Platform business models allow enterprises to set up powerful industry-focused, cloud-based ecosystems for value exchange and innovation among participants. After their initial debut in the tech sector, they\u2019re now appearing across almost all industry verticals, including finance, healthcare, manufacturing, public sector, telecom, transportation and utilities.\nIn \u201cPlatform business models: A primer\u201d, I wrote about how these models are fast becoming the dominant business model for companies with advanced digital transformation strategies. The business model is particularly compelling because it converts traditional, linear value chains into multi-dimensional value networks and ecosystems. Organizations employing platform business models are often seeing a 4x multiplier in terms of average valuation compared to more traditional business models and McKinsey noted recently that \u201ccompanies pursuing offensive platform strategies yield a better payoff in both revenue and growth\u201d.\nSince building a thriving platform ecosystem doesn\u2019t happen overnight, in this follow-on piece we\u2019ll explore four key steps for implementation and take a look at how GE achieved this via their Predix industrial Internet platform. If you\u2019re considering building a platform business model yourself, or are already in process, here\u2019s four key steps which may help in your implementation approach:\n1. The foundation: establish the business and technical foundation for your platform\nWhile platform applications, just like other applications, need to follow a customer-focused, outside-in approach, it\u2019s critical at this early stage to get your foundation right in terms of the set of \u201cbase platform services\u201d that will underpin your application environment. In the platform world, these services often consist of core service layers for cloud infrastructure, security and governance, data and analytics, service integration and management, as well as a service catalog and industry-focused microservices.\nIt\u2019s important to note that, from an IT perspective, a platform is much more than just API management, and much more than just an integration layer, it\u2019s a complete environment for developing and deploying applications with value-added services from core partners that can speed time-to-market and generate cost savings from standardization and reuse.\nGE\u2019s Predix platform is an edge-to-cloud distributed environment for the Industrial Internet and serves their customers in the energy, aviation, and oil-and-gas businesses among other verticals. It\u2019s also the flagship offering of GE Digital \u2013 GE\u2019s software division that is on track to achieve 20%+ topline growth in orders annually for the foreseeable future.\nIf we look at the Predix platform, we find that some of their major service layers include the cloud foundation, security and compliance, data and analytics, as well as their industry-focused microservices.\n GE\n\nGE Predix Architecture\n\n\nAccording to Gytis Barzdukas, VP, Head of Predix Product Management, Licensing & Business Operations, GE Digital, customers had a clear need for an industrial application platform that provided a good balance of IoT functionality (e.g. data ingestion, time series data management, asset-centric analytics, etc.), scalability (e.g. various data types and volume), data security and compliance (e.g. largely regulated industries, tenancy management, role-based access controls) with an overall lower cost of ownership (e.g. OPEX over CapEx with public cloud infrastructure). Additionally, given the operational technology environment in industrial companies, they had to build and enable their customers with an edge-to-cloud distributed platform.\nIt\u2019s at this stage in your platform roadmap that you\u2019ll begin collecting requirements for your initial target applications, establish the vision and foundational architecture for your \u201cbase platform services\u201d, and determine initial core partners that are essential for platform operations. As an example, you\u2019ll likely need core partners for your \u201ccloud infrastructure\u201d layer, as well as for your \u201csecurity & governance\u201d and \u201cdata & analytics\u201d layers. The Predix cloud, for example, runs on AWS and will be on Azure later this year.\nThe business value of this stage is that IT can start to make build versus buy decisions at each layer, review available 1st party software assets, and develop early partnerships that will be key to success. This is the essential ecosystem plumbing that needs to be in place prior to your first application. \u00a0\n2. The internal app: align internal stakeholders and demonstrate platform value\nThe next step is to align internal teams and stakeholders, often working across business units, to establish a common vocabulary, roles and responsibilities, and to deploy an internal application to demonstrate the business value of the platform approach.\nAs you communicate with stakeholders, it will be essential to have a crisp definition of your platform and an equally crisp definition of exactly what\u2019s included within the platform \u201cstack\u201d. This taxonomy can help establish a common understanding across various teams so that it\u2019s clear what\u2019s in scope of the platform, the envisioned service layers, and where each team can play a role.\nAs an example, you may find that you want the IT team to own and develop the \u201cbase platform services\u201d and your industry business units to develop your industry applications that will reside on top and utilize these underlying, highly-reusable services and infrastructure. Funds will need to be allocated accordingly so that each group can succeed, with roles and responsibilities clearly outlined.\nAccording to John Gordon, former Chief Product & Marketing Officer, GE Digital, GE\u2019s approach was to prove the business value internally (i.e. \u201cGE for GE\u201d) and then to take to customers (i.e. \u201cGE for the World\u201d). GE had seen that there were no platforms in place to give industrial companies like themselves a shared view of industrial assets and equipment to drive decisions. They began by using data and analytics to change GE and then found that their customers were interested in deploying these same capabilities.\nAt this stage in your platform roadmap, you\u2019re deploying an internal application that can deliver value to the business units (such as asset optimization in GE\u2019s case), but which also allows internal developers and IT test the lower part of the platform stack to ensure services are consumable and interoperable and can help to speed development and time to market.\nThe business value at this stage is that you\u2019re demonstrating value internally and establishing the right organizational connections to adopt this strategy as a firm-wide initiative. IT can start to productize the platform early on, without building out the full ecosystem, and can start to achieve cost savings from standardization and reuse by avoiding re-inventing the wheel for each application.\n3. The external app: build success with initial apps for external customers\nThe next step is to offer initial applications to external customers. This provides a means to gain early feedback from customers and to fine-tune and adjust the business and technical approach based on real-world observations. \u00a0\nGE\u2019s approach for Predix was to gain success with existing customers such as Exelon in existing verticals, then to move towards prospects in the same verticals, and finally towards prospects in adjacent areas. They started with asset performance management (APM) as their initial application, targeting use cases related to wind turbines, pipelines and other heavy assets and were able to improve asset availability by 2-6%, reduce production losses by 30-40%, and lower long-term incident costs by 10-20%.\nAccording to Gytis Barzdukas, Head of Predix Product Management, what GE quickly learned was that customers needed tools for the lower end of the stack as well. Customers were still in the mode of how to build the apps and needed more help with developer tools. This gave rise to Predix Studio, a visually-oriented environment to help \u201ccitizen\u201d developers build applications to run on Predix.\nAt this stage in your product roadmap you\u2019re deploying single applications and starting to expose APIs for developers to enhance and extend these applications. Your platform approach is becoming more visible to customers, but an extended ecosystem of partners is not needed at this stage since the focus is on customizing and extending your initial applications.\nThe business value at this stage is that you don\u2019t have to wait for a rich ecosystem before taking the platform to market and generating revenues. You can turn on one application at a time and start to build momentum. Even one application can become the beginning of your industry cloud and allow you to experiment with this new business model with its associated pricing-, operating- and support-model.\n4. External scale: scale the ecosystem with more apps and services\nThe final step is to scale the ecosystem by introducing more 1st party and partner applications and associated services so that customers can select from a wider range of apps and services in the service catalog. More selection brings more participants to the table, creating a virtuous cycle. Your business model doesn\u2019t necessarily have to be an exchange model where you\u2019re consummating matches between buyers and sellers, producers and consumers, and on so. While that\u2019s one of the common models, another equally valid model is the innovation model whereby you\u2019re simply providing apps and services and opening up the ecosystem for industry participants and developers to innovate on top.\nAccording to Mr. Barzdukas, GE\u2019s approach to scaling their Predix ecosystem has been to house 3rd parties within their service catalog, allow these 3rd parties to brand services as their own offerings, and evolve their various service layers in parallel. That is, rather than build out each service layer sequentially, GE has purposefully evolved each layer simultaneously.\nThe Predix developer ecosystem went live in February 2016, in tandem with Predix general availability, and the Predix.io catalog has been mostly developer-focused around microservices. GE is now looking at ways to make ready-made apps more visible and plans to focus more on applications this fall. The focus is shifting more to enterprise buyers interested in applications.\nAt this later stage in your product roadmap, as you move into building external scale, you\u2019re introducing more ready-to-run applications, introducing more services within your service catalog, and onboarding more customers onto the platform.\nThe business value at this stage is that you\u2019re now benefiting from the network effects of the ecosystem and the customer value proposition is increased due to a larger number of apps, services and constituents (i.e. buyers and sellers). IT benefits from this virtuous cycle via the cost efficiencies of operating at scale.\u00a0\nCharting your course\nThese four key steps for implementation provide a pathway to scale, with business value delivered early on via initial applications that are focused on market needs and requirements which can be used to guide both business unit app developers as well as the IT team managing the underlying platform stack. In addition, these initial applications can be used to pull the platform along with them, so you can start small without requiring an extensive ecosystem from the start. \u00a0\nSome of GE\u2019s key learnings in this journey have been to start small, be focused and be flexible. It\u2019s important to start with a premise on how to monetize, but be ready to adjust course using an agile methodology and to receive customer feedback sooner rather than later. GE\u2019s FastWorks methodology has been one way they have maintained this faster, smarter, and more customer-focused approach.\nPlatform business models are a very different animal from traditional IT projects, they often involve the entire company and require multiple business units to come together and stay in close alignment. Excellent communication and collaboration is therefore a vital ingredient as well as having a shared understanding of the strategy, vision and end-game of the platform approach, its value for all parties involved, and a shared commitment to its success.