Platform Innovation // Blog

Interview: Sangeet Paul Choudary on How to Build a Successful Platform

Sangeet Paul Choudary is a Singapore-based platform expert who researches digital markets and platform economics and has extensive experience leading early stage ventures. I recently had a chance to ask Sangeet a few questions about platforms and why he thinks they’re having such a big impact today. Check out part two of the two-part Q&A below. You can find part one here.

1. What are the key metrics for platforms?

If I were to summarize my thinking on platforms in one line, it is simply this: We are in the business of enabling interactions. We are not in the business of creating goods or services. We’re also not in the business of building software. We are in the business of enabling interactions.

So what’s most important is that the key performance indicator should track a platform’s ability to enable interactions.

Let’s look at Airbnb. They don’t track the number of listings on the site or number of users visiting the site. What they track is number of nights that are booked. That is the real measure of the interaction between producer and consumer. And that’s really what matters for platforms. You need to ensure that you track not so much producer growth or consumer growth but interaction growth.

Even if you think of WhatsApp, the 450 million users that they have is an important measure of success. But the billions of messages that they send every month are what is most important. It’s the interaction that matters, because there’s no point in having a lot of users if they aren’t interacting.

Facebook also does this. It tracks the ratio of daily active usage to monthly active usage, which is again a measure of interactions – the more you log in, the more likely you are to be interacting.

In this case, the key metric Facebook pays attention to is not just how many people are on their platform, but what percentage of them are highly active.

So any proxy for interactions is really important. Now, interactions usually consist of not just isolated producer activity or consumer activity but activity happening between the two groups.

2. Which side of your ecosystem is more difficult to get on board initially: consumers or producers?

There are cases where you can have producers on board first. For example, OpenTable got restaurants on board first before it got the consumers on board. Then you have examples where you have the consumers on board first. If you think of Yelpthe consumers came before the merchants came. So, there’s no real rule over here. It really depends on the implementation that you have and the kind of market that you have.

Usually, it depends on your first iteration. Is your first iteration something that is targeted to the consumer or to the producer? Who is it offering value to first? The question that you really need to answer is which side you can bring on board and hold onto for sometime before you bring the other side. And that tells you whether it’s the consumer or the producer.

The question is really something you need to answer every time you build a platform. Who can I bring on and hold onto while I try to get the other guy?

3. What are some of the key strategies for attracting users to your platform in its early stages?

Well, let’s look at the producer-first strategy. If you want to get producers first, you need to give them production tools. Think of OpenTable. They gave the producers restaurant-management software. And once the restaurant started managing their inventory using OpenTable, the consumers were allowed to start booking that inventory.

Think also of Kickstarter. It’s a producer-focused tool. It allows producers to host a project and get funded. So, they approached producers first. They provided producer-focused tools as one of the ways to get producers on board.

Another other way of gaining producers is to enable them to market themselves. Kickstarter is an example but an even better example is Udemy, which allows a teacher to market himself and get a student base. So, that again is a case of producer first.

You get consumer first in certain cases, specifically if you have some content to kick-start the platform. That was the case with Yelp, where you had that initial directory of merchants that they could use to kick-start the platform. And once the consumers came on board, the actual producers also came on board.

Another method of getting consumers first is to fake it. A lot of dating sites fake it to attract one side. Reddit faked it. PayPal faked it to attract one side. So that’s another way of getting these guys on board. If you’re really successful, whatever you fake initially will not affect you in the longer run. But if you continue to be unsuccessful, MegaUpload being an example, you will have to kind of continue faking it in the long run.

4. How large does your ecosystem have to be before you reach critical mass?

Critical mass is not a number. Critical mass is actually a state that is achieved when there is a high degree of overlap between producers and consumers. You could have a million producers and consumers with very little overlap and you will not have critical mass. Or you could have just a hundred producers and consumers and because they have high overlap, you have reached critical mass. So critical mass is not a number.

Let me illustrate that with a very poignant example. If you think of Facebook, Facebook is a great example because it launched within Harvard University and that allowed for a huge overlap between users who are interested in each other. Essentially, if it had launched globally and gotten a million users around the world, it would still not have reached the critical mass.

What’s important is not a number but your ability to maximize overlap.

So instead of tracking a number, what I personally prefer tracking is an ability to see interactions happening. At critical mass, interactions will happen. Before critical mass, interactions won’t happen.

Suppose you’re building the next YouTube and you see people coming on and uploading videos to the site but none of those videos are being watched. What I want to see is what percentage of my videos are getting a minimum number of views because that is really what shows critical mass rather than a certain number of users.

5. What are some of the core technologies that successful platforms use?

I don’t want to get into the details of these specific technologies because platforms are really broad business models and depending on what you are building – the technologies vary.

What I want to mention is that from a technological perspective, there are a few things that are important.

  • You need to have really good data and you need to have a really good ability to process that data. That’s why the data scientist role is becoming so important today.
  • The next thing is intuitive interfaces. Platforms are self-serve systems. Users create value by themselves on platforms. This means that the platform needs to be extremely intuitive and low friction. Interfaces need to be really good.
  • As platforms are open and self-service, they can be gamed. You can have the wrong behavior happening on the platform, so you need to ensure that your platform has rules built into its technology that prevent this behavior. And that again happens through creating the right algorithms.

That’s really how I think about the technology architecture of a platform. Once you go beyond that, you get to the basics of exactly how you do it, what technologies you will leverage and all of that. But these are the top three things that I would think of.

6. What are some methods for testing and improving your platform on the technology side?

The most important thing is that you need to always track interactions, no matter what the stage of the platform. You need to track interactions or you need to track some proxy of that interaction.

An actual interaction may not happen. For example, when OpenTable was first launched, the consumers were not on board yet. But as long as restaurants were updating their inventory, that meant that interactions could happen in the future.

The most important thing of course, is to set your analytics and your dashboard in a way that it’s always focused on the interaction rather than tracking hundreds of random metrics.

The second important point is that platforms have network effects, so platforms are usually not single-user offerings. They are multi-user offerings. You just can’t go to one user and show him something that nobody else is using and say, ‘’Are you finding value in it?’’

The traditional methodologies of testing start failing because of this. You need to ensure that the network value is simulated before you start testing the platform with a particular user. And there are many different ways of doing it.

When Quora started, it simulated the community through internal editors. There have been cases where the platforms paid people on Amazon Mechanical Turk to start creating initial buzz on the platform so that they could test it with users. Whatever the methodology of creating that value, it’s important that you create that value before you test it.

Weekly Industry Newsletter

Top Posts

  • Platform Business Model – Definition | What is it? | Explanation

    Read more

  • Platform vs. Linear: Business Models 101

    Read more

  • White Paper: Development Platforms and the Next Wave of App Innovation

    Read more

  • Amazon Business: Understanding the Threat to B2B Distributors

    Read more

  • The Value of Digital Transformation: How Investors Evaluate “Tech”

    Read more