The platform economy is an API economy. And artificial intelligence needs APIs

0

A spectre is haunting economy –  the spectre of digitization. All the powers of old economy have entered into a holy alliance to exercise this spectre. But other than the communism (obviously the first two sentences are stolen from the Communist Manifesto by Marx and Engels from 1848) – this spectre is about to stay, to grow, and finally to take over huge parts of the whole economy. Accepting this as a hypothesis the most famous examples cited often include these huge tech giants Facebook, Google, Alibaba, Uber, AirBnB, Tesla and Netflix. What do those very different companies have in common? They created, maintain, and grow platforms. What is the techncial idea behind: to connect supply side with demand side via coded applications based on standards set by the platform owner. In other words: APIs. Lets have a deeper look into this. 

Why platforms win the game

It never changed: competitive advantage by real differentiation (better product, better service, lower prices) will lead to increased market shares. When Ford leveraged the slaughterhouse idea of highly automatized production in form of assembly lines they could deliver a product to a much better price to a totally new customer segment.

But, after a short period of time the competition learned the same, developed something similar and introduced more or less the same product to a similar price. Sorry to say, but generally speaking there is no really important difference between a Mercedes-Benz, an Audi or a BMW. The differentiation comes not from the product itself but from marketing aspects: brand positioning, market presence, distribution power, etc. Market leaders in those markets own 10 %, 20%, maybe 30 %. Toyota, General Motors, Volkswagen and all the others have their place in the market.

Having the great idea to build a platform to or connect, lets say “book producers” and “book buyers” – something totally new rises on the horizon. Amazon is not a retailer for books, its really about this platform idea. They created the perfect experience for all book buyers by allowing and promoting reviews, automated recommendations, calculated satisfaction measurements for suppliers, most comfortable customer services etc. So, every buyer of a book (or any other product at Amazon or Alibaba) will regulary return to this. This attracts suppliers, special offers, promotions – and competition on the supplier side, but everything on the same platform! So continuously the customer experience will be improved.

With this mechanism in place platforms will grow almost endlessly – or until they own the market. You can see this today in existing platforms like Amazon or Alibaba. They own 50%, 60%, 80% of the market in different categories in the markets they are active.  An instructive overview for online retailers in Western Europe can be found here. In 2015 the biggest 19 ecommerce retailers together had an revenue of 43,1 bill. EUR, none of them bigger 6,4 bill. EUR – whereas the number 1 alone had 38,5 bill EUR. I bet, that in 2016 figures Amazon will have beaten the 19 combined ones.

What platforms do better

They make use of data. As simple as it as: they collect data, analyze data, gain insights, and take decisions with one goal in mind: a better customer experience. Once a company started the journey of digitization their main aspect of their big data labs or incubators should be a better customer experience. This might have the consequence of making the production lines more intelligent or a predictive maintenance strategy or a better fraud detection in online transactions.

I attended a lot of presentations in a German working group of leading vendors and service providers around Big Data. 90 % of all use cases cited where around internal processes. Car producers presented their stories around egine development. No doubt, it is fascinating when one is able to collect massive amounts of test data in motor development much faster than before because of a new in-memory database or something else. Sometimes the engineers even reported that they found something interesting in the data. But rarely I heard of concrete decisions taken because of this data. And if there where decisions they developed at the end the same engine – a little bit fast, a little bit different. Incremental improvements.

But when Amazon collects all their data they have, they listen to the customers voice and immediately take decisions out this. Where to place which promo, whom to deliver which next best offer, how to recognize a customer in danger of churn – and how to respond to him. If you have a “machine” like a well-established platform in place, you can start experimenting. How do customers react on this pricing strategy, this discount, this offer – every click gives additional insights. And how does these machines do this? They learn patterns.

That’s why platforms tend to be learning machines.

Why platforms use APIs

Application programming interfaces, APIs, are technology that allows software programs to talk and interact with one another. This simple idea helps to combine very different skills in one customer centric application. Take the example of Uber as a platform. They create a fantastic customer experience.

When I visited New York City over Christmas last year for the first time, I could give it a try – in Germany, where I live, a holy alliance exorcised this spectre for the time being. I wanted to go to Whitney Museum, no glue how far it was and how much a taxi would cost – and always in doubt whether the taxi driver would try to benefit from my inexperience as a first time tourist in NYC. So I opened the Uber app. Google Maps knew where I was, asked me where to go, and offered me a price. I accepted, was informed about the drivers car, where he was, and how he looked like. And I knew: he accepted the price as well! And he would not have to fear not be payed – since Uber does this based on my reputation (= credit card) at Uber.

So Uber used the Google API to create an app on this. This increased the usage of Googe Maps and provides Google with additional insights. They can use this to sell location based advertising on the map or afterwards in their search application (e.g. when Whitney is closed on Tuesdays and I was looking for an alternative around the corner. Which, by the way, finally ended up in a nice walk in Highline Park).

Where is the link between APIs and Artificial Intelligence

Digitization creates a lot of data – which will end in platforms. Did you read the announcements about IoT platforms in the last months? There will be many platforms. And they all need to analyze the data which hopefully will flowed these platforms, these learning machines. The platform owner will have to leverage artificial intelligence technologies like machine learning algorithms, pattern recognition, decision trees, artificial neural networks, deep learning etc. to make use of this data. Wouldn´t it be great to have an analytics engine available via APIs to which you could talk to in several programming languages to be able to build your own applications? Wouldn´t this allow developers to make use of all those advantages (a fully functional map, like Google Maps) without having to code it from scratch? Would this perhaps be a sufficient explanation why SAS launched SAS Viya and makes it available via RESTful APIs? Would love to hear your thoughts!

Where do you see the impact of APIs heading?

There can be no doubt APIs are playing a more significant role in how we provision and consume intelligence. But how far, and how fast can we really go?

Suggested articles to read:
How to add analytics to your application development pipeline
Could APIs provide advanced analytics for the masses?

Download a complimentary White Paper: Understanding data streams in IoT

Share

About Author

Thomas Keil

Director Marketing

Dr. Thomas Keil is a specialist for the impact of technology on business models and on society in general. He covers topics like Digital transformation, Big Data, Artificial Intelligence & Ethics. Besides his work as Regional Marketing Director at SAS in Germany, Austria and Switzerland he regularly is invited to conferences, workshops and seminars. He serves as advisor to B2B marketing magazines and in program committees of AI-related conferences. Dr. Thomas Keil 2011 came to SAS. Previously, he worked for eight years for the software vendor zetVisions, most recently as Head of Marketing and Head of Channel Sales. Dr. Thomas Keil beschäftigt sich mit den Folgen des technologischen Wandels für Geschäftsmodelle und für gesellschaftliche Veränderungen. Dabei geht es ihm um Themen wie Digitale Transformation, Big Data, Künstliche Intellligenz und ethische Fragestellungen. Neben seiner Arbeit als Regional Marketing Director für SAS in Deutschland, Österreich und der Schweiz ist er regelmäßiger Gast auf Konferenzen, Workshops und Seminaren. Er ist Gutachter im Bereich Fachpublikationen im B2B-Marketing und agiert als Programm-Beirat für Konferenzen in seinem thematischen Umfeld. Dr. Thomas Keil kam 2011 zu SAS. Davor war er acht Jahre für den Softwarehersteller zetVisions tätig, zuletzt als Head of Marketing sowie Head of Channel Sales.

Leave A Reply

Back to Top