Platform Revolution by Geoffrey Parker, Marshall Van Alstyne and Sangeet Paul Choudary

The Rabbit Hole is written by Blas Moros. To support, sign up for the newsletter, become a patron, and/or join The Latticework. Original Design by Thilo Konzok.

Key Takeaways
  1. The platform business model
    1. One which uses technology to connect people, organizations and resources in an interactive ecosystem in which incredible amounts of value can be created and exchanged.
    2. Network effects are the main source of value creation and competitive advantage in a platform business - It encourages an open, cooperative interaction which removes friction from matching suppliers and consumers, also removing barriers of time and space.
    3. Frictionless entry to the network is key.
    4. Because they create value with resources they don't own or control, they can grow much more rapidly.
    5. Side switching, ability for consumers to easily become producers, is also important
  2. Traditional linear businesses, aka pipelines, cannot scale as efficiently as platforms since platforms eliminate gatekeepers which typically slow down flow of information from producers and consumers. Platforms can also "unbundle" services so consumers can get exactly what they want for less. Everyone wins except for the traditional players. Platforms bring new sources of supply online and pits high fixed cost companies against low fixed cost companies (Hyatt vs AirBnb)
  3. In platform markets, the nature of supply changes and harnesses feedback from consumers and the community which only used to consume. Platforms use database tools to create community feedback loops in which the users' reputations are always at stake and to determine which projects are most promising
  4. Platforms often invert traditional business models by focusing more on functions outside of the business than internal to it
  5. Network effects - can be both positive and negative. Positive effects occur when additional value is created for each new member that joins a network. This is the principal competitive advantage of platforms
  6. Traditional businesses dominated through supply economies of scale - efficiencies in production from scale. New businesses tend to dominate from demand economies of scale - take advantage of efficiencies from technological advancements on the demand side such as efficiencies in social networks. This is one of the key reasons network effects are so powerful
  7. Two sided markets feed the network effects. At Uber, riders attract drivers and drivers attract riders in a powerful positive feedback loop
  8. Important to distinguish between network, price and brand effects. Network effects leads to a virtuous cycle that is more sustainable whereas price and brand can be fads or not as sustainable. Most platform failures rely mostly on price or brand effects
  9. Virality about attracting people to join network, network effects about delivering more value once on platform. Temporary vs. sustainable
  10. Skillful curation helps mitigate negative network effects (OK Cupid and how originally all men flocked to most beautiful women who then left, and then men left because there were no more attractive women)
  11. Data driven network effects - the more users and therefore the more data you have, the more effectively you can curate
  12. Same side network effects - the effects consumers have on other consumers and producers on producers
  13. Cross side network effects - effects consumers have on producers and vice versa
  14. 4 broad buckets for different companies - asset providers (Ford), service providers (United Healthcare), technology providers (Microsoft), network organizers (platform businesses and are the most efficient value creators)
  15. Every exchange between producers or consumers involves 3 things - information, goods or services and some sort of currency (money, attention / reputation / other forms of social currency)
    1. Value creation for platform directly depends on what kinds of currency are being exchanged and how much of it the platform can capture
  16. The why of platform design - core interactions.
    1. The core interaction is the most important form of activity of the platform and has 3 core areas - participants, the value unit (any info which helps users decide if they want to proceed - price and description of item on eBay) and the filter (search query or any other filter which effectively provides users only value units they're interested in). These 3 must be designed very well to make the core interaction as frictionless and value accretive as possible
    2. Platforms don't create value units so platforms are information factories where producers can provide value units
  17. The how of platform design - pull, facilitate, match.
    1. Platforms face a chicken and egg problem where consumers won't come if there are no producers and vice versa. Most platforms fail because they don't effectively pull consumers and or producers into the platform. Finding and leveraging feedback loops which drive engagement is vital (FB changed focus from getting new members to helping current members establish new connections)
  18. Modularity in a platform is important - structures which are designed independently but can all function together. Subsystems can interact in a way that yields complex adaptive behavior without any one subsystem being too complicated
  19. Inevitable that users will use platforms in ways the designers never anticipated. Sometimes the best design is anti-design - allowing space for the bizarre and unanticipated to grow and expand
  20. Software eating the world is evolving to platforms eating pipelines. Internet no longer just a distribution channel, a pipeline, but also acts as a creation infrastructure and coordination mechanism. Physical and digital are also rapidly merging. Platforms enjoy two main advantages - superior marginal economics of production and distribution and network effects allows platforms to scale much more quickly
  21. Platforms eliminate barriers to entry, bringing in more supply and therefore competition
  22. All platforms struggle early on with quality due to abundance but as the community grows and curation improves, quality typically improves
  23. Delinking assets from value allows B2B to move to more profitable B2C
  24. Re-intermediation is the process of adding new, nimble, automated, customer rated, value-add middlemen to transactions. Separating ownership from control
  25. Nike is one of the most successful pipeline companies turning themselves into a platform. Fuel band, apps, Apple Watch and more helps connect products with platform businesses - fueling growth and keeping customers engaged in a new ecosystem. Eventually allowing them to make better products. Under Armour attempting same thing with purchase of MapMyFitness  MyFitnessPal and Endomondo- all about platforms, data and users, not products.
  26. IoT allowing industrial companies like GE potentially create a viable platform
  27. Chicken and egg problem with platforms occurs when both sides of the market are equally valuable and PayPal overcame by reducing friction involved in online transactions, making very user friendly, gave new customers money for just signing up ($10-$20).
    1. Getting users to sign up is step one, then must realize value and become regular users - user commitment more important than user registration. Multiple positive feedback loops fed at this point and explosive number of customers fueled sellers to promote their acceptance of PayPal. Enabling service on eBay made it even more visible and reduced friction
  28. Platforms must rely on pulling in customers rather than push like most pipelines do. Marketing must be built into the platform
  29. Knowing value proposition of competitors, even if seemingly similar product or service, can help you structure your own
  30. Strategies for beating chicken and egg problem
    1. Follow the rabbit - build on top of already successful pipeline or platform, attracting both consumers and producers (Amazon Marketplace, Intel)
    2. Platform from scratch
      1. Staging value creation - attract initial users who attract more users and so on. Huffington Post started with very high quality content, attracting first consumers to engage more and therefore attracting more users
      2. Designing platform to attract one set of users - critical mass on one side will come to attract other side
      3. Simultaneous onboarding - value for those who first join but increasingly more valuable as more join (Facebook)
    3. Piggyback strategy - connect with a user base from an an existing platform and stage the creation of value units to recruit those users
    4. Seeding strategy - create value units for at least one set of users who will then attract other users because they want to interact. Platform owner can be first customer, leading way in showing how to take advantage of value units and what types of rules and interactions are recommended (Google and $5m prize for best apps when first launched Android; Quora at first would ask and answer their own questions, showing how it's done)
    5. The Marquee Strategy - sometimes one side of the market can make or break the platform so targeting them is key (may purchase marquee producer like Microsoft did with Bungee which eventually became Halo)
    6. Single side strategy - develop platform for only one side and then entice other side to join
    7. Producer evangelism strategy - platform helps producers better serve consumers and cross-pollination can then ensue (Kickstarter, Udemy)
    8. Big Bang Adoption Strategy - use one or more push strategies (marketing, etc.) to attract a high volume of interest, creating a nearly fully formed network almost instantaneously. Less effective in today's world with so many ads and distractions
    9. Micro market strategy - target a tiny market which is already engaging in interactions - enables platform to act as a market even in earliest stages of growth
    10. Viral growth (complements all strategies above) - encourages users to spread the word about the platform to other users, network thus becomes the driver of its own growth - positive feedback loop in action
  31. Key pillars
    1. Sender (love the platform and often get some benefit such as money or notoriety)
    2. Value Unit (spreadable value units key, can't be secret or hard to spread)
    3. External Network (Instagram leveraging Facebook)
    4. Recipient (responds and spreads word if find enough value)
  32. Monetizing platforms - about capturing a portion of excess value created
    1. Determine value of platform is step 1 - only way to be sustainable is if it doesn't hamper network effects and even better if it reduces possibility of negative network effects
      1. Consumers - access to value created on the platform
      2. Producers - access to a community or market
      3. Both - access to tools and services which improve interactions
      4. Curation mechanisms - connecting right to consumers with producers through curation is key
    2. Often a good strategy is not to charge either side early on but simply take a small transaction fee. Transaction will occur and feels like only a small tax on the service provided by the platform
    3. Another good tactic is to charge companies for postmortems to help them understand what they did well and where they can improve
    4. Charging for access or enhanced access. Alibaba charges no transaction fees but fueled network effects by paying members who recruited others
    5. Freemium - charge full price to a certain set of customers and give to free or at least subsidized to those who value it less
    6. Charging one side will often drastically reduce volume but drastically improve quality and engagement and discourage second rate participants. Deciding which side to charge and how much is one of the most crucial decisions
    7. Should try to keep as many monetization opportunities open as possible
  33. Openness - Determining what users can and cannot do is key
    1. No restrictions placed in participation of platform's development, commercialization or use or any restrictions and applied uniformly to all participants
    2. Openness encourages innovation but obviously gives up much control
    3. Determining how open and which areas to leave open is important - can keep key pillars closed but leave other areas open for others to play with and innovate on. Facebook platform opened up massive innovation of apps
    4. As platforms by definition derive value from outside producers and consumers, key to find right level of openness. Defining exactly who should have access to the platform and how they can participate is absolutely vital with huge strategic repercussions and why openness at the top of every platform manager's agenda
    5. 3 key decisions about degrees of openness - decisions regarding manager and sponsor participation, developer participation, user participation
      1. Proprietary (Apple), licensing (Google with Android), joint venture (Visa when first started), shared (Linux)
      2. As a platform manager, cannot let an outside developer drive too much of the value creation on the platform - buy the app or the company that created at this point (Apple bought Siri)
      3. Absolute openness is usually not chosen in attempt to provide highest quality but getting a lot of user participation often leads to very engaged and sticky customers - AirBnb customers can also be suppliers
    6. Platforms in similar arenas may choose to differentiate themselves through varying levels of openness
  34. Data aggregation can be very effective but it must be done appropriately to not feel intrusive or creepy
  35. Good governance - set of rules of who gets to participate in an ecosystem, how to divide the value created and how to solve conflicts
    1. Always create value for the consumers you serve, don't use your power to change the rules in your favor, don't take more than your fair share of the wealth (don't make Keurig's mistake of blocking out all competition from coffee making platform)
    2. The scale and scope of today's largest platforms like Facebook and Alibaba often have direct or indirect consequences on tens of millions of people and hundreds of billions of dollars. They can learn much from cities and states - namely, how best to create wealth and distribute it fairly
      1. 1% drop in a state's anti-corruption leads to ~1.7% rise in GDP! Multiplier effect of loyalty, trustworthiness! Singapore is the prime example and has grown GDP at 6.7% for over 50 years. Good governance matters
    3. Absolute openness doesn't work in companies or states because it can't always be relied upon to be fair and satisfactory for all
    4. Governance failures occur because of information asymmetries, externalities, monopoly power and risk
    5. Good governance increases trust and transparency thereby enabling good interactions to occur, allows people on different sides of the market to find each other more easily, minimizes congestion when too many people are involved or quality is too low and minimizes repugnant activity.
    6. 4 tools of good governance (as used by nation states but should be adopted by platforms) - laws, norms, architecture and markets
      1. Laws of platforms are its explicit rules which determine behavior by producers and consumers
      2. A dedicated community is one of the most powerful forces a platform can have
      3. Trigger - Action - Reward - Investment is the process platform managers use to entice behavior they want and improve engagement and stickiness
    7. Important to give outside stakeholders as much of a voice as inside stakeholders or else decisions will inevitably be made more for the platform's benefit
    8. Act consistently - commitments to act or not act must be able to be counted upon
    9. Don't surprise people and don't play favorites with news
    10. Don't promise not to change, simply promise early notice
    11. Must have skin in the game
    12. It is alright to provide differentiated access and value but must clarify what qualifies
    13. Promote welfare and health of partners, especially smaller partners
    14. Fairness creates wealth in two main ways - sharing of ideas, wise allocation of resources (less fear of being taken advantage of)
  36. Metrics - how to measure what really matters
    1. Cash flow, inventory turns, operating income, gross margin, overhead, ROI - the efficiency through which value flows through the pipeline
    2. Must be able to get close to solidly understanding positive network effects and what drives them
    3. Goal is to measure the rate of interaction success and the factors which contribute to it - most powerful metrics quantify the success of the platform in fostering sustainable repetition of desirable interactions
    4. Pipeline more concerned with flow of value through pipeline and platform manager about value creation for whole ecosystem, all users - both on and off platform
    5. Revenues, cash flow, profitability are key to pipelines but largely irrelevant for platforms during start up phase. Once critical mass is attained, conversion of active users to customers can take priority
    6. Crucial to measure extent to which both producers and consumers are interacting on the platform and increasing their participation over time
    7. Metrics during start up phase - core interaction and value it creates for both consumers and producers
      1. Liquidity - first and most important, minimum number of producers and consumer and percentage of successful interactions is high; interaction failure minimized and intent of users to interact is consistently satisfied within a reasonable period of time
        1. Most important metric early on is one that helps determine when liquidity is reached - tracking % of listings which lead to interactions within a given time period
        2. Must also track illiquid situations (such as when an Uber user opens app and sees no cars available)
        3. Most meaningful metrics are comparative ones - either between groups of users or over periods of time
        4. Don't fall into the trap of over-measuring. What matters is having customers who love, rave and repeatedly use your service
      2. Matching quality - accuracy of search algorithm and the intuitiveness of search tools to connect with users to start value add interactions
        1. Achieved through product or service curation
        2. Sales conversion rate helpful - % of searches that lead to interactions
      3. Trust - degree to which users on a platform feel comfortable with the level of risk associated with interacting on a platform. Achieved through excellent curation of participants on both sides of the platform
      4. Actual metrics used to measure these 3 key areas must be relevant to the type of platform, the types of users and producers, forms of value being created and exchanged and so on
        1. Engagement per interaction, time between interaction, % of active users, number of interactions, interaction capture (for platforms taking a stake of every interaction), market access, producer participation
    8. Metrics during the growth phase - best metrics will change as company grows and it is important to determine when these inflection points occur
      1. A proxy for interaction success can be measured by the ratio of producers to consumers. The ratio of failed interactions and producer fraud are also important to monitor
      2. Using various metrics, producer and consumer lifetime values can be calculated and various strategies tested to see their effects on these important values
    9. Metrics during maturity phase - incremental innovation measured through key metrics must be closely monitored
      1. Studying extensions created by developers is important to keep on top of to see if any changes or adaptations are necessary
      2. Metrics must be actionable, auditable, accessible (comprehensible)
      3. Again, in the end, the most important metric is the number of happy customers on every side of the network who are repeatedly and increasingly involved in positive interactions
      4. Are people happy enough with the ecosystem to continue participating in it actively?
  37. Huge advantage of platforms is ability to incorporate products and services of outside partners into activities and capabilities of the platform
  38. Believes that sustainable advantages are illusory in today's world with how fast technology is progressing. However, platform businesses tend to expand the pie rather than taking market share of a fixed market (like Amazon was able to do with Kindle and self-publishing) or goes sideways and creates new markets with new supply (like Airbnb did with supply of lodging). However, winner take all situations do lend to longer lasting moats as they encourage users to abandon other platforms - network effects, supply economies of scale, high switching costs and lack of niche specialization
  39. Platforms seek exclusive access to essential assets and create the platform to discourage multi-homing (same behavior on different platforms) as this facilitates switching. Apple and not making Flash compatible, Alibaba and not allowing Baidu's bots to crawl their site so that they alone could sell ads to their customers
  40. Data is the new oil - fuels growth and aids ecosystem optimization
  41. The policy, regulation and tax regimes will need to adapt and evolve to take platforms into account since they provide so much value to both producers and consumers
  42. Future trends where platforms can disrupt - industries with non scalable gatekeepers (publishers), highly fragmented industries, important information close to the source (media and telecom , extreme information asymmetry
  43. Industries likely to fight off platforms - Industries with regulatory control (healthcare), high failure costs (banks), high resource industries (energy and mining)
  44. Education, finance, healthcare and renewable energy may be the most ripe for disruption by platforms
What I got out of it
  1. One of my favorite business and technology books of all time - shows the power of platforms and a roadmap to build or analyze them
Choudary's website is worth checking out and has a good introductory page