Key Takeaways

• Intellias recommends choosing the right API type depending on the overall strategy and goals, as opposed to favoring one of them. Open/public APIs can increase adoption or provide an additional revenue stream via commercializing access. Partner APIs, which only give access to licensed partners, can strengthen strategic collaboration. Internal APIs, however, "improve connectivity between proprietary apps."
• Platforms consist of microservices, which are modular and reusable blocks that can be combined into various business applications thanks to APIs. Typically, as the platform grows, so does the number of microservices that need to be managed. Due to the complexity of orchestrating them, platforms require automatic coordination and management of their systems. Those systems should be organized into the following layers: core systems and services, data and analytics, integration, and partnership.
• According to Deloitte, a platform business should be capable of five types of analytics, including descriptive analytics (i.e., what happened?), inquisitive analytics (i.e., why has it happened?), predictive analytics (i.e., the ability to anticipate), prescriptive analytics (i.e., suggesting a response), and pre-emptive analytics (i.e., recommending further actions).

Introduction

Basic Principles of the Platform as a Business Structure

• The first best practice provides a high-level overview of elements that should be the base of any platform as a business structure. We included it among best practices because it was highlighted as important by Gartner and Deloitte. While Gartner's model is from 2018, it is mentioned as relevant in multiple recent articles by Deloitte. (Note: However, we couldn't track the original publication by Gartner, which might be paywalled, which is why the data comes from a secondary source.)
• Gartner selected five crucial elements of a platform as a business, including:
• IT systems supporting "back office and operational areas"
•  Customer experiences, such as customer employee portals and apps, and commerce channels
•  Data and analytics (described in detail in one of the sections below)
• Internet of Things (IoT), "which connects the physical assets of core and operational systems, in order to promote connectivity, analytics, and integration by monitoring, optimizing, controlling, and monetizing"
•  Ecosystems to support any connections external to the business, e.g., communities and marketplaces.

 

Source

 

• Deloitte expanded on the model by adding key technology elements that need to be factored in, which are:
• Cloud with an emphasis on an efficient and unified cost model, as opposed to multiple overlapping charge structures
•  Data and analytics (described in detail in one of the sections below)
• DevOps, which is "the materialization thread between the platform business strategy and innovation, and the platform and all its services"
•  Agility at scale, i.e., being able to iterate swiftly based on customer feedback
•  Partnership, which is essential to stay relevant with the fast pace of innovation
•  Cybersecurity, i.e., a strategy to prevent cyber fraud that involves all parties.

Having an API Strategy

• Several sources, including Intellias and Platform Thinking Labs, suggest that neither open nor closed API ecosystem is superior. The choice depends on each platform business's needs and capabilities. Both sources, as well as ABB, emphasize that an API strategy is crucial and/or highly beneficial, which is why it was included among the best practices.
• Platform Thinking Labs states that the successful use of APIs is vital to reaping the benefits of the platform business model. However, it notes that the benefits are not tied to a particular type of API. If used properly, all of them can grow reach, enable new business models, foster innovation, build an ecosystem, and facilitate the use of mobile channels.
• According to ABB, APIs "drive the platform economy," which underscores the importance of having an API strategy. Still, it cautions that building a strong strategy is a time-consuming task, which requires understanding market requirements and collecting feedback from all parties involved (i.e., partners, end users, and communities).
• Intellias also focuses on choosing the right API type depending on the overall strategy and goals, as opposed to favoring one of them. Open/public APIs can increase adoption or provide an additional revenue stream via commercializing access. Partner APIs, which only give access to licensed partners, can strengthen strategic collaboration. Internal APIs, however, "improve connectivity between proprietary apps."
• While each type has its unique advantages, Intellias believes that the most successful platform businesses integrate multiple types. It gives an example of Expedia, which is an aggregator thanks to combining internal and partner APIs, and had about $12.3 billion in 2020 revenue.
• Furthermore, it notes that Amazon successfully uses internal APIs to connect its microservices while also having a strong external API strategy. However, for many businesses, legacy architecture is too big an obstacle to follow Amazon's example.

4Ps of API Governance

• 4Ps of API governance is an approach to API governance recommended by MuleSoft to avoid conflicting goals in the governance process, with the final layer of platform governance dedicated to platform business models. It was included among best practices due to being mentioned both by MuleSoft and Intellias.
• According to MuleSoft, four categories (or "4Ps") of API governance include program, product, portfolio, and platform. Program governance is at the top of the structure. Responsible for overseeing the execution of the overall API strategy, it monitors progress on the overarching goals and adjusts them as needed, as well as champions the adoption of the API-led model.
• Product governance focuses on individual strategies and business models for each API product, such as assessing risks, establishing ownership, gathering feedback, and tracking the key performance indicators (KPIs).
• Portfolio governance is particularly relevant to platforms that have a lot of APIs. Some of its goals are to ensure consistent API development and security standards, identify low-value of redundant products in the portfolio, and pursue scaling opportunities.
• Finally, platform governance, as per Intellias, is "for a rapidly maturing API portfolio." At this stage, "key concerns shift toward automating your portfolio management and technically enhancing the operational environment for your APIs."

 

Source

 

Well-Structured Microservices/Modular Architecture

• Well-structured microservices/modular architecture was selected because companies in the field, such as Intellias and Quenit, describe it as essential to effectively managing the platform. Furthermore, Accenture outlines a model that is consistent with Intellias' recommendations.
• Platforms consist of microservices, which are modular and reusable blocks that can be combined into various business applications thanks to APIs. They may include native platform features and API-based components, among others. Typically, as the platform grows, so does the number of microservices that need to be managed. Due to the complexity of orchestrating them, platforms require automatic coordination and management of their systems.
• According to Intellias (and Accenture), modular architecture should be organized into the following layers:
• Core systems and services: The infrastructure for them can be custom, bought, or rented. The last option is recommended in the early phases, not only for the startups that may lack funds to develop their own solutions but also for established players, so they can test out different options.
•  Data and analytics: At a minimum, it should consist of a data ingestion mechanism, a data lake storage platform, and an analytics engine (either rule-based or machine learning-based).
• Integration: It requires strategies for identity, platform security, and API management.
• Partnership: It's an additional layer, which should be designed after the other three. It may include complementary services, features that allow partners to build solutions on top of the platform, and services in other verticals.

• Quenit provides Zalando as an example of a slightly different microservices architecture based on three key elements: urbanization, modular services, and development platform. All elements of the architecture are defined here. Zalando's overarching goal is "to enable autonomous teams to iterate fast on value delivery."

 

Source

 

Strong Data and Analytics Capabilities

• Strong data and analytics capabilities are included among the best practices because they were mentioned as crucial by multiple expert sources. Specifically, Deloitte described them as "the lifeblood of any platform business, allowing the business to perform key functions and function optimally." Additionally, Intellias notes that superior analytics can be a key differentiator for such businesses.
• According to Deloitte, a platform business should be capable of five types of analytics, including descriptive analytics (i.e., what happened?), inquisitive analytics (i.e., why has it happened?), predictive analytics (i.e., the ability to anticipate), prescriptive analytics (i.e., suggesting a response), and pre-emptive analytics (i.e., recommending further actions).

 

Source

 

• Furthermore, there are four key touchpoints at which data should be collected and consumed, which are onboarding, matchmaking, product fulfillment, and aftersale review.

 

Source

 

• The matchmaking stage typically uses machine learning algorithms, including clustering and the regression model. When combined, they can group customers based on certain characteristics and suggest products based on the results. For aftersale reviews, it is recommended to use Natural Language Processing (NLP) for sentiment analysis because reviews are usually submitted as text.
• A platform business should track key metrics during the entire customer journey. The overarching categories, along with sample metrics or data analytics models, are:
•  Acquisition (e.g., number of downloads and registrations)
•  Activation (e.g., customer churn analysis, activity/behavioral analysis, product holding analysis)
•  Retention (e.g., "customer churn analysis, minimum revenue per customer analysis, next best officer/product recommendation")
•  Referral (e.g., viral coefficient, i.e., "the number of new activities each customer refers to the platform")
•  Revenue (e.g., Customer Lifetime Value, Customer Acquisition Cost)

Additional Findings

• As noted in the introduction, Deloitte published a series of five articles on the platform as a business concept. Three additional parts will be published in the near future. The sources can provide further insights into structuring a platform business.

Research Strategy