Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
While there are different approaches to implementing platform engineering with the Platform Engineering Capability Model, we've found from user research that most Microsoft customers fall into one of three customer segments. The customer segments are emerging innovator, strategic builder, and platform pioneer. This article walks you through a case study for a real customer in each segment. Company names are removed for privacy.
| Customer segment | Focus areas | Team size | Organization traits | Frequency |
|---|---|---|---|---|
| Emerging innovator | Rapid product development, automating manual processes, tackling inefficiencies | 1-5 (from DevOps or cloud infrastructure teams) | Identifies bottlenecks to improve delivery, beginning to realize need for organization-wide solutions | Second-most common |
The large insurance company realized that they had different infrastructure spread out across a large tech stack. There were multiple platforms and environments, and not many ways for developers to get started without relying on other teams. The company needs to reduce its growing workforce costs and have more standardized systems.
"The tipping point was pretty much straightforward. Given that we have multiple engineering platforms, multiple infra environments, including hybrid, no self-service developer portal capabilities, and three massive different stacks across our architecture, we had to bring in something like Terraform or an enterprise-level player like GitLab or GitHub. To manage end-to-end containerized platforms, we considered something like OpenShift, Ansible for workflow automation, and Backstage for the IDP. We did a massive evaluation to bring synergy across such a large tech stack...This is a very simple cost case of taking a bit of reducing the workforce or the developer base by 30%." - Chief Architect, insurance company
Challenge: Their primary challenges are rising cloud costs, compliance issues, lack of infrastructure engineering expertise, misaligned processes, and inconsistent team communication.
The insurance company plans to implement a standardized platform for all development and deployment activities to foster collaboration, speed up project setup, and simplify governance. The company will focus on growth across all five key platform engineering drivers.
Investment: The company is working with an external partner to implement platform engineering using a build, operate, and transfer (BOT) model. The external partner develops and operates the platform before transferring it back to the organization once they gain the expertise and capacity to manage it internally.
Adoption: There's significant internal resistance to adopting new practices. Developers don't want to shift from traditional methods to newer platforms and toolsets. To overcome this, the organization's leadership will push platform engineering adoption by tying it to productivity benefits and making it part of employee goals.
Governance: The enterprise planning and deployment (EPD) team is responsible for compliance and security. The centralized governance structure is deliberate to maintain high security and avoid vulnerabilities, making decentralization a challenge. There's a push towards democratizing deployment to developers, while maintaining governance protocols to prevent data breaches and ensure compliance. The goal is to strike a balance between security and agility.
Provisioning: The company will improve efficiency and reduce provisioning times by adopting a more integrated and self-service model. The potential reduction in time and resources spent on provisioning is a key driver for change.
Interfaces: The organization is adopting Backstage for its open-source flexibility, cost-effectiveness, and developer familiarity. Cortex was also considered. The decision to choose Backstage was driven by its cost and integration capabilities.
Measurements and feedback: It's been hard to move to a more meaningful feedback system because the company has a legacy measurement system and needs to align technical metrics to business KPIs. The company plans to work on aligning engineering efforts with business outcomes for a more integrated measurement approach. During this transition, the company will add tools and platforms that provide real-time analytics and observability.
| Customer segment | Focus areas | Team size | Organization traits | Frequency |
|---|---|---|---|---|
| Strategic builder | Collaboration, reducing redundant effort, shared solutions, standardization, cost management | 1-15 technical experts (developers and infrastructure specialists) | Leadership views developers as customers, partially integrated platform engineering features (self-service not fully adopted) | Most common |
The financial institution is at a mid-level of DevOps maturity, with some reusable central artifacts, standardized guidelines, and basic automation managed through code. The organization has reached a point where the size of its development teams and the diversity of its tools and practices create significant costs. The institution had thousands of custom tools used across the company and many complex organizational needs. The bank plans to offer developers a "golden path" to improve productivity that has flexibility built in while avoiding a one-size-fits-all approach.
"So the idea was that we will show them [developers] that this [golden path] is one way of doing the thing which will improve your productivity, but this is not the only way. Right? So we wanted to leave enough room for the developer to feel that they are empowered to make changes to this path that we are telling them. So when these paths are getting defined in the CTO team, the question is always, what are the paths to be defined which will work for the majority of the people in the bank? As I said, we are very complex. There are thousands of tools used across the bank. So one size fits all was always the biggest problem." - Executive Director, financial institution
Challenge: Their primary challenge is high costs and inefficiencies due to many different tools and practices. The company wants to make sure that the platform meets each team’s specific needs without causing problems or being an overly directive approach that could hinder adoption. The financial institution also lacks the expertise to develop custom platform solutions in-house.
The financial institution plans to focus on growth for three key drivers: adoption, governance, and provisioning and management. The bank wants to increase adoption of the platform engineering solution, better integrate governance, and build automated resource provisioning tools.
Investment: The financial institution has a central engineering team with 120 people spread across multiple locations worldwide. About 20 members make up a center of excellence (COE) team. The COE team rolls out engineering best practices, the platform, and DevOps practices across all the other business divisions.
Adoption: The platform engineering team focuses on enforcing policies set by the COE-team to guide engineering operations. The company also plans to motivate teams with publicly visible performance metrics. Overall, the bank wants to grow platform usage without relying on strict directives and metrics. However, they face challenges in upskilling the COE team to handle the variety of technologies used across engineering teams. A major obstacle is the concern that the platform might not meet individual teams' specific needs, potentially causing problems.
Governance: The platform engineering solution is an internally developed portal that acts as a central hub for developers, offering tools, guides, coding standards, and videos. The solution includes a quiz on minimum enterprise requirements (MERS) to ensure compliance before coding begins. The portal features a version of Stack Overflow for support, certified engineer profiles, and an onboarding journey to familiarize new developers with standards and tools. The company plans to streamline resource management and integrate governance into the development lifecycle, removing bottlenecks and attracting top technical talent with a modern toolset.
Provisioning: The COE team created "happy paths" for developers to boost productivity while maintaining flexibility. The goal is to offer an efficient path while allowing customization. When designing these paths, the CTO team aims to cater to most developers, but the bank's complexity, with thousands of tools in use, makes implementing a standardized approach. To scale the platform, the organization plans to implement automated resource provisioning to meet the diverse needs of their many engineering teams.
Interfaces: The internal developer portal was built primarily in-house. It's referred to internally as the DevOps portal, although it encompasses broader platform engineering functions beyond just DevOps. The portal serves as a centralized resource for developers and includes various tools, learning materials, videos, and trainings, as well as access to automation tools, self-starter guides, and containerized images for development. The portal is also integrated with security tools like Sonatype for code scanning and includes a registry of approved images and boilerplate code.
Measurements and feedback: The COE team is open to feedback and actively solicits it from engineering teams. Developer advocates and ambassadors also collect feedback on behalf of the COE team. The feedback process is mostly informal.
| Customer segment | Focus areas | Team size | Organization traits | Frequency |
|---|---|---|---|---|
| Platform pioneer | Treating developers as customers, managing platform as a product, strong developer experience | 16+ with specialized groups | Emphasizes accountability, empowerment, and innovation, promotes self-service and minimal context switching | Least common |
The software company is at a high level of DevOps maturity. The company's developers can self-provision cloud services in compliance with corporate guidelines. The company's large platform team with over 250 members successfully developed custom platform engineering solutions for the organization. The company plans to investigate how to continue improving their organization through platform engineering moving forward.
"How do we let our developers to deliver better software faster and (cheaper)?..We still need to investigate and invest in what could be that ideal solution which could work for our multicloud strategy...is there a one system that can scale to the diverse needs of the developers?..We are using generative AI and AI-driven solutions built in-house for the documentation and information discovery..Our goal is to make the developers accountable." - Senior Engineering Leader, software company
Challenge: The company's primary challenge is figuring out how to continue refining their already strong platform engineering practices in ways that save money, explore generative AI, increase adoption, and work for a multicloud environment.
The software company plans to focus on grow for four key drivers: investment, adoption, provisioning and management, and interfaces. The software company already functions at a high platform engineering level and wants to continue. The company plans to explore ways to integrate generative AI (with governance), increase platform adoption, and implement metrics-driven feedback loops.
Investment: The platform is funded and supported through a collaboration between the CTO and CFO offices. A dedicated platform team, formed by reallocating resources, includes 250 to 280 members like architects and engineers. The team oversees compute, runtime, CI/CD, tools, and observability, with a focus on cost efficiency. They're exploring generative AI for infrastructure scalability but recognize further research and investment are needed.
Adoption: Developers initially adopted the platform primarily for cost optimization and efficiency, driven by the pandemic. Internal campaigns, including hackathons, promote the platform, showcasing benefits like service maturity insights. The platform team has had difficulty convincing some teams to move from their existing setups to the platform.
Governance: The governance model for the platform is structured around a central platform team that manages core elements. Individual service teams contribute plugins. There's a review process for all contributions to verify that they align with organizational standards and meet broader needs. The platform team maintains a service catalog and service map to track metadata and dependencies, which help ensure accountability and resource management. Additionally, a dedicated governance body was established specifically for AI applications to manage their use and ensure compliance with standards.
Provisioning: The platform team provides a centralized yet flexible platform for resource creation, deployment, and management. The platform is built on Kubernetes and uses Argo CD for CI/CD. The tool offers custom-built templates and predefined workflows. The platform includes a developer home where users can manage their infrastructure lifecycle from provisioning to deployment. Teams contribute tailored plugins to enhance functionality. The goal is to manage multicloud infrastructure seamlessly with a scalable platform.
Interfaces: Developers use the developer home in the platform to manage infrastructure, provisioning, and their entire development lifecycle. The platform’s plugin-based architecture allows for customization, while generative AI enhances documentation and searchability.
Measurements and feedback: The organization gathers feedback through surveys and uses metrics like DORA (deployment frequency, lead time, change failure rate, and mean time to recovery) to assess platform effectiveness. These metrics are categorized into agility and stability to pinpoint bottlenecks and improve outcomes.