The graphitational pull
The market power gained by Facebook and Google is the first manifestation of a new organism every industry soon will learn to know: The graph.
What resource is driving the growth of the most successful businesses in the last decade? Not oil, not steel, not gold, but metadata. To put it simply: Information from the real world. The more raffined and structured, the more valuable.
The phenomenon is not very well described in management litterature or classical economic theory. There are new natural forces on the block. Right now, the circulation of bits means way more for the making, redistribution and accumulation of financial value than the circulation of atoms. The tables have turned. Facebook is stealing the power of definition from the media industry, Uber is stealing it from the taxi industry, Spotify from the record companies.
Many industry leaders don’t seem to understand what is really happening. In my home country Norway, editors are turning emotional towards Facebook. The high margins are moving away from traditional business models, and many companies seem to have a hard time imagining new strategies to regain the power. The phenomenon they are trying to put into words has changed the reality we designed our language to describe.
Virtual fungus
The first thing we need to acknowledge, is that businesses like Google and Facebook do not fit into the scheme. They are not mutations of publishing companies. They are not software companies. The are not digital service providers. They are something different. If they are going to be challenged, morally or commercially, we need to understand that we are talking about a new species in the jungle of business.
Metadata-creatures.
Should they be compared to something we know, from nature, the most precise metaphor would be virtual fungus. Mushrooms are not primarily living on the surface. More than anything else, they are a complex system of branching, threadlike roots in the soil. The parts visible to us, the fruit, or the interfaces, are just small parts of a much larger underground organism.
Many of the so-called platform services have become important and useful to society. They have become infrastructure. Many users and a high frequency of use means everything to these businesses, because both the strategy and the business model requires access to good listening posts. In a business perspective, the free services provided are trojan horses. By using them, you give away data about yourself. And you get targeted ads in return.
The same way fungal roots, hyphae, sucks nutrients from the soil, the new giants of the internet suck knowledge from the use of its own, and partly other providers’, services. But in the binary soil, there are no roots.
There are graphs.
Knowledge equals power
Google was the first company to go through this transformation. We may call it a metadata-morphosis. Until 2010 the gigantic backlog of search phrases and other metadata was a badly exploited ore of gold. Google organised their knowledge in the knowledge graph, a gigantic and complex data structure that connects the reality of language with the reality of reality. It connects word and terms in semantic models. Different languages are aligned. Names of places are referenced with physical locations. Sentences are connected to intentions. By the use of the graph Google started making more sense of the stuff you write in to the search engine. They created the best map of the world, a massive index of media content, structured way better, and in more dimensions, than the publishers managed themselves.
They are forerunners in machine learning. Today they know a lot, and way more than most competitors, about how the users think, act and articulate their intentions. Next generation of services will get an even heavier tool, the knowledge vault, rumoured to be an otherworldy library of information, still wrapped in mystery. The myth of the machine brain is probably a part of the brand strategy, but the services are clearly becoming smarter by the day. The search engine and an growing spectre of other applications are becoming faster, friendlier, more intuitive. The patterns hidden in the metadata gives access to more insights. Insights give room for improved services. Better services give more metadata. And thus, more power.
Facebook, on the other hand, has established the the social graph. A similar knowledge infrastructure about the social reality. It maps out individuals, groups of individuals and the information that flows in between.
Facebook is also a friendly and intuitive service. Just as Google, they started out with a useful, simple application. And just like the search engine, this application has transformed into a listening post. They are trawling for metadata. Again we see the same feedback-effect. Useful services returning small pieces of information. The graph connecting the dots, giving insight. Improved services. More power. That can be used to do good stuff. Or bad.
Scary, smart — and here to stay
A graph is a technical term. It describes a spesific way of organising data in a database. A graph organises data in nodes, and also the relationship between these nodes. It can look a bit like the structure of the brain: Neurons and synapses. A comparison may be a map connected to road sensors. A city is a node. The roads between them are connections between the nodes. By measuring the traffic on the roads, you can find out wich way is shortest or fastest or most popular. You can rank the cities by economic importance. And, if you have a really smart sensor, you can register the flow of different products in a regional economic system. A wide range of conclusions can be drawn by mapping the structure and observing what happens within it.
As humans we live somewhere, we have an age, a gender, relations to other humans, we have interests, political views and good and bad habits. A graph database can organise information along all these dimensions. It is not only concerned with who we are in demographical terms, but how we change, and what we can turn into, especially as consumers. The flow of information is important. By analysing data along a timeline, algorithmic models can also predict behaviour. And the service can place marketing with a high degree of precision into our search results or personal feeds.
The graph is scary. It has laid the foundation for what some critics call surveillance capitalism. It challenges democratic principles. The surveillance regime revealed by Edward Snowden was all about tapping into different sources of metadata, and organising them in a stately owned mothership graph.
At the same time, the graphs are making the world better, on a daily basis. This principle of organising data lays the foundations for important research on systemic stuff such as climate change, and important breakthroughs in the health sector. The groundbreaking collaborative journalistic research in the Panama Papers was possible thanks to a graph database. The graph can help bring along a green energy shift. It can fix transportation problems in our cities.
We need this kind of digital infrastructure, so-called platforms. They are useful to us. And platforms need the graph. Just like the newspaper needed the printing press. Just like the car needed the combustion engine. Just like the lightbulb needs the electricity grid and the turbine. Graphs are here to stay. And there will be more of them.
This is why I suggest this new word: Graphitation. A term that describes a natural phenomenon. A natural law for the post industrial digitized reality.
Graphitation is in my eyes a natural force in any network where binary information changes hands. A force we have to learn to live with, take advantage of, adapt to, build protection against, embrace. Just like we do with other forces of nature.
The law of graphitation works like this:
From messy to structured data
When a process is digitized, machine-readable metadata is released. In the analog process of the pre-digital world, this data is locked in and inaccessible. In the early stages of digitization the metadata is very chaotic. The same things have different names. Many things are named the same. The relations are uncertain. Precision in the system is low. There is lots of noise. Computer engineers are frustrated.
Sooner or later someone will be in the position to structure the data. They give the data unique identifiers and organise them in a relational data model: A graph.
Graphs give access to patterns. Processors can automate pattern recognition. Algorithms built into the service can use the patterns to improve the service. The service improves. More people use them. More metadata is being fed into the graph.
The graphitational pull increases.
If your graph are becoming dominating within a value network, you can open up the structure of the graph for third parties. It becomes, due to sheer dead weight, a “standard” just like OpenGraph has become for everyone who wants their content to look nice in the Facebook feed. The metadata starts flowing into the graph by its own free will.
The graph is galvanised by trust and user habits. End users and third parties need to feed it through a set of services. When we started letting the Facebook feed into our everyday life, we also let a set of routine actions into our everyday lives, we included the service in the way we interact and the way we speak.
We like stuff, we follow people, we block and we unfriend. And if we wonder about anything, we google it. Every action is recorded in a graph.
At the same time the fungal root system keeps growing. And it grows past the fence of the garden from which it originated. We accept terms that enable the graph to get access to the sensors of our devices. The graph knows where we are, who we are with, in the physical world. Many of the online newspapers that now fear Facebook, have integrated their services in ways that enable the graph to track their users while browsing the news.
The media have been feeding their own monster.
All our tiny actions are beeing harvested, transported in real time into big data centers and converted into structural insight. The owner of the graph will, at least in theory, know more about parts of your life than you do yourself. This is certainly problematic. But it might just be part of the deal.
When trust is established, user habits encoded into our brains, the infrastructure has become default and metadata flows orderly into the structure, the owners of the graph will be in a position to manipulate the services for their own benefit. By defining rules in the algorithm the graph can be used to nudge the relationships between the people in the network. By opening up for external parties, like companies that are not a part of the native social structure, the platform can start charging access fees. You can get access to the interaction on commercial terms.
Owning a dominant graph is the most scalable business model thinkable. Owning a graph that scales like Facebook is like owning a money printer. That is why Mark Zuckerberg has become so rich. Together with a handful of other players he has been alone in this landscape, unchallenged by universal standards, smart competition and regulations.
That will probably, and hopefully, change in the years to come.
(This is the english version of a blogposts that I recently published in norwegian. I have also written a follow up-piece — called “strategy in the time of graphitation” — still only online in norwegian. I will translate the second part into into english in the near future.)
