construction) was going to run content on the internet. Or at the very least be the number one player in the field. Maybe not everyone thought that. But that was the narrative.

In their grand vision, Netflix became a gatekeeper for content on the internet. Meaning that they gave you money, and you gave them the new opium of the masses (TV). But the reason why that was never going to work is that Netflix cannot have a monopoly on good ideas. They cannot just out-content everyone forever. So all it took for someone to compete or even beat them was just better content available on the internet as well.

This means that the only advantage Netflix ever had was the technology gap between them and old media. Said formally, Netflix had a first-mover advantage. Because they were on the internet first they did not have to compete with anyone, so we’re able to establish a subscriber base.

But the problem with the first-mover advantage is that (like everything else) it fades away. The issue with always was/is for Netflix to beat Disney they need to become Disney. While to beat Netflix, all Disney has to do is register an IP and take out some server space.

Netflix does have one advantage of traditional gatekeepers for now. Instead of overproducing everything, they give people a lot of money and don’t get too involved. Which definitely gives Netflix a competitive advantage.

Machine learning and video streaming:

There used to be a lot written about how omnipotent Netlfix’s algorithm is. But it turns out that algo could be improved upon. 

In order to do machine learning, you need to be able to sort data into categories to then learn from it. Because an unsupervised machine learning algorithm which is able to group and sort tv and film just by watching it is still very far off. Because of that we need to rely on humans to fit these movies into different categories by manually watching them.

Then based on the different categories customers are put into different taste profiles. These taste profiles then instruct what type of content Netflix recommends to you.

Netflix estimates that these algo recommendations account for 75% of their viewer activity is driven by recommendations. I am hesitant to take this as a marker of effectiveness. Because if you present any human being with a menu of choices they are going to pick from it at a very high rate. It is also worth keeping in mind how a metric like that could be gammed (which is often the case).

If 75% of viewership activity comes from recommendation then the other 25% comes from custom searches on Netflix. And I personally believe most of the edge is in the remaining 25%.

Data and revealed preferences:

Netflix knows the type of content that their consumers want to see but can’t because it is not on Netflix. This is called an intervention (more on that later).

So unlike the method of association that Netflix currently uses. Which recommends what people want to see based what is already there. The method of intervention allows Netflix to know what you want to watch, so it could either acquire the rights to it (where their margins are worse). Or could make things right with the universe by rebooting the Matrix franchise. But from the second movie because that is when things when downhill. And with Keanu reeves playing young Keanu Reeves because he still looks the same.

With this data, Netflix can use these searches to figure out what people want, and then produce their own content to match their need. This way, Netflix would always be on the pulse of its customers. Making their offering more complete than that of their peers.

The reason why the intervention is superior to the association is due to the fundamental limitation to data itself and beyond the current scope of this article. Where, in reality, data is fundamentally limited to what questions it can and can’t answer. The current understanding seems to be that through statistical analysis, we are able to tease out the answers to our questions. But the problem has always has been that data can only tells you what happened in the past, nothing else.

Current machine learning algorithms run on association. Which is squarely in the bottom rank on the ladder of causality. So in the case of Netflix, they are looking at what types of shows you watched and trying to find a good enough association from another show to recommend it for you. 

  1. Association (seeing): Finding variables which are related. Example: Given you already watched Friends, what TV have people in your taste profile watched most afterwards. 
  2. Intervention (doing): Allowing the customer to reveal their preference to you by searching. Example: Given you wanted to watch the Matrix, here are what movies it is in a similar category to.

However, the data you collect in the Matrix example falls under the second rung on the ladder of causation. Meaning that your customers are revealing their preferences to you directly. 

So you do not have use data to get an answer. As the first rung data will still be worse in predicting the preference of your customers. When compared to allowing them to reveal their preferences to you

The closest thing to this in traditional industry would be focus groups. But that isn’t even doing this data suggest. Because unlike focus groups, Netflix gets to observe the hedonistic consumer in his natural habitat.