The BL Interview. Data is the hero of platform model: Sangeet Paul Choudary

It’s the age of the ‘platform’ business model, which creates value by facilitating exchanges between consumers and producers. Facebook, Airbnb, Uber, Apple are all successful examples of businesses founded on platform thinking. This is radically different from the old ‘pipes approach’ — the traditional linear model where value creation happens inside a firm through employees. Sangeet Paul Choudary , Founder and CEO of Platformation Labs and author of Platform Scale and Platform Revolution  explains to Chitra Narayanan the major shifts taking place in production now.

Has it become a business imperative for everyone — even non-internet companies — to switch to a platform model? Traditional companies like Kellogg seem to be in the midst of a platform driven digital transformation?

I don’t think everybody should switch to a platform model largely because it’s not possible for everyone to become an aggregator. However, it is important for everyone to understand platform business models for two specific reasons. First, whether you become a platform or not, you might have to participate on platform-enabled ecosystems so it becomes important for you to understand how they work, how the governance and policies work. For example, not everybody will be a Flipkart or Snapdeal, but they may sell through these platforms.

Second, while everyone may not become a platform, and this is especially true for large traditional pipe businesses, it’s important to understand concepts of how you open up for communities to participate; how do you create an open business model for the larger ecosystem of partners to come in, and how do you govern and regulate such open participation. 

As an example, even a mining company can apply theories of a platform model in the way it does business. Increasingly, mining equipment is connected to the cloud, and mining companies are operating through centralised intelligence. Thinking of mining as a data industry before it is a resource industry helps us identify new mining opportunities and greater collaborations with current and existing partners. So, while implementing the platform business model may not be a business imperative, a study of it should be an imperative from the perspective of participating in platform companies and from the perspective of leveraging open production and collaboration.

It seems a bit topsy-turvy that the aggregators — who are, after all, middlemen — gain more than producers and consumers in this… 

Yes, a lot of the value that is coming from the democratisation of production accrues to the aggregators. But these aggregators are not the traditional rent seekers; they create value for producers and consumers as well. Let’s take a few examples. If you look at what happened in the handset manufacturing industry earlier, when Nokia and BlackBerry would put out the handsets, they would seek applications from developers who would get only 30 per cent of the revenues that accrued. But when Apple came into the picture and opened up its App store as a platform, it could pass on 70 per cent of revenues to developers. That’s why the developer ecosystem coalesced around Apple and, subsequently, Android. What platforms do is that they are more scalable as intermediaries and gatekeepers and are able to redistribute value more efficiently back to producers and consumers.

If you take the Apple example again, it shows how more value also accrues to consumers in the platform model. In the past, Nokia and BlackBerry would launch handsets with set features that worked for a broad set of consumers. But with the launch of Apple’s App store, every individual could personalise his or her phone, and make it into a machine that fitted their needs. The platform model allowed Apple to give greater choice to consumers by letting them personalise the phone they wanted.

In fact, platforms that rely on rent-seeking instead of new value creation are less likely to encourage widespread participation from their ecosystem. Accordingly, platform governance plays a very important role, something that is missed by players who see this only as a technology play. 

Another way the platform model benefits producers is by allowing the creation of entirely new forms of production. If you think of Airbnb, it allows people to monetise a spare asset that they have — a spare room or apartment. Similarly, look at how independent media houses came up on YouTube, because the platform allowed them an easy access to market and handed out tools of creation. 

Finally, platforms also actually increase transparency in markets, because they aggregate social feedback. Think about what TripAdvisor does; now you get a clear indication of which hotels are good and which are not, based on actual inputs from travellers. 

So, in these four different ways platforms actually accrue much of the value back to producers and consumers. First, they redistribute value more efficiently; second, they allow consumers more choice; third, they allow producers entirely new opportunities because of access to a market and, fourth, they create transparency in these markets for consumers to make decisions.

You talk about the death of process and rise of interactions in the platform model. Isn’t that a dangerous thing as interactions can become uncontrollable?

Yes, in fact, sometimes the tyranny of uncontrollable interactions is worse than the tyranny of process because unlike processes you don’t own or direct interactions. When people start interacting in ways you never imagined, different types of issues come up. For instance, people have used Airbnb in new kinds of ways that flout laws. Then there are case studies of how terrorists have used platforms like eBay to communicate information between themselves. These interactions are way beyond what the platform owner would have anticipated. So it becomes important to keep monitoring actual data on how the platform is being used and create checks and balances. It not only helps you control interactions but it also helps you shape the future direction of platforms.

If the interactions are out of your control, how do you do quality control in a platform business model? 

In pipe business model there is centralised quality control. And a centralised management decides what is right and what not. In a platform business model quality control takes on entirely new dimensions. There are three or four specific ways in which quality control can be done on a platform model. For instance, how you choose to do quality control depends on the impact poor quality will have on your business. If interaction failure leads to high costs, you might choose to do it sooner rather than later. A wrong video on YouTube that does not appeal to a viewer may have less impact versus a healthcare aggregator that matches the wrong doctor with a patient. This is why in healthcare there is a lot of pre-curation of content. On media platforms, there is post-curation. The spread of content on a media platform is dependent on how the content is up-voted or down-voted. Media and healthcare are two extremes – in between there is banking and education. 

Increasingly, platforms are also scaling their quality control by looking at data and seeing how people are using the platform. For example, Airbnb when they started began with upfront quality control. They sent photographers to apartments and certified them by taking pictures and making sure it was the right place to stay. Then they scaled up by looking at data of how people were using these apartments. If a new host comes on to the platform and gets uncharacteristically high number of bookings and great ratings, it is possible that the host is gaming the system. 

Quality control also takes different dimensions in B2B platforms versus B2C platforms. For instance, on app stores, more apps are very good. But for B2B users, it’s better when there are fewer apps in a controlled environment. So the way SAP’s app store approaches quality control is different from Apple’s app store, as businesses only want to invest in few applications which are highly useful and highly resilient. 

How do you build a network effect that stays strong? What happens if your network migrates to competition?

Network effects are an important topic in platform models but they are often misunderstood. A lot of people think network effects are about more people using that utility. This is how telephones were used – more the people using the network more useful it was. That’s the fundamental model when you think of network effects. 

When it comes to platforms built on top of the Internet, however, it is not just about access to more and more people. It’s also about the fact that more people using the platform create more content and data which makes the platform more efficient as it gives greater choice to consumers. This data also allows the platform to become more intelligent and able to make better matches between producers and consumers. By leveraging the network effect, the platform can build unique advantages that a competitor cannot. The competitor will start with zero content, zero data. It’s because of these tools that value is being created on the platform. That’s what makes network effect defensible.

 So, data is the hero of the platform model?

Yes. Data helps to inform the platform about how the ecosystem participants should behave in future. So based on your behaviour in the past, your behaviour in the future can be guided and helped by platform. As an example, if you look at the newsfeed on Facebook, your user data actually helps it become more relevant and more targeted and directed. On the other hand, Amazon uses data from a lot of users to recommend new behaviour in other users. Amazon has a feature which says that customers who bought this also ended up buying that. So it uses data of other users to recommend what kind of activities you should participate in. These are all examples of new value creation using data. 

The third role data plays is it allows platform to determine who is a good producer and who is a good consumer. It allows platforms to assign a reputation to producers and consumers. The way reputation helps is that by asking you to rate the quality of producers it provides a new form of decision support to other consumers leveraging the data you provided. 

A lot of the platforms are about match-making between producers and consumers. If value creation is happening outside the platform, how does a platform owner monetise?

Charging a lead generation fee is one way. But a better way to do it would be if the platform figures out a way to create additional value so that it brings participants back on to it and retains the repeated value exchange.

For instance, if you were to create a platform that allows you to find a local plumber, once somebody finds a good plumber then they may not come back to you but will transact directly. This is less the case in an Uber or Airbnb where you care less about finding right people rather than using the utility. On some platforms, once a match has been made, there is no incentive to come back. 

But freelancer platforms, which faced the same issue, have solved this well. These platforms may match a designer to a client but they also provide many tools — such as time-management, monitoring of work etc, for the client to manage the freelancer on an ongoing basis. This ensures people come back repeatedly to the platform. So it boils down to how much of the value exchange do you own. 

COMMENT NOW