Platform Scale by Sangeet Paul Choudary

  1. Choudary explains how technology, democratization of connectivity and rise of data-driven decision making systems are enabling a new type of business model – platforms. Platforms are so powerful because they enable efficient interactions, create excess value and are able to scale rapidly
Key Takeaways
  1. The Platform Manifesto
    1. The ecosystem is the new warehouse
    2. The ecosystem is also the new supply chain
    3. The network effect is the new driver for scale
      1. Platform scale is achieved by maximizing the repeatability and efficiency of the platform’s core interaction. Interactions must be executed smoothly and in a manner which kick-starts the next interaction organically
      2. Achieving platform scale requires the ability to scale value creation to scale value exchange – the ability to scale production and consumption simultaneously – and to repeat the two so that each reinforces the other
      3. 5 drivers of platform scale – minimal marginal costs of production and distribution, network effects powered by positive feedback, behavior design and community culture, learning filters, virality
    4. Data is the new dollar
    5. Community management is the new human resources management
    6. Liquidity management is the new inventory control
    7. Curation and reputation are the new quality control
      1. One of the platform’s main focus is limiting poor behavior and interaction risks
      2. Quality control (screening, curation) is vital. Can be done through an in-house editor, through algorithms or through social signals (rating, voting)
      3. 3 factors governing platform adoption – network effects (most important), curation of content, curation of participants (through ratings, reputation, incentives – indicating quality and reliability)
    8. User journeys are the new sales funnels
    9. Distribution is the new destination
      1. New focus on how to distribute its experience into multiple user contexts
    10. Behavior design is the new loyalty program
      1. 3 core principles to platform design
        1. Start with defining the value that is created or consumed, the core value unit
        2. The core interaction – the set of actions that enable the creation and consumption of that value – should be laid out around the core value unit
        3. The design of the platform’s features, functionalities and management should stem from the design of the core interaction
    11. Data science is the new business process optimization
    12. Social feedback is the new sales commission
      1. Platforms often create new behaviors and reward/reinforce the most beneficial
    13. Algorithms are the new decision makers
    14. Real-time customization is the new market research
    15. Plug-and-play is the new business development
      1. Platform as an enabler of interactions – plug-and-play business design, balancing value creation for both producers and consumers, strategic choice of what is “free”, pull/facilitate/match, layering on new interactions, enabling end-to-end interactions, creation of persistent value beyond the interaction
      2. At their core, platforms enable a plug-and-play business model. Other businesses can easily connect their business with the platform, build products and services on top of it, and co-create value. Platforms primarily benefit not from internal production but from a wider source of open co-creation and open market interactions. This ability to drive interactions through a plug-and-play infrastructure is a defining characteristic of platform scale
    16. The invisible hand is the new iron fist
  2. Business model transition from pipes to platforms
    1. Choudary calls traditional companies like manufacturing, “pipes.” Pipes build products or craft services, push them out, and sell them to customers. Value is produced upstream and consumed downstream, creating a linear flow of value, much like water flowing through a pipe. In effect, pipes were designed to enable the flow of value in a straight line
    2. Three forces today are driving a whole new design for business, platforms – increasing connectedness, decentralized production and the rise of AI. These businesses create a plug-and-play infrastructure that enables producers and consumers of value to connect and interact with each other in a manner that wasn’t possible in the past
    3. In this new design of business where the firm is no longer the producer of value, platforms perform two specific roles
      1. They provide an open, participative, plug-and-play infrastructure for producers and consumers to plug and interact with each other
      2. They curate participants on the platforms and govern the social and economic interactions that ensue
    4. Shift in markets from consumers to producers (both can and do add value on platforms whereas only one side typically did in the past)
    5. Shift in competitive advantage from resources to ecosystems
    6. Shift in value creation from processes to interactions
  3. The Broad Goal of Platforms
    1. Goal of platforms is to enable interactions between producers and consumers repeatedly and efficiently
      1. Build platforms with an interaction-first, not a technology-first mindset! Technology should be built only after understanding the interaction that needs to be enabled. Without this in mind, one often ends up with a platform that nobody wants to use.
    2. The movement from pipe-based, user-first view to the platform-based, interaction-first view is best captured through the following shift: We are not in the business of building software. We are not in the business of selling products and services. We are in the business of mediating and enabling interactions
      1. The importance of an interaction-first approach to building platforms cannot be emphasized enough. Focusing on the actions involved in an interactions helps us design the tools and services as well as the rules required to facilitate the interaction. Understanding the players participating in the interaction and their motivations helps us design the actions and rewards that create pull on the platform. Finally, only by focusing on the core interaction can a platform know what data it needs to capture
  4. The Core Value Unit
    1. The core value unit is the minimum stand alone unit of value that is created on top of the platform. It represents supply or inventory created on top of the platform and without this, the platform has very little value in and of itself
      1. For network/marketplace/community-dominated – goods, standardized services, non-standardized services
      2. Infrastructure dominated – apps
      3. Data dominated – data helps the platform become more efficient overtime, data itself is the source of value
      4. To increase platform scale, focus on increasing the quality and quantity of core value units on the platform. However, platforms are unique in that they don’t control this inventory as this is produced outside the platform
      5. All actions in the core interaction fall into one of the following buckets – creation, curation, customization, consumption. The keys to platform scale lie in simplifying each constituent action in the core interaction
      6. Information exchange has 3 components – The producer creates a core value unit, the consumer sets up a filter of some combination of overlap and data, the value unit that best passes through the filter is served to the consumer (based on good data and filters). Filter can be point in time (search) or cumulative (taking account of past history or behavior) or some combination
  5. 6 elements of execution
    1. Choice of the overall interaction space – connection, content, clout, coordination, competition, culture and code
    2. Production incentives – tools/access/both, simplify production process, great curation, clear, democratic and equal access path to the top, great conversion rates, good feedback mechanisms, removal of skill, time/effort/investment, resource, access barriers (removal of frictions)
      1. Frictions can sometimes be useful when trying to discourage the repeatability of undesirable interactions and can indicate quality, superior signaling or a barrier of some sort
    3. Building long-term cumulative value – reputation, influence, collections, learning filters
    4. Strong curation mechanisms and trust
      1. 7 Cs of Trust – confirmed identity, centralized moderation, community feedback, codified behavior, culture, completeness, cover
    5. Strong filters and relevance
    6. Ownable interactions – more difficult for platforms offering nonstandardized services (TaskRabbit) but in order to own the interaction, all platforms must create more value than they capture
  6. The Chicken and the Egg Problem
    1. All platforms must overcome the chicken and egg problem until they reach critical mass, the minimum network size at which there are enough producers and consumers of value on the platform to ensure that interactions spark off reliably
    2. Solutions to the chicken and egg problems have a few defining characteristics:
      1. Breaking the vicious cycle – platform should have standalone value, users to derive value even without other users
        1. The standalone mode, for producers, should encourage the creation of value units on the platform, which can then be used to pull in the consumption side
        2. Faking initial supply may often help kick start network effects (YouTube had pirated content early on) – seeding and weeding, seeding demand, seeding supply
        3. Identifying a group of power producers and providing them with tools and incentives to better “harvest” their following can solve the chicken and egg problem very effectively
        4. Get more difficult side on board through curation and incentives
        5. Often, the solution to finding adoption lies in providing backward compatibility with existing solutions
        6. Focus on value-creating interactions and then scaling those interactions instead of focusing entirely on scaling the user base. Small user bases with thriving interactions trump large user bases with low activity
          1. Solve a pain point for a niche segment, target a micro-market where small is good, leverage existing interactions in the micro-market, find a micro-market that encourages spread, find a micro-market that is representative of the final market, a micro-market may be a thin-sliced use case, make a two-sided market one-sided
          2. A platform can scale well only if it encourages interactions within a small user base before attracting a large number of users
      2. Positive feedback
      3. Maximizing overlap between consumers and producers
      4. Getting the harder side in first (through incentives)
      5. On-boarding of two distinct markets
    3. Five design principles for solving chicken and egg problems
      1. Finding a compelling bait to start the loop
      2. Ensuring there is no friction in the feedback loop
      3. Minimizing the time it takes for the startup to reach critical mass
      4. Incentivizing the role that is more difficult to attract
      5. Staging the creation of two-sided markets
  7. Scaling & Virality
    1. Scaling strategies
      1. Bump – non-sustainable exposure such as PR, advertising and events; important for initial traction
      2. Engines – an internal engine of growth and designed to grow as a consequence of usage
      3. Also needs to create the hooks and motivations that will enable and incentive users to expose the offering to others, every time they use it
    2. Misconceptions about virality
      1. Virality and word of mouth are two names for the same phenomenon – virality a consequence of users using the platform, not loving the offering. Virality does not need fans, it merely needs users who are encouraged to bring in other users
      2. Virality and network effects are the same and lead to rapid growth – open platforms like email do not benefit form network effects whereas closed ones do (but both can have virality)
      3. Virality is all one needs for a growth strategy – should be complemented with other user-acquisition models
      4. Virality involves manipulating users to send out invites to other potential users
    3. Networks spread like diseases do
      1. The sender – a user on the platform sends out a message about the platform
        1. Sender incentives – why will the sender send units out of the platform?
      2. The core unit – message is typically the core value unit
        1. Spreadable unit – what is the minimum transferable unit on the platform that one can move on an external network?
      3. The external network – units spread on an external network, connecting people
        1. External network – Where will the unit from the platform meet current non-users
      4. The recipient – recipient on the external network interact with the unit and is brought back to the original platform
        1. Recipient incentives – why will a non-user on an external network convert to a user on the platform?
        2. The recipient, if interested, then joins and becomes a sender and starts the process over
    4. Virality is a design problem, not an optimization problem. Take into account:
      1. Sender incentives, low friction in creating core value units, high percentage of producers, spreadable core value units (triggers an interaction on an external network), plays on the producer-as-sender dynamic, the spread of the unit helps to complete an incomplete interaction)
      2. External network – choice of network which takes into account relevant interactions, relevant connections, relevant look and feel, add value to users on this external network, create an unfair advantage and make integration as easy as possible
      3. Recipient incentives – unit should serve as a compelling pitch to the platform and a call to action embedded within the unit
      4. 4 key optimization priorities for achieving sustainable viral growth
        1. Send: maximize outflow of units from the platform
        2. Spread: ensure that units spread on the external network
        3. Click: maximize clicks on an external network
        4. Convert: minimize cycle time
    5. Producers never spread the word about the platform, they merely spread the word about their creations
    6. Platforms that succeed with viral growth reward users with accelerating social feedback
    7. Network effects can work against platforms if higher adoption gets in the way of interaction efficiency and repeatability, reducing interaction quality. To achieve sustainable scale, a platform needs to scale both the quantity and the quality of interactions that it enables
      1. A scaling strategy for platforms should involve scaling of production, scaling of consumption, strengthening of filters through ongoing data acquisition, scaling social curation, scaling community culture, minimizing interaction risk
      2. Lack of curation scaling is very common when platforms fail. Platforms need to ensure that access and creative control, as well as curation and customization, scale well as the platform scales
      3. Platforms must encourage cross-cluster interactions as well as cross-cluster incentives
  8. Other
    1. Platforms aren’t truly software but they are eating the world – efficient social and business interactions, mediated by software
    2. Value creation still dependent on aggregation, but not of labor or resources. Rather, the ecosystem is the new warehouse, supply chain and scale through network effects. Shift from culture of absorption to data absorption. Manage community incentives and governance
    3. Must ensure there is never unfilled demand
    4. While platforms can be incredibly different, the following three distinct layers tend to emerge repeatedly: data, infrastructure, network-marketplace community. These 3 can play varyingly large or small roles depending on what the platform wants to achieve, how to differentiate itself and what the key drivers of value are
    5. The single most important decision in testing is the choice of the hypothesis to be tested. Without clarity on this, one can waste a lot of time testing irrelevant hypotheses and optimizing poor design. Laying out the overall architecture of the platform helps us understand the key points of failure for the ensuing platform business and shows us what needs to be tested. All design decisions should ensure the repeatability and sustainability of the core interaction that the platform enables
    6. The platform canvas is a framework for makers to build interaction-first platform businesses and includes the value-creating interaction, the platform that enables the interaction, a mechanism for value capture, enablement of a plug-and-play business model  through channels (websites and apps) and access control for producers and filter creation for consumers. The platform must provide tools and services of creation, curation, customization and consumption.
      1. Value is derived from charging one side to access the other, charging a third party for advertising, charging producers and consumers for premium tools and services, charging consumers for access to high quality, curated producers and charging producers for an ability to signal high quality
    7. The TRIE Framework – tools and rules, interaction, experience
      1. Platforms allow the users to shape their own experience and not just accept the maker’s ideas
      2. Platforms must allow for emergent behavior to arise, some of which may redefine the architecture and lead the platform in entirely new directions
    8. Everything old is new again! The answers lie in using the old to interpret the new
    9. Platform strategy involves 3 primary priorities, aligned with the three layers of the platform stack – pull, facilitate, match
    10. Two critical factors will determine the success of a company in the on-demand economy: multihoming costs (ease of switching between platforms) and interaction failures
    11. Best way to launch a platform business at a conference is to ensure that the core interaction on the platform is organically embedded into the conference experience and that it fits in with the activity at the event
What I got out of it
  1. A dense and extremely insightful book on how to design, think about, build and spread successful platform companies. At the core of it, platforms must make sure they enable their core value unit to foster interactions which are as frictionless as possible in a repeatable, efficient and effective manner