Platform Operating Model
A platform pays off when product teams choose to use it, not when they are told to. Run it as an internal product with an owner, a roadmap and a feedback loop, or it turns into a ticket queue that slows delivery.
Part of Cloud, Infrastructure & Applications, in the Adopt mode. 8 sources, checked .
- of organizations have adopted at least one internal platform (DORA 2025 survey)
- 90%
- of organizations have adopted at least one internal platform (DORA 2025 survey)
- Google Cloud, DORA, 2025(opens in a new tab)
- lower change stability for internal platform users than for non-users (DORA 2024)
- 14%
- lower change stability for internal platform users than for non-users (DORA 2024)
- Google Cloud, DORA, 2024(opens in a new tab)
- higher productivity when developers can finish tasks without an enabling team (DORA 2024)
- 5%
- higher productivity when developers can finish tasks without an enabling team (DORA 2024)
- Google Cloud, DORA, 2024(opens in a new tab)
The problem
Internal platforms are now the default way large engineering organizations ship software. Google's DORA research found that 90% of organizations have adopted at least one platform and 76% have at least one dedicated platform team1. Adoption is no longer the open question. Governance is: who owns the platform, what it must provide, what product teams still own, and where security sits. DORA describes the leadership challenge as moving from adoption to governing many platform teams without creating new silos1.
The returns are real but uneven. In DORA's 2024 report, developers using an internal platform were 8% more productive and teams performed 10% better, yet throughput fell 8% and change stability fell 14% compared with teams without one2. Where respondents had to use the platform exclusively for the whole application lifecycle, throughput dropped 6%. Independence mattered most: productivity rose 5% when developers could finish their tasks without waiting on an enabling team2.
AI raises the stakes. DORA's 2025 report found that AI's positive effect on organizational performance is strong when platform quality is high, and that organizations treating the platform as an internal product see significantly greater returns1. The CNCF sets the goal of a platform as reducing cognitive load on product teams through self-service, consistent interfaces, documentation and secure defaults3. Team Topologies defines the platform team as one that provides a compelling internal product to speed up delivery by stream-aligned teams5.
Good looks like the upper levels of the CNCF maturity model: the platform funded as a product, users choosing it over alternatives, self-service interfaces, and measurement that combines usage and satisfaction data4. DORA names the patterns to avoid: the platform built without user research that becomes a ghost town, the ticket-ops team that works like a vending machine, and the ivory tower that pushes developers into shadow IT1.
What you get
- 01
Platform charter and ownership map
A written split of what the platform team, product teams and security each own, with a named platform product owner, decision rights and escalation paths.
- 02
Team interaction model
Which teams collaborate, which consume the platform as a service and which need facilitation, set out with Team Topologies interaction modes so handoffs are deliberate.
- 03
Paved roads and service catalog
A ranked list of golden paths and self-service capabilities drawn from developer interviews and ticket data, with security and change controls built into the defaults.
- 04
Platform scorecard
Adoption, satisfaction and delivery measures for each capability, tied to DORA's software delivery metrics, so you can see whether the platform is helping.
- 05
Funding and roadmap model
How the platform is funded as a product, how requests are prioritized against the roadmap, and how unused capabilities are retired.
Where it goes wrong
The failure modes we design against from the first week, and the control for each.
Building a platform nobody asked for
Teams build what they assume developers need, without user research, and the platform becomes a ghost town1. Start with interviews and ticket analysis, and fund only capabilities that have a named consuming team.
Turning the platform team into a ticket queue
When every request is a ticket, the platform team spends its time on one-off work and never builds self-service1. Track the share of requests met through self-service and set a target to raise it each quarter.
Mandating the platform before it is fit for purpose
DORA found a 6% drop in throughput where teams had to use the platform exclusively for the whole lifecycle2. Offer paved roads that are easy to choose, allow documented exceptions, and earn adoption before you require it.
Shipping faster through weaker quality gates
Platform users in DORA's 2024 data saw change stability fall 14%, and instability combined with a platform was linked to higher burnout2. Put automated tests, policy checks and rollback in the default path and track change fail rate per service.
Standards and rules
DORA software delivery performance metrics
Primary text on dora.dev(opens in a new tab)Use change lead time, deployment frequency, recovery time, change fail rate and rework rate per service to judge whether the platform helps, not to rank teams.
CNCF Platform Engineering Maturity Model
Primary text on tag-app-delivery.cncf.io(opens in a new tab)Gives a shared scale for investment, adoption, interfaces, operations and measurement, so you can agree an honest target level for the platform.
NIST SP 800-53 Rev. 5, CM-3 Configuration Change Control
Primary text on csrc.nist.gov(opens in a new tab)Defines controlled changes, security impact review and change records; a platform can enforce these in its pipelines instead of in a separate approval queue.
NIST SP 800-218 Secure Software Development Framework
Primary text on csrc.nist.gov(opens in a new tab)Sets secure development practices for software producers; the platform is the natural place to supply them to every team as defaults.
How to measure it
- Share of product teams that use the platform by choice for new services
- Share of platform requests met through self-service rather than tickets
- Developer satisfaction with the platform, surveyed each quarter
- Change lead time and change fail rate for services on the platform versus off it
- Time for a new team to ship its first change to production
Questions buyers ask
What is a platform operating model?
It is the set of decisions that make an internal platform work as a service: who owns it, what it provides, how teams request and consume capabilities, how standards and security controls are applied, how it is funded and how success is measured. The technology is usually the smaller part. Most platform problems come from unclear ownership, a missing feedback loop or a mandate that arrived before the platform was ready.
Should we make our internal developer platform mandatory?
Usually not at first. Research from Google's DORA program found throughput fell where teams were required to use the platform for the entire lifecycle. A better approach is to make the paved road the easiest option, with security and compliance built in, and allow documented exceptions. Mandate specific controls, such as change records or secrets handling, rather than the whole toolchain. Adoption by choice is the clearest sign the platform is working.
How do we know if our platform team is working?
Look at three things together. Adoption: are teams choosing the platform for new work? Experience: do developers rate it as easy to use and responsive to feedback? Delivery: are lead time, deployment frequency and change fail rate improving for services on the platform? If adoption is high but stability is falling, the defaults are too weak. If satisfaction is low, the team is probably working tickets instead of building self-service.
Sources
Every figure on this page comes from the sources below. We checked each link on .
- 1.State of AI-assisted Software Development 2025(opens in a new tab)Google Cloud, DORA, 2025 · services.google.com
- 2.Accelerate State of DevOps Report 2024(opens in a new tab)Google Cloud, DORA, 2024 · services.google.com
- 3.CNCF Platforms White Paper(opens in a new tab)Cloud Native Computing Foundation, TAG App Delivery, April 2023 · tag-app-delivery.cncf.io
- 4.Platform Engineering Maturity Model(opens in a new tab)Cloud Native Computing Foundation, TAG App Delivery, November 2023 · tag-app-delivery.cncf.io
- 5.Key concepts and practices for applying a Team Topologies approach to team-of-teams org design(opens in a new tab)Team Topologies · teamtopologies.com
- 6.DORA's software delivery performance metrics(opens in a new tab)DORA (dora.dev), January 5, 2026 · dora.dev
- 7.SP 800-53 Rev. 5, Security and Privacy Controls for Information Systems and Organizations(opens in a new tab)National Institute of Standards and Technology, December 10, 2020 · csrc.nist.gov
- 8.SP 800-218, Secure Software Development Framework (SSDF) Version 1.1: Recommendations for Mitigating the Risk of Software Vulnerabilities(opens in a new tab)National Institute of Standards and Technology, February 2022 · csrc.nist.gov