How small ideas like Twitter, WhatsApp and Airbnb create large value through network effects

| February 26, 2014 | 12 Comments

This is post #5 in the ongoing series on Platform Design, and is probably the most foundational post for all the work on this blog.  


Twitter: You can use this to type 140 characters

WhatsApp: Just like SMS but doesn’t cost much

Airbnb: Rent out your spare mattress

Uber: Just another taxi company to many

Over the last week, I’ve been hearing people in the startup world repeatedly scratching their heads on how a simple app like WhatsApp got bought for $19B. There is a lot of speculation in non-tech circles on why a software that does nothing but let you type 140 characters is valued at multiples of billions. What’s the big deal about Uber when it seems like just another taxi company?

The fact is that when it comes to platforms, the technology often isn’t the reason people use it at all. I don’t intend to discuss the above valuations and their legitimacy in this post. Instead, I’d like to talk about the real value that is created on platforms.

Most platforms look like stupid ideas when they have no users. 

When startups complain that simple apps like Whatsapp get bought, they completely miss the point. In fact, even valuation experts miss the point when they start valuing companies based on their user base and the rate of user growth. None of these things matter.

What really matters – the one metric that platforms should really be judged on- is the rate at which value is created on the platform. Airbnb is valuable when it powers a liquid marketplace, with millions of transactions. Twitter is valuable when tweets are created, retweeted and responded to.

The value of a platform is created by the interactions on top of the platform. 


The value is neither in the technology, nor is it in the rate of user growth. The value is in the activity that exists on the network of users. What matters about WhatsApp is not the fact that it is a ‘stupid messaging app’ nor just the fact that it has 450 M users, but the fact that this network of users creates tremendous value for each other through interactions; enough to hook 72% of its user base on a daily basis. (I was even called on primetime television to discuss this.)

But what does value really mean? And if it’s not about the technology, what is the platform’s role in enabling this value creation?

Over the last few days, I’ve tried on various occasions to explain WhatsApp, Twitter, Instagram, Uber and Airbnb to skeptics, most of whom aren’t directly involved with tech, and in the process have ended up using multiple metaphors to explain how value is created on platforms.

From Lego to Soccer and everything in between. If nothing else, the one point that we need to embrace as platform builders is this:

We’re not in the business of building software! 

We’re in the business of enabling interactions!


Lego: When Platforms become Child’s Play

At The Lego Movie a few days back, my mind again wandered to platforms, as is its wont. Typically, while building entire towns and physical spaces out of LEGO, one tends to create a base first and then build out other elements (buildings, trees etc.) and attach them to this base.


Platforms work in a similar way by providing an underlying infrastructure on which others build and add value. Android provides the underlying infrastructure for creating applications on top, YouTube provides the base for creating videos on top and Etsy provides a base for creating product listings on top.

In all these examples, the base has no value without the units that are built on top of it. These units represent value. The platform becomes more valuable as more units get added on top of it. These units become inventory for the platform, if you will.

Since LEGO is physical and has tangible building blocks, one is tempted to think in terms of content platforms or app platforms where content or apps created on top of the platform have some tangibility. However, this metaphor extends to anything that can be added on to the platform that increases its value. E.g. Uber drivers stream their availability and location, thus adding value (and inventory) to the platform.


FedEx: Movement on a network

Let’s move away from the notion of platform-as-underlying-infrastructure and think of the platform-as-a-network-of-transfer.

I’ve talked about this before. FedEx works like a network, it delivers packets from a source to a destination. Here’s how FedEx gets it right most of the times (and I’m again oversimplifying for illustration):

  1. It gathers data from the sender (source) about the recipient (destination) (Address etc.)
  2. It converts this data into some internal code to give routing instructions to everyone involved in the transfer
  3. It delivers the package

Routing, a concept from network theory, is the process of selecting the best path on a network.


There are two key aspects to networks like FedEx:

  1. Packages move around
  2. Packages carry some information that enables the network to send them to the right destination

Every platform is responsible for the transfer of packages from source to destination.

Twitter transfers Tweets from tweet creators to tweet readers.

YouTube transfers Videos from Video creators to Video viewers.

LinkedIn transfers Jobs from Recruiters to Professionals.

Airbnb transfers Apartment listings from Hosts to Travelers.

Uber transfers a Cab request from a Traveller to a Cabbie.

All of these platforms rely on the ability to correctly transfer information (and content) from the source to the destination.

While it is important to think of YouTube as an infrastructure on which users host videos, it is also important to see YouTube’s role in transferring videos to the right consumers. This is where the platform needs information to correctly recommend and serve videos.


Sparks, Catalysts and Reactions

Let’s think again about these units that we have been talking about. These units always start off an interaction between a producer and a consumer. You need an Uber driver to be available and around for him to be matched to a traveler. You need a listing to be put on Ebay or Etsy to enable any commerce. And going deeper, you need a seller to set up an instance of her product on Paypal for a buyer to buy it using Paypal.


In all these cases, the unit sparks the interaction. David Evans, in his book The Catalyst Code, talks about sparks that start exchanges and catalysts that enable these exchanges but never gets quite specific about what the spark is.


Soccer: Thinking social and interactions

Think of a game of soccer or ice hockey. The game always has a central point of focus: the ball or the puck. Spectators do not view the stadium or rink as a whole, they are solely fixated on three things:

  1. The ball or the puck
  2. Who has control over the ball or puck now
  3. Who is likely to receive/intercept it next

The ball or puck is the centre of a social interaction here. Social interaction between two or more parties always has a center. The initiator and the recipient of the interaction are determined relative to this centre.

Sébastien Chabal

Moving back to platforms, every platform facilitates interactions between producers and consumers. These interactions always occur around a center, much like the ball or puck.  Much like the rules of play in a game and the role of the referee, the platform’s purpose is to provide the conditions that are necessary for such interactions to occur efficiently.


The Nature of Order: Ask the Architect

When thinking about design – of products, platforms or business overall – I’ve always found it useful to learn from those who do this for a living: Architects. While the physicality of real world design is different from what we look at while building digital systems, the fundamental aspects remain the same.

In The Nature of Order, Christopher Alexander makes a great case for inverting how we think about design. Here’s the gist: If you’re going to design a building, you don’t start by thinking of the building. When architects get down to thinking on paper, they start with a center and build things around that and in relation to that.

Minnesota History Center Ceiling

The actual argument is much more nuanced but what I find extremely compelling is the notion that we might be missing the point in thinking in terms of wholes. When we look at a website or an app, we see it as a whole. We don’t see a center there.

The more complex a system, the more difficult it’s going to be to understand it as a whole. Thinking in wholes forces us to start thinking about features. This is where it becomes super-helpful to think of centers.

When building a platform, it is super-important that features are not the starting point. Features should, instead be built around a center. The center determines how platforms users will interact, and leads to what kinds of features the platform should have to enable these interactions.

If we think of YouTube, Airbnb or Theadless, every feature that is built out there is built in relation to the video, the listing or the graphic. You don’t have features for the sake of having features. In the initial days of any internet startup, in particular, a definite way to mess things up is by thinking in terms of features and ending up with something clunky. The closer one aligns each and every feature with the center, the more intuitive the platform’s use is.

This ties back to how we think about Pipes (linear businesses) as well. Toyota builds out a factory starting with the car (end product) in mind. If it goes about adding things to the factory without the car in mind, it ends up with bloatware. The same thing happens on digital platforms.



If you’ve read the previous posts in the series, you would have related by now that these units are the ‘seeds’ that I refer to in the previous post.

The idea of this unit of value as a building block of the platform is very central to understanding platforms. 

However, I’m not entirely convinced that ‘seed’ is the best way of describing this unit.

At this point, I’d like to throw this open. Based on the above, what would you call this? What name would you give this? There are several elements to what these units are and we’d need to come up with a name that captures these elements.

So let’s recap the various properties of these units:

  • Unit of value: What adds value to the platform and without which the platform has little or no value?
  • Unit of exchange/transfer: What do platforms carry and transfer between producers and consumers?
  • Represents the job to be done: What is the information required to get the job done? The unit represents the job to be done. It is the information that is required to get that job done. (Credit: Kevin Carroll, comments below)
  • Unit of production and consumption: What is created by producers and consumed by consumers on the platform?
  • Something created on top of the platform: What is created on top of the platform? What can be attached on the platform that complements and extends its functionality?
  • Unit of information required to start an interaction: What is the basic unit of information that needs to be created and consumed for the exchange/consumption of content, goods and services to happen?
  • Pre-condition to activity: What is a pre-requisite on the platform without which producers and consumers do not interact with each other? What is that spark or ignition to start it?
  • Center of interaction: What is the centre around which the interaction between producers and consumers gets structured?

Should we call it the seed, the value unit, the spark, the core, something else? Which word do you think makes the most sense? Would you mind choosing one at the link below that most closely explains what we’re trying to say here?

Click here to answer this question

Thanks for helping out above!


In Summary…

This is the enduring message about platforms. Value them for what value really is when it comes to platforms: not technology, not user growth, but network effects. And network effects are determined by these units that are created on the platform. These units are the fuel that run platforms and make them worth billions of dollars.


Tweetable Takeaways

WhatsApp value is in the interactions (stickiness), not in the technology or user growth alone.  Tweet

Twitter, WhatsApp and Instagram change how we think of value in the Platform Economy. Tweet

The Platform Economy isn’t about technology, it’s about interactions that are enabled by technology. Tweet

To read the rest of the posts in this series, click here

Image Source: Flickr/Creative Commons

About the Author:

  • Ngoc Anh

    Relax with the entertainment you here okay.

    Thank you. Wish you a very effective working days.

    Friv 200


  • software7001

    For sure, there are some lessons to be learned here.Holdings and liabilities group together with man’s steps, certain names display etch an indication individually aided by specific innate top personality also aided by the subsequent to men and women command.000 millones de d lares, segun analistas privados.You can still make your cheap nike air max 90 day by wearing your chelsea shirt.One girl from Sichuan International Studies University said, “I never buy a new things just because something is broken, as long as the things I like, I will not consider whether I had bought something like that”.Coordination with personnel through the whole game is absolutely vital.

    Pain relievers and anti-inflammatory drugs are used to ease immediate symptoms, but by themselves will not cure the condition or keep it from recurring.0, 5.I feel mind f***edBased on the calories I have left after working (which for me is standing/walking the whole time) and running earlier today I should be eating 2500 calories to maintain the deficit!Signature links are also indexed by the search engines and will provide you more links to your site thus increasing your ranking in the search engines and driving more traffic to your site!- DON’T ask anything too personal.The spoils constantly shift to those that innovate and execute the best.

    Getting a company with good name can be quite necessary for just about all.If something says Hang Dry or Line Dry, don’t dry it in the dryer.There appears to be a repetitive pattern that as long as NKE’s dividend yield reached the 1.Since corporations don’t have the same responsibilities as us, they shouldn’t have the same rights.These gels may vary in strength, meaning some are only effective in preventing chafing during normal activities while others will last longer and are made for those who wear active gear such as a bike chamois or a wet suit for endurance activities.Soon enough, you’ll have the dream wedding cake for your wonderful wedding.

    (from Nike Consumer Affairs packet, 1996)Nowadays, it not fresh to see pop music celebrities wear Jordan shoes in all kinds of business cheap nike air max activities.You can also monitor your heart rate, but to do that you need to buy an additional heart rate monitor.Humbly now, with no vaulting dream of achieving a vast unity, I wanted to try to build a bridge of words between me and that outside world, that world which was so distant and elusive that it seemed unreal.80% to 36 cents/share.Trying to catch snowflakes on her tongue.There would be a huge demand for IT security professionals to grow in the coming years with the advancement in technology.

    Anti Aging And Its SecretDuring its 3Q13 call, the company mentioned that it is taking the necessary decisive actions to re-establish a pull market in China.Before developing your formal business plan, you must make sure you have put together a solid management team.This Nike Air Max+ 2009 running shoe has been developed for women who have a neutral to under-pronated gait.These kinds of earrings needed a large amount of precious alloys plus gemstones therefore had been extremely hefty, which meant that many women who dressed in them suffered from elongated nike air max 1 sale earlobes.Nike’s swoosh trademark was thought out by a student with the name of Carolyn Davidson majoring in graphic-design in Portland State University.

  • moovd

    Nice piece, Sangeet, and very instructive on how to become THE hub, rather than another node in someone else’s hub!

  • Sangeet Paul Choudary

    I agree, Hemant! One can’t help but believe that even FB wouldn’t have valued WhatsApp at this level. Some of the valuation is definitely a fallout of a bidding war and the fact that FB probably didn’t want to lose another deal the way it lost the Snapchat deal.

    I don’t think anyone can comment on the justifiability of $19 Bn at this point. NPV calculations are difficult simply because FB itself has proven skeptics wrong on the mobile by ramping up its revenue over there in such short time.

    But I think one thing we can confidently state is the fact that everyone valuing it for its technology or user base alone is missing the point. Those seem to be the issues usually. We heard the same things during Instagram. It’s a bit ridiculous to compare a network effects company with a S/W company (like Symantec) and say that the tech isn’t comparable.

  • Hemant Bhargava

    I think you’re spot on with that last point … “not be about how much money it can make but how much money it could take away from Facebook were it not acquired”. Though a marginal cost thinking economist would say that the purchase price should have been the loss Whatsapp could have caused FB minus the (NPV of the expected) loss that some upcoming next-app will cause FB in the next few years.

  • Sangeet Paul Choudary


    Thanks for such a detailed comment. I completely agree with the fact that the greater fools theory is at play. I would term the WhatsApp acquisition a gamble at best because (as you mention) of the poor lock-in. The easier it is to replicate a network, the lower the defensible value of the network effect, one might think.

    So overall, I do have an overtly optimistic take in this post. My main intention of taking such a one-sided stance was to illustrate the fact that the value is not in the app or the user base, it’s in the network effect. I see a lot of techies complaining about how simple the app is and a lot of excel gurus comparing companies on valuation/user when the definition of user isn’t constant across companies. Hence, I wanted to bring out that difference while also illustrating value creation on platforms. I merely used the WhatsApp story as a hook to explain this to people who may otherwise not care about platforms.

    So I’m wholly in agreement with where you stand on this, with one addition. As far as I see it, the WhatsApp acquisition may not be about how much money it can make but how much money it could take away from Facebook were it not acquired. Only time will tell whether the second problem gets solved with this acquisition or whether a Telegram or WeChat comes in to spoil Facebook’s party. Given the low defensibility and high multi-homing, I tend to believe this was over-priced.

  • Hemant Bhargava

    Sangeet – I’m not going to be a naysayer or a Luddite – as you know I do get network effects :) But yet, I find the Whatsapp valuation to be inexplicably high. (The one reason I could be convinced its not too high is that most of the valuation is in Facebook money!)

    But seriously, here’s how I see this. You mention above,

    “Most platforms look like stupid ideas when they have no users. ”

    Ok, so we’re not arguing about that. Whatsapp has half a billion users, and lets agree they will soon get to a billion and more. But lets ask what kind of NPV per user would justify Facebook paying $19 billion? After accounting for the fact that there is already some user overlap between WA and FB, lets say for simplicity we’re looking at a valuation of $100 per user. Now, that seems high knowing that the big chunk of that user base is in countries where $100 means a lot of money. But, more importantly, we need to think of

    1) the “shelf life” of Whatsapp (how many years before it gets displaced by some other interaction technology), and

    2) the absence of substantial lock-in effects in Whatsapp? Yes, there is some lock-in because of your address book but that is easily replicated. The bigger lock-in would be if the content that is exchanged during these interactions is both protectable and presents ever-lasting or long-lasting value to the user. I don’t see that. (Compare against, say, the lock-in effect of having hundreds of dollars worth of apps on your iphone; or hundreds of documents created in a high-end document deisgn software.)

    Given my intuition about (1) and (2) I feel that an NPV of $100 per user — on the average — over hundreds of millions of users from developing countries will be very hard to achieve. What I see happening is a repeat of the “greater fools theory” that was at play in the dot-com boom and bust. Facebook built lots of users, got a huge valuation, then ran into headwinds over monetization — so decided to buy the next new toy around with that money — and that next new toy will run into headwinds in a few years — and they might buy the new toy that comes after that — but when will the real monetization occur ?

  • Kevin Carroll

    Agreed! In my design I capture “intent” then call it an “order”. After all, orders do communicate intent and they can have conditions. Think stock exchange orders.

  • Sangeet Paul Choudary

    This is a great comment! You’re spot on with the JBTD and the rJBTD because that’s how I think of the UNIT and the other associated objects. E.g. YouTube has to first get its video right before it goes around solving for comments. Platforms like MySpace lose focus when they go around solving fro rJTBD.

    Facebook’s current competition problems are also coming about because there isn’t one single UNIT and INTERACTION on Facebook. There are many. And whenever a competitor takes one of these UNITs and provides a better INTERACTION, Facebook has a tough time.

    But trivialities aside, this is a great comment and helps me focus further on explaining the unit.

    What would you call the unit? I think it’s important that people get what the unit is above without a lot of explanation. A good name will go miles in helping understand how we think about platforms.

  • Sangeet Paul Choudary

    This is fantastic! In the next article, I want to talk about how the unit enables the core interaction. I’m going to leverage the metaphor and design principles of gameplay to talk about it.

    The goal is to remove the fluff while building platforms. And then layer the fluff that helps without making or breaking the game, as you put it.

    Thanks again!

  • Kevin Carroll


    The puck is Clayton Christensen’s “Job-to-be-Done”. Getting the puck in the net is Nir Eyal’s “Morphine” problem. I call the unit the #JTBD which is distinguishable from the related JTBD or #rJTBD that needs Eyal’s “Aspirin”. Catching a pass, stick handling, skating, etc. are #rJTBD activities.

    If you optimize for #rJTBD you end up with Christensen’s sustaining, not disruptive, innovation. Classic case is higher education industry which, at present, is optimizing for “cost reduction” and “learning quality” at the same time. Solution providers are feeling quite disruptive b/c money is being thrown their way. HOWEVER assuming the market is accurate and honest (which it is) it’s greatest pain point is “placement at best-fit” which makes “learning quality” a #rJTBD. Albeit a very necessary #rJTBD.

    Being great at #rJTBD activities is fine until a competitor comes in and provides morphine for the greatest pain point (#JTBD) to attract the market then starts providing aspirin for the lesser pain points (#rJTBD). OUCH!

    Moving along with the ice hockey analogy: For a goalie the #JTBD is to prevent the puck from being delivered. Imagine a goalie so big that the puck could not be delivered into the net. Then we could care less about the #rJTBD that involve goalie skating, rebound control, flexibility, etc. We would pay the goalie a lot of money b/c he’s great at completing the #JTBD even though he is terrible at everything else which are #rJTBD .

    So, to reduce #bmgen sustainability risk the most the UNIT should be the way to capture the most important “intent” which is “eliminate my greatest pain point”.

    I hope I helped.

  • Jordan Kohl

    This is a common idea in game design, usually called core gameplay loop or game mechanic. It’s the central part of the game, which has to be fun, and gets repeated, iterated, and built on throughout the game. For example, in Super Mario, the core gameplay mechanic is running and jumping. If that part isn’t fun and well executed, then the rest of the game fails.

    This is why simple games like Tetris can be just as popular as complex games like Grant Theft Auto. They both have a core gameplay mechanic (putting together a puzzle before space runs out vs. stealing cars). It also partially explains why “cheap” games like Minecraft are wildly successful. The core gameplay mechanic: building with life-size LEGOs, is a lot of fun. All of the other “fluff” helps, but they don’t really make or break the game.