I thought the best way to open a series of posts about platforms is to actually talk about what is a platform. Software platforms have been around for a long time and many of the same principles and dynamics apply across the changing times and technologies.
I want to start off by recommending some homework. I found a book early in my time at Twitter, called Invisible Engines, that helped me greatly in understanding the dynamics and economics of platforms. A lot of how I think about platforms was shaped by that book and you’ll see many of the concepts repeated here. While it looks long, you can focus on a few key chapters to make it an easy read. You can download it as a free eBook from the MIT Press.
Now, let’s go back to June of 2007 — the iPhone has just launched. The only applications allowed on the phone were the ones that Apple built and pre-installed. No Facebook, Pandora, Twitter, Evernote, Uber, Angry Birds, Netflix, Square, and so on. Each one of those applications went on to thrust the iPhone into huge new markets. Square suddenly turned your phone into a money making device for small business owners. Angry Birds made the iPhone an indispensable gaming device. Netflix let you stream movies and TV shows right to your device. Facebook and Twitter turned every phone into something that let you share the world around you — from your child’s birth to a plane landing on the Hudson. The iPhone would not be the monumental success it is today if Steve’s original vision of Apple building every app had played out. Each of those apps, and countless others, have created immense value for Apple that they never would have realized on their own. By becoming a platform, they enabled developers to build applications that would make their device more valuable to users, thus selling more devices. As more devices were sold it created more revenue potential for app developers, thus drawing more developers to iOS. This created a very powerful network effect that drove growth of both sides of the business (developers and users) where growth of one side directly benefited the other.
“Most successful software platforms have exploited network effects between applications and users: more applications attract more users, and more users attract more applications.” — Invisible Engines
Platforms, as we’ll reference here, are software-driven, multisided businesses or marketplaces. A lot of the same dynamics can be applied to more general marketplaces like eBay or TaskRabbit, but for the purposes of these posts we’ll focus on software platforms. iOS, Android, Microsoft, Facebook, and Twitter serve at least two sides of the business and sometimes more (partners, advertisers). While all platforms theoretically have network effects, some have stronger network effects than others. This is something that changes over time depending on what services the platform provides and what policies are in place to govern behaviors.
An example of a platform with strong network effects is iOS. Apple provides great discovery of applications to their users through the App Store. I think they do the best job in the industry of merchandizing valuable applications to users and providing a fairly frictionless process for paying for and downloading an application onto a user’s phone. This in turn delivers great distribution and revenue opportunities for their developers, making users happy as they get more value out of the iOS device through new applications and the application developers become more successful through the distribution and revenue opportunities.
An example of a platform with weak network effects was the Twitter Platform in it’s early days. At that time, the ecosystem was made up of some great applications, but the platform didn’t directly provide distribution or revenue opportunities. For the most part developers had to do their own marketing to get users and had to provide their own means of capturing revenue. For example, CoTweet, one of the earliest applications on the platform, built their entire business around Twitter data, yet had to do their own marketing to find customers and had to setup their own billing in order to capture revenue. This creates a lot of friction for app developers and thus loses some of the value of those network effects. As Twitter has brought the platform into the core product (Twitter Cards) they have been able to deliver better distribution opportunities for developers while also increasing utility to the user experience.
The lower the friction to connecting the various sides of the business, the greater the network effects and the greater the value to the platform provider.
An API does not a Platform Make
One final note that I think is worth throwing out there for debate is what isn’t a “platform”. A lot of people talk about Amazon Web Services, Urban Airship, Twilio and other similar services as platforms. However, I think of them more as software as a service (SaaS) instead of a platform, since they aren’t multisided businesses and only provide services to developers. These can still be incredibly valuable businesses, but they don’t get the benefit of the network effects that come when you connect multiple sides of the business together. On the flip side of that, just because your business is a platform, doesn’t mean it is provided with some magic dust to make it successful. In fact, trying to build up two sides of a business can be much more difficult than just signing up developers to use and pay for your service.
Ryan Sarver 
Pingback: Things I Learned Working on the Twitter Platform | Ryan Sarver
Whenever you’re building a multi-sided business model – a marketplace, a brokerage service, etc. – there is always a chicken and egg problem at the beginning. There is the illusion of “if you build it, they both will come” but that is never the case. The most successful marketplaces are those where, as you’ve alluded to, the platform subsidizes outsize value for one side as a means of attraction. Classic example being Google: by subsidizing the costs of indexing the world’s information, they attract massive search user inventory and can build AdWords as a means of brokering that intent to the highest bidder. If Google had say charged users for search (a valid suggestion given the value and costs), they would have never grown to where they are today.
As for Twitter, I agree with your assessment of the early days. Without the ability to harness the growing inventory via analytics and provide embedded transaction services, Twitter was/is a lossy platform – lots of uncaptured value. Thankfully for them, this is changing quickly and purposefully.
Looking forward to your next post.
Pingback: Quickthink » Blog Archive » More on Steve Blank “Blowing Up the Architecture”
WordPress ate my comment, so re-posting. Sorry if the previous one shows-up.
Looking forward to that series, Ryan! Couple of questions:
1. When you say that the platform needs to have network effects, are you talking about Twitter users or developers on the Twitter platform?
2. Would you say that the Twitter platform is a pretty controlled one, by design? Why are there so few monetization opportunities on top of it?
Glad you like the first one. Many more to come 🙂
1. Correct. The network effect is between the users and the developers. As more users join the service, the more desirable it is to developers. Then, the more applications that are built and provide more functionality, the more users are interested in the platform.
2. It depends on what you mean by “controlled”. I would actually say there are a number of monetization opportunities on the platform, but payments aren’t something the platform provides natively, which might be what you mean. Providing payment infrastructure is incredibly complicated and risky, but as you probably saw, Twitter recently hired Nathan Hubbard away from Ticketmaster, where he was CEO, to be VP of Commerce. So my guess would be to stay tuned.
Re #2, I meant that you had to tighten the screws a bit on the API last year in August, so you could better monetize the Twitter streams. But, aside from Advertising via the API and selling the firehose, I’m trying to think what other monetizable apps are there now, on top of Twitter?
I didn’t mean to imply Payments, although that might be an interesting move with a Bitcoin plan maybe 🙂
I’m hoping to learn from others’ mistakes in this community as I embark upon a phased approach to building a platform for a 3-sided marketplace; renting fitness/yoga studio space by the hour for live classes, where the interests of space owners, instructors, and students intersect.
I call it a platform because it’s designed to allow any of these three groups to create live classes anytime, anywhere: allowing existing studio owners to monetize unused space, convert to a better business model or open new locations. It allows instructors to independently host classes without having to audition for a job or open a studio. Finally, students can practice when and where they want, tapping into a directory of spaces and instructors, designing their own schedules without having to rush across town to make it to class on time.
The student value prop is the mass consumer product that emerges at scale on top of the platform, but requires the foundation of what looks a web-based Saas product for studios and teachers to schedule, promote, and process payments for space rental and class promotion & tickets.
By necessity I am approaching this in phases, first targeting independent studio owners with a product that provides cash flow where they otherwise didn’t have any, renting to teachers on their existing roster or elsewhere who are eager for more classes… This approach provides value for early users regardless of whether or not anyone else is using it, and allows us to start to build critical mass to achieve network effects that can lead us on a path to becoming a true platform.
I’m currently in the weeds of building the web based saas for studios & teachers, and as a first time entrepreneur, this approach makes sense to me, but I know the collective experience and wisdom here is far greater than my own. A bit of open kimono to the public here, but I’m familiar enough with the complexity of execution that I’m confident there’s more value in community feedback than there is the risk of competition at this point. I thank you sincerely for any guidance.
When you say that Amazon Web Services is not a platform, are you talking the AWS Marketplace into account? It seems to fit your definition: server application developers sell a product (their app packaged for an EC2 instance) to other users; Amazon manages the marketplace and handles billing and payments too.
Benjamin, I hadn’t been and was unaware of that marketplace. Is that just for other developers to get EC2 instances off the shelf or is it marketing the end services to end users?
We can all name platforms that started with the users first (of the chicken/egg issue), which platforms that you know of started with the developers? Aren’t all successful platforms created based on their ability to get the user side first?
I can’t think of any examples off the top of my head. I think that’s because user adoption is the hardest part of it and therefore the critical path to building up one side of the business. Developers go where the users are.
XBOX comes to mind.
Great post Ryan. Interesting insight from you re iPhone as platform : “The iPhone would not be the monumental success it is today if Steve’s original vision of Apple building every app had played out. Each of those apps, and countless others, have created immense value for Apple that they never would have realized on their own. By becoming a platform, they enabled developers to build applications that would make their device more valuable to users, thus selling more devices. “
Aside from Twitter, Facebook, and Apple…what are the true platforms that exist in the consumer world (I would add Salesforce, but that is clearly in B2B)? What scale do you have to get to to be considered a real platform? Curious if you think a dominant “platform” will emerge in every vertical?
Looking forward to your next posts in this series.
You list what I consider the dominant platforms for consumers, save that you left off Microsoft. Amazon’s consumer services could also be considered. They are working on building up an ecosystem around Kindle, but it’s still very early.
As for scale, I’m not sure that is a requirement of the definition of platform. It’s more about the mechanics of how it’s setup. Scale leads to relevancy in the users and developers mind though.
Would i be correct in understanding that the implication here is that a platform is not just having the API. A platform, as per this post, is when it allows one to build applications and get users for it. Since you called Twitter’s platform a platform even without the mechanism that helps developers connect to users i guess, a platform only gets better if it provides a good mechanism to developers to market their wares and perhaps collect revenue. So why isn’t a set of APIs that lets one develop a SaaS app and market it either via a central, optimized proprietary marketplace OR via own efforts. I couldn’t figure out the reason for bashing APIs and letting Twitter (without the twitter cards) be called a platform. Are we complicating stuff here when there is no such ambiguity or debating an ambiguity that ain’t worth any $ unravelling.
There is no bashing of APIs. All of those API businesses that I list are great businesses and services. I’m just trying to clarify what I consider to be the differences.
Can you also please give couple of examples for Platform as a service (PaaS)? Is it right thinking that Apple, Facebook etc provides Platform as a Service to developers?
Platforms emerge and they primarily build user side first (I agree with @webjoe’s comment). Once you have one side customers (users or developers) then you can open up yourself on multiple sides to evolve as Platform.
So may be some SaaS player today can become platform of future.
Great post Ryan. I work at a company where everything is a platform. 🙂 There’s an identity platform, a notification platform, a payments platform, a wallet platform, a demand-gen platform, a developer platform, a reporting platform, …. – you get the picture. 😉 I think what you said about multi-sided business is right on. Payments as a Service can be done by anyone in many different ways – but being a payments platform to developers while being a wallet provider for consumers is a different game all together and there is a lot to do for it to become a true platform than simply providing the APIs!
Not much depth in the definition of what is a platform? I was expecting this to be on similar lines as Steve Yegge’s post on platforms.
There are so many definitions of the term ‘platform’ these days, but yours I like quite a bit. No, APIs platforms. Still, (relatively) open APIs, SDKs, and ecosystems tend to result in network effects and pretty vibrant platforms like Amazon, Apple, Facebook, and Google. It’s a a big emphasis of The Age of the Platform.
Great post, looking forward to the series. I’m most interested in how you define your platform so that you get the right kind of behavior on it — said another way, how you decide what parts of your ecosystem to expose, and what parts you keep internal-only.
Fascinating thread already, thank you. So, a platform is not so much a product of scale or scope, as it is the presence of the ying and yang of an offering — the ability to balance both sides of the user/developer weighing scale. Adding capability, content or value to one side, empowers, interests and cements the commitment of the other.
At Twitter, you obviously had the scales weighed heavy with users, so it was your job to counter that weight with valuable “somethings” that would start users looking AT twitter, rather than simply using twitter and moving on.
Would that be a reasonable definition?
Thanks, and please keep these posts coming!
We see things the same way. You would like my 2011 book on the subject.