Platform Revolution

  • Need a product first, platform second. That’s how you build a side of the market
  • Underappreciated benefit of platform model (besides scale and unit economics and variety) is you get to run massive amount of small experiments with real world data on what customers want
  • Redraws the lines of the business - influence outside matters more than control inside
  • Moved from supply economy of scale to demand economies of scale
  • Two sided network effects -
  • Ways to attract
    • Price - can disappear
    • Brand - fickle
    • Network - virtuous cycle
  • Viral is attracting new, network is making it more valuable for existing
  • Need to curate to avoid negative network effects from undesirables crowding out others
  • 4 types of network effects
    • Cross side and same side effects, on both consumers and producers
    • Can be positive or negative on both
  • 4 types of companies, broadly: asset builders, service providers, technology creators, network orchestrators
    • Network orchestrators are by far the most efficient value creators (2x or 4x better ratio)
    • Far easier to scale network effects outside a firm, so focus of organizational attention inverts
    • Management of externally is
    • Functional integration and network orchestration
  • Platforms are complex and getting started is hard - focus first on the fundamentals. What are you trying to facilitate
  • 3 things that are exchanged
    • Information
    • Goods/services (value unit)
    • Currency (incl attention and reputation)
  • Design one interaction at a time
    • As easy as possible to exchange - the core
    • Well designed filters are critical - give users the value units that are valuable to them. Eg a good search fn
  • Pull, facilitate and match
  • Pull - attracting consumers is very different than pipeline biz. Chicken and egg.
    • Once they are there, make sure they can find value
      • Feedback loop is critical
      • Single user feedback loop - targeted value unit to that consumer
      • News feed is great feedback loop with likes/etc for producers
      • Lower barriers to entry
  • Platform must be simple and clean to let users and producers rely on it, no bloatware
    • Low variety and stable base
  • Integral architecture can be easier, but is much harder for external parties to build on. Network effects are much harder
  • Software is eating the world
    • V1 was improved digital pipelines
    • V2 is platform model
  • 3 ways to unlock value and create platform
    • Delink asset from value (let it be used by anyone who can find a use)
    • Reintermediation - more efficient pairing of value units to consumers..also can leverage social feedback and reputation
    • Market aggregation - centralized markets to serve dispersed markets
  • Designing for viral growrh
    • Pull not push strategies
    • Push doesn’t work bc it doesn’t activate and create feedback loops
    • Strategies that bake marketing and appeal into the platform, feedback loops that reinforce the right usage and add value
  • Incumbents vs upstarts - not about huge initial user base, but about designing the right interaction model
  • What works for one platform does not typically work for others
  • YouTube’s early focus was entirely on content creators
    • Created network of producers which is more sticky than viewers are
    • Know your value proposition and those of your competitors
  • Solving chicken and egg problem
    • Start with an existing pipeline business then open to external
    • Staging value creation - attract initial users who then create new value units
    • Focus on one side then use that to attract the others
    • Simultaneous onboarding in a burst
  • Example variations
    • Piggyback off another customer base or provider base (scraping Craigslist, ebay PayPal)
  • Seed the initial value by being the first provider
  • Incentives to key users
  • Exclusive access
  • Induce producers to bring their customers
  • Big bang strategy for simultaneous onboarding
    • Can do this in small networks too, not just globally
  • Micro market - bring people who are already likely matches (fb at Harvard)
  • Viral growrh requires 4 components
    • Sender
    • Value unit
    • External network - medium where something spreads
    • Recipient
  • Viral growth incentivizes every user to become a marketer and spread value externally
  • How to monetize the business
    • Cannot charge either side - would discourage entry, data etc
      • Unless you need to add friction to curate
    • Charge at deal completion - once value is proven
    • Or charge for something tangential
  • What value is created on the platform? 4 types 
    • To consumers - access to value created on the platform (apps, vids)
    • To producers - access to a market
    • For both - access to tools and services that facilitate interactions
    • For both - access to curation mechanisms
  • Plenty of times that monetization may not be possible - need to produce enough excess value through the platform that can be captured by the platform
  • More users is not always better
    • Meetup succeeded by charging organizers, making for better interactions
  • Find a way to monetize in a way that does not destroy network effexts
    • Charge for enhanced curation
    • Enhanced access to users (broader reach, sponsored content)
    • Subscription
    • Charge for completed transactions
    • Charge for advertising that is beneficial to both sides (eg LinkedIn recruiting)
    • Charge for enhanced tools that aid the interaction beyond the match
  • Who to charge
    • All members - when you want to curate the full network
    • Charge the side that values the interaction more
    • Subsidize ‘stars’
    • Charge the less price elastic group
    • Never take first money - only once both other sides have gotten value
  • Platform is basically an infrastructure to facilitate exchanges that are valuable
    • Start with the basic then later on over time
  • How to decide how open to be?
    • Three kinds - manager and sponsor particiaption, developer participation, user participation
    • Manager - different than sponsor. Apple, Android, linux all different
    • Devs - core, extension, data aggregators
      • Core - app store
      • Extension - apps; opened through apis, which can be closed . Salesforce gets 50 percent of revenue through APIs
      • Data -
      • If an extension app can become a platform itself, seek to own it
      • If a function is often reinvented by extensions, own it and make it a free api( (camera, record)
    • Users - compete openness usually doesn’t work, need to filter for quality
    • The unique value of a platform lies in it’s ability to make connections between users outside the platform itself
  • Governance
    • Don’t take more than your fair share
    • Don’t change rules in your favor
    • Always create value for your customers
  • 4 ways to build healthy markets
    • Increase safety - transparency, wualitu, insurance
    • Thickness - discoery
    • Minimze congestion
    • Minimize repugnant activity
  • 4 tools to control behavior
    • Laws - forbid smoking
    • Norms - mkae it uncool
    • Architecting - filters that reduce the impact
    • Market - taxes
  • Use all 4
  • Vibrant community is one of the most important things a platform can have
    • Promote the best members
    • Have a natural role progression to more authority
    • Trigger - action - reward - investment
      • Unexpected rewards are most habit forming
    • Let users have a hand in designing the system
    • Use social credit system to get community involved
    • Tell community what you are building and help them financially to avoid unnecessary competition 
    • Insurance exacmple - transaction benefit and freedom outweighs fraud
  • Metrics
    • Successful interactions and value for all users
    • Early -creation, sharing of value.
    • Later - retention and conversion
      • What portion is happening on vs off the platform
    • Which groups to incrnticiz, to bring onto the platform, to charge..
    • At maturity, innovate to enhance the value proposition against competitors — track producers and consumers participation closely
    • In phases
      • Startup -
        1. Liquidity - percentage of successful interactions is high. There is stuff to do! Percent of listings that lead to interactions within a time period.
        2. Matching quality - successful curation ; daily interaction percentage
        3. Trust - comfort with engaging in interactions on the platform
      • Growth
        • Producer to active consumer ratio
        • Producer - frequency, offerings, rate of failed interactions, fraud; can calculate LTV of producer
        • Consumer -
        • Interaction conversion rate
        • Consumer conversion to producer (issue attestation)
        • Odd one - distance to consumers. How integrated into lives?
      • Maturity
        • Measuring incremental changes to the product
        • Drive innovation - identify what developers are doing that’s valuable to absorb,
        • Signal from noise - don’t over measure/underprioritized
          • The more you measire, the less prioritized you’ll be. No vanity metrics
          • Actionable, simple, auditable
        • Resource allocation -
  • Strategy
    • In Porter’s 5 forces - build a moat to avoid competition, and push all others to compete; horizontal and vertical integration
    • Resource based view - control of a key resource
    • Sustainable advantage disappearing today with such rapid iteration and competition
    • Purpose of business is to create a customer
    • Can remake markets through network effects
    • Platforms turn businesses inside out, moving resources outside the walls
    • How platforms compete
      • Limit access to control more value
      • Foster innovation and capture it’s value. Eg sap dev community. Don’t seek to own all value, but do own the most important resources
      • Leverage data -
      • M&a - get to observe activity on the platform first
      • Platform envelopment - scan the horizons for adjacent platforms
      • Enhanced platform design
    • When advantage is sustainable
      • Supply economy of scale
      • Strong network effects
      • High switching costs
      • Catering to niche tastes/specialization