Why platforms leak

There was a time when institutions, both public as well as private, were intent on vertical integration. And it made sense. For a number of reasons. Firstly, it was the industrial age, and linear repetitive processes with low standard deviation was considered good. Secondly, land, labour and capital, the things we used to call the “factors of production”, were intrinsically immobile. Thirdly, the process of vertical integration genuinely reduced friction in manufacturing processes: transaction costs were lower as a result. And finally, consumer choice was unheard-of; any colour you like, as long as it’s black.

That was a long time ago. The bulk of humanity has moved out of the agricultural sector, leaving only the mega-corporation and the boutique, the military-industrial agronomist and the guerrilla smallholder. Soon this will be true of the industrial sector as well: the low-touch cost-leading hyper-global, and the high-touch quality-leading hyper-local will soon be all that remain.

For some time now,  the services sector has been where the action is. We’ve experienced the Information Age for over half a century; the knowledge worker has replaced her predecessors. Capital has been mobile for some time now, and labour increasingly so. A different world.

And now of course we have the internet. And the web. With their concomitant flattening and democratising. Where everyone’s an actor, with the power to write, to edit, to delete, to publish, the tools of trade of the knowledge worker. A worker who now has the power to access, edit, alter and disseminate information at high speed and low cost to all and sundry, in the language of their choice, to the device of their choice, when and where the recipient wants. A world where the distinction between information consumer and information producer no longer exists.

Which is where the problem lies. We’re going through a process of horizontalising of everything, of “small pieces loosely joined”, of “high cohesion and loose coupling”. Platforms are now no longer hierarchical, they’re closer to being independent layers, often of different sizes and shapes. Almost counterintuitively, the glue that reduces friction between layers is the API, that which allows the small pieces to be loosely joined, that which ensures that the small pieces are loosely coupled. It is this loose coupling, this high cohesion, that allows for the flexibility that underpins adaptive systems, and makes that which appears to be complex simpler.

In the Information Age, these small pieces start acquiring new roles; some consume information; some produce it; some do both; some make lists of information, index them, tell people where to find information; some translate information, change format, change language, render for different devices; some make the devices that support all this; some make the software that empowers everyone to do the consumption, editing, publishing; some transport the bits; some manage access and use rights.

All this happens in a digital world, where reproduction and transmission of information is becoming cheaper by the minute.

Attempts to implement end-to-end control in such environments are doomed to fail; in essence there is no point in attempting to tighten what is designed as loose coupling, it doesn’t work. Which is why platforms leak.

You can reduce leakage, by concentrating on keeping the information cohesive within the layer, and making sure the APIs work securely. You can reduce it further by increasing inspection and usage, by ensuring that the APIs are open. You can reduce it even further by testing each layer against as many conditions as possible, by avoiding lock-in between layers, by making sure the pieces remain small.

The horizontalising nature of the internet and the web, of the digital age, needs to be understood. Layers must be independent of each other. Where they are not, the joins will come apart. And leakage will happen.

Of course, given what happened with Wikileaks, given what happened with Egypt, there will always be attempts to recreate vertically integrated control.

And more leakage will happen. Because the internet, and the web, route around obstacles. By design.

[Note: I will attribute the photographs in this post in half a day or so, I closed the tabs by mistake; my apologies to the people whose works I have used, I will rectify as soon as possible].

11 thoughts on “Why platforms leak”

  1. Interesting. Then horizontal knowledge work is inherently leaky. Yet security in IT systems such as ERP was designed for yesterdays hierarchical vertical work.

    Now clear why the security models are at cross purposes with work.

  2. I’m thinking about it this way: what is the shortest vertical integration we need to succeed. Just as when we take half the clothes and double the money when we travel, so we might need half the organization and double the agility.

    I think a mental thing takes over like we sees with fans of a sports team who are so desperately thinking about the final score that they don’t actually watch the game let alone enjoy its ups and downs.

    Or to go back to traveling – the entire point of the trip becomes showing the pictures on our return. The trip has no point in and of itself.

    Well there is a market for score obsessed sports fans and travelers who don’t enjoy traveling. -and there is a market for people who want vertical integration!

    And opportunities for people who enjoy the engagement with their economy, polity and culture.

    Or something like that!

  3. JP:

    Thought-provoking as usual.

    Since you mentioned the exodus from agrarian models of economy, I couldn’t however help wondering if such “horizontalisation” will ever solve the food security problem that is rearing its head every where in the world.

    Also for industries, that deal in physical goods, information-as-metaphor has its limitations. And for better or worse, much of global trade is still dependent on those physical goods (much ink spilt over the classification between manufacturing and services already so I wouldn’t attempt a repeat). I would be keen to hear your views on how the more “basic” problems in the world could be addressed meaningfully and sustainably using horizontal platform thinking. Thanks.

  4. Great post. The “horizontal” business model as the new optima in the information age is at the heart of our investment thesis here at anthemis but of course you know that… ;)

  5. And we can’t forget that this knowledge available to all is really only available to the 3 out of 10 people who actually have internet access.

Let me know what you think

This site uses Akismet to reduce spam. Learn how your comment data is processed.