On platforms and sharing


I signed off last year with a couple of posts asserting that 2013 was going to be The Year of the Platform. So I thought it behooves me to get this year started with the same message.

When I speak of platforms, I tend to speak of open adaptive constructs that allow people to layer value around and “on top of” the foundation. As stated before, platforms come in many forms, shapes and sizes. A closed platform is likely to evince linear growth at best, as postulated in Sarnoff’s Law. More open platforms are capable of supporting geometric or exponential growth, initially quadratic, as in Metcalfe’s Law, and then tending towards exponent values greater than 2, as in Reed’s Law. The principle is the same: the more open the platform is, the more likely it is that relationships and interactions between platform participants will create new and differentiated value.

I am not an expert on platforms; at best I’m a passionate amateur. If you want to dig deep into open adaptive platforms you should start with the internet and with the Web, what Tim O’Reilly called The Architecture of Participation. So you should delve into the writings of people like Vint Cerf and Tim Berners-Lee; understand the thinking of people like George Gilder, Bob Metcalfe, David Reed, Andrew Odlyzko; get into the books of people like Richard Schmalensee and David Evans. Maybe even read the comments to this post, there may be some good pointers there.

Enough scene-setting. Here are my initial thoughts on platforms and sharing. Have at them, let me have your comments and views and criticisms, let’s all learn in the process.


Platforms create value by enabling social interactions between participants

The simplest way that platforms enable sharing is by helping people communicate with each other. The traditional post and telegraph companies, forerunners to the modern telcos, are classic communications platforms. All you need is a directory, some way to look up who’s there. If possible, you can reduce the cost of searching through the directory by grouping/sorting, in alphabetical, regional, functional, whatever. When directories were analog this was very important, but now with digital search capabilities this is less so. Nevertheless, the ability to form groups and to have those groups discoverable is important.

Telco 1.0 was about enabling primarily voice-based means of communications (though telegraph played a part for a while). Directory updates were analog and annual, grouping capabilities only available to the directory publisher. Telco 2.0, led by Microsoft, added e-mail to the functionality, and with that form of messaging allowed the scheduling of events to become reality. Telco 3.0, led by Facebook, made all this real-time, available as a feed, and with exposure of APIs to make it a real platform.

Once you have a platform that allows simple social interaction, the next simple thing is to enable sharing of digital things. Photographs. Documents. Music. Film. Blog posts. Smaller blog posts. Micro blog posts. No surprise why we all spent time going through all this, with the consequent impact on traditional analog intellectual property structures.

The value was not just in the sharing of the digital object per se, but in making that object “social”, as people like Jyri Engestrom reminded us. By allowing people to participate, new forms of value emerged. Everything could be tagged: folksonomies emerged, new ontologies and taxonomies could be formed, more flexible and more adaptive than their predecessors. Discovery of the object became easier.

It went beyond the tagging and labelling. Everything could be rated or reviewed. In fact even the rater or reviewer could be rated or reviewed.  So the next stage of value came from participant ratings (as in eBay) reviews (as in Amazon) comments (as in blog posts).

There was an obvious next step: lists. Top 5s, top 10s, favourites, whatever. So that became the next basis for adding social value.

And then of course you needed a way to share the tags, the comments, the reviews, the ratings, the lists, with a larger group of people. So we saw more ways of publishing come to the fore, with Tumblr and Twitter. But unlike the past, the growth in participants did not need to arrive at spam. Because these new publish mechanisms came with new subscribe mechanisms. Pub-sub was here to stay.

All that was phase 1: the sharing of digital objects, the facility for participants to enhance those objects with comments and feedback, and the capacity to publish or distribute the enhanced objects.

From that point platforms have continued to evolve in at least two ways: allowing people to build platforms themselves (by exposing infrastructure and software and data as services); moving to engaging with analog objects, not just digital (where now only the icons and labels of the shareable things were shared within the electronic community). Sharing physical inventory became valuable, with recent examples like Mealku and FlightCar as part of a category that includes kiva, airbnb and for that matter Mechanical Turk; they’re all exchanges where one person’s excess meets another’s scarcity, for a small fee. Sharing applications built on top of the platform is another step, which means the platform needs to offer APIs and some form of app store. WordPress, Discogs, Etsy, Minecraft, these are all examples of communities that can be built with APIs and apps.


Sharing also creates value by reducing waste: the growth of the 21st century scavenger

I want to emphasise this by sharing some recent examples that came across my radar screen, examples that excited me a lot : Mealku, which came to me via Jerry Michalski, and FlightCar, which came to me, I think, via Pat Phelan. I think of the two as identical. With Mealku, you get a chance to do what the New York Times called “making reservations for leftovers”: you have an excess of food prepared at home, and the platform allows you to get rid of that excess meaningfully. With FlightCar, you lend your car to others while it would normally have been parked at the airport waiting for your return, and they return it in time for you. I think Doc Searls will love it, it is so VRM. [Incidentally, I mentioned Jerry and Pat for a reason: they’re perfect examples of how “social” improves filtering. They knew me, knew what I would be interested in, and were connected to me in ways that allowed them to share the information simply and meaningfully. Pat lives in a different country, hundreds of miles from me, but in the same timezone; Jerry lives eight timezones away. I’ve known both for years. We see each other reasonably often, but our primary interactions tend to be digital.]

Resource scarcity is a reality for us now. We’re already facing challenges with energy, water and minerals, and there’s more to come. Reducing waste is no longer a nice-to-have option. So services that allow us to match one person’s excess with another’s shortage are going to become more and more important. When I was a child in India I remember being told that every day, the USA throws out enough food to feed Canada. I have no idea whether that’s true, I haven’t been able to verify it. But anecdotally it appears to make sense.

There’s a lot of research that’s already been done on how scavengers and rag-and-bone men and their ilk were part and parcel of city ecosystems in the past. Those were all about physical cities, bounded by physical spaces. Now those constraints have gone, with a new dimension of freedom added: the concept of renting scavengeable inventory. Borrow my car while it’s in the car park. Eat my leftovers. It’s an exciting area, disrupting many markets, and allowing for new forms of intermediary to form and thrive.


All this sharing creates Big, Small and Open Data

Big Data is by itself nothing new. We’ve had distributed devices for decades, collecting transactional information, and allowing for pattern analysis through aggregation and visualisation. The credit card is a classic example.

As we moved from small numbers of dumb devices to somewhat larger numbers of smart devices, the changes that took place were at least twofold. One, we moved from transactions to conversations and activity streams to intention signals: the “tense” of the activity stream moved from past to present to future-as-well-as-past-and-present. And two, this information came wrapped in metadata about time and place and actor and anything else the sensor could “lay hands on”: altitude, temperature, rainfall, friends present, devices used, ambient conditions, whatever.

There was a third change, a material change. We started involving people by connecting them. A whole new world, no longer machine to machine. Lots of data. Lots and lots of data. You’ve heard the similes. They go something like “In year X the world had a total of Y data. Today we add that in two days.”

What’s important is that we understand that there are at least three types of data from a platform perspective. Aggregated forms that allow us to act as groups or be understood as acting in groups. Drilled down forms that allow us to act as individuals, or to be understood as acting individually. And conventions, definitions, labels, part numbers, classifications that allow us to have meaningful aggregations or drilldowns.

Both Big Data as well as Small Data come in open and closed forms.

We understand the value of the Big and the Small, but we have only just begun to scratch the surface of the Open. [And that’s something I will dwell on in later posts].


Machines filter. It takes a human to curate

By now most of you have probably come across the kind of problems of an algorithm-driven world. If you haven’t done so already, check out Kevin Slavin’s TED talk on the subject. Similarly, you’ve probably understood most of what is there to understand about the benefits and risks of filter bubbles. If you haven’t done so already, you should hear Eli Pariser’s TED talk on the subject. And speaking of TED talks, I guess my whole philosophy about all this can be understood in what Sugata Mitra had to say, also in a TED talk, this time on education.

Mitra said “A teacher who can be replaced by a computer should be replaced by a computer”.

So it is with all of us in all our walks of life, as knowledge workers or traffic wardens or surgeons or rocket scientists. There are things that we do that can be replaced by computers. And they should.

And then there are things that can’t be replaced by computers. Those are the things we should concentrate on.

Technology is best when serving us, when our actions are augmented by the tools.


The enterprise context: everything starts with the customer

If you’re a regular reader of this blog then you probably know I spend a lot of time thinking and writing about the enterprise, and about how information lives within and beyond the enterprise.

Open adaptive platforms, underpinned by robust social networks, aided and abetted by strong analytical capabilities, are here to stay. They’re some of the reasons I joined salesforce.com in 2010, convinced by the importance of the multitenant model, the platform and Chatter.

The traditional lock-ins for customers have disappeared in most markets, making the loyal customer a scarcity. Understanding customer engagement has become an imperative. Not just how the company engages with the customer. But, more importantly, how the customer engages with the company or companies. So it is no surprise that the focus on investment has moved from traditional back-off systems to “front-office”.

But there is a subtler change. Customers are also investing in “systems”, platforms that simplify their engagement with the business world at large. How they identify themselves (login, Facebook, Twitter, LinkedIn). How they share their personal information. Who their friends and advisors and recommenders are. What devices they’d like to use. How they’d like to pay. Their preferences in terms of devices, approach to contracting (no contract versus term, prepaid, pay monthly, the lot), approach to payment (card, PayPal, Square, Amazon Payments, etc).

Those “systems” are often based on services provided by platforms, and companies must now build services that take those systems and services into account.

When you start thinking from the customer perspective, it all starts with C2B. It may look like C2B but often it turns out to be C2B2B2B2B. Soon it may start looking like C2C2C2B2B2B2B.

Platforms. Open, adaptive.

2013. The Year of the Platform.



12 thoughts on “On platforms and sharing”

  1. Lovely post JP – this will enlighten and contextualise for many the environment we live in in 2013.

    Resource scarcity – time. I firmly believe that for many people the management of their time for themselves and indeed for their self is paramount over and above all other considerations. This has come about, continues to come about through all the myriad emergent behaviours you list above.

  2. Interesting & thought provoking examples of practical relevance. Will we ever in reality move towards implementing them – am reminded of the important pillar of “trust” you oft talk about. Am sure that will be required as a building block for the platform.
    Will then the “the year of platform” in affect create the “base of trust” for interactions between B2B2B2C2C2C2….

  3. Well worth posting and reading JP, enjoyed it. But what is the economic ‘worth’ of the platform? Started in Telco1 with subscriptions and user fees, Telco 2 similar, Telco 3 initially as sheer hope funded by VCs and investors, moving towards advertising revenues, but is there a limit or constraints that the harvesting of economic worth places on the platfrom? Also, worth considering other competing standards and platforms. Have you read shapiro, ‘The Art of Standards Wars’? http://faculty.haas.berkeley.edu/shapiro/wars.pdf

  4. Entering the Wall Street subway station in NYC, thinking about JP’s platform posts, I was surprised how often “platforms” appears on the signage. But then the penny dropped, subway resonates with Tim O”Reilly’s architecture of participation.

    The platform I really grok is GitHub which made Linus Torvald’s Git code sharing repo site friendly to use for social coding. http://www.wired.com/wiredenterprise/2012/05/torvalds_github/

    Got me thinking why do some NoSQL databases not really take off until someone makes the core technology friendly to use. Facebook’s user experience built on top of MySQL and other NoSQL databases being a case study.

    Thus I’d posit many technologies may be poised to turn into platforms if they are made very friendly to use. The pattern at Heroku.

  5. Yes, I have written many times about social objects, just use the search box if you want the references. I make the reference to Jyri Engestrom, who got that ball rolling, and then Hugh picked it up. In other posts I refer to Hugh regularly/

Let me know what you think

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