Understand what stands between a feature and its value
A feature can be licensed but not configured. People might not know how to use a configured feature, and use might not produce a safer or faster result. Calling every condition “adoption” hides the next work.
Availability, readiness, use, and results answer different questions about the same change. Keeping them separate helps you choose an effective response. Low use might call for a missing permission to be fixed, clearer instructions, or a different workflow. More training won't repair an integration, and a well-attended training session doesn't show that daily work became easier.
Follow one use case through all four questions
Ask four questions about one use case
Describe the task you want, not only the feature name. “Review delivery trends from complete deployment records” is easier to test than “adopt analytics.” Define the eligible projects and people before collecting a percentage or count.
| Rollout question | What to check | Likely next work when information is missing |
|---|---|---|
| Availability: Does the scope have the required tier, runner, integration, permissions, and settings? | A representative project check and links to settings or integration records. | Fix access, setup, or the data path. |
| Readiness: Can a user complete the task, recover from a common failure, and find help? | A completed task, support path, and points where unusual help was needed. | Improve instructions, training, ownership, or support. |
| Use: Do eligible teams follow the intended path in real work? | Recent GitLab events and a sample of representative work, including teams that didn't use the path. | Check workflow fit, usability, competing tools, and local barriers. |
| Results: Did the intended delivery, quality, security, risk, cost, or user result improve? | A starting point, result measure, safety checks, and a decision made after review. | Change the use case or trial rather than declaring success from usage. |
Keep the four conclusions next to each other and separate rather than combining them into one adoption score. Recheck them after a pilot and after each wider rollout wave.
Example: DORA metrics
Suppose a platform team wants to use GitLab's DORA reporting across a group.1
- Availability: Check that the team can access the reports and select a representative project. Deployment frequency needs successful production deployments. Lead time for changes needs merged merge requests connected to successful production deployments. The stability metrics need incidents opened and closed in GitLab. If another tool deploys software, check that it creates records through an integration or the Deployments API.
- Readiness: Ask the project owner to trace one merged change into production and explain how a production incident enters and leaves the GitLab incident workflow. Record who fixes missing records.
- Use: Compare recent deployment and incident records with the systems that performed the work. DevOps Adoption by group can show whether a group or subgroup recorded deployment activity in its reporting period,2 but it doesn't identify the contributing projects or prove that every DORA input is complete.
- Results: Name the decision the review should support. A populated chart shows that reporting works. Follow a finding backed by records and team accounts through to an improvement someone owns, then check what changed in both delivery speed and stability.
Example: security scanning
Suppose a team wants to find security problems earlier without leaving developers waiting for help. Choose representative projects and check their application risk, existing scans, runner capacity, and the time available for triage and remediation. A scanner rollout can fail through an overloaded response queue even when every scan runs correctly.
For one scanner, availability covers the required GitLab plan, configuration, pipeline, and runner capacity. Readiness shows whether developers and security staff can assess and route a finding in the vulnerability report3 or their existing tracking system. Use appears in recent pipelines and handled findings. Results might include better coverage or faster handling, checked beside pipeline duration, false-positive burden, and developer wait time. Consider wider security policy3 enforcement after the pilot shows a workable process.
Find the next piece of work
When a rollout stalls, check scope and eligibility, then availability, readiness, use, and results. This order helps narrow the investigation, but several gaps can exist at once and an earlier gap is not always the only cause. Training can help a readiness problem; it can't repair a missing integration or make an unsuitable workflow valuable. Record what supports each conclusion before choosing the next improvement.
Use reports to locate questions, then investigate
Different workloads might need different capabilities or timing. Compare projects with similar eligibility and setup, and keep exclusion reasons with the result. Avoid a universal target when the use case is unnecessary for some teams.
Product reports answer only part of the rollout question. DevOps Adoption marks a feature as adopted when a group or subgroup used it in a project during the last full calendar month. It doesn't measure the proportion of projects using the feature: one active project can produce the signal even when other eligible projects haven't started. It also doesn't assess user confidence, data completeness, workflow quality, or business results. Keep the event definition and reporting period beside any conclusion drawn from it.
Readiness findings should come from people who perform the work as well as platform or security owners. Platform teams can confirm that a scanner runs, Security can define how to prioritize its findings, and application teams can explain whether they can apply that advice in a normal change. A rollout works when those responsibilities meet in a usable path from finding to decision and follow-up.
Questions to ask
- Which projects and users are eligible for this use case, and why are others left out?
- What tier, setup, permissions, runner capacity, and integrations make the capability available?
- What task can a representative user complete to show readiness?
- Where does that user get help or request an exception?
- Which GitLab event shows use in real work, and what work can the report miss?
- What result should change after teams begin using the capability?
- Which safety checks could reveal an unwanted cost or delay?
- What points to the next improvement?
- What would make the team pause, change, or stop the rollout?
Further Reading
Notes
-
GitLab's DORA reporting is available in Ultimate. It calculates delivery metrics from GitLab's deployment, merge request, and incident records. ↩
-
DevOps Adoption by group is available in Ultimate. It summarizes adoption signals across groups; teams can also review activity directly in individual projects. ↩
-
The vulnerability report and security policies are available in Ultimate. Teams using available scanners on other tiers can assess and track findings through job reports, issues, and merge requests. ↩ ↩2
- Helping teams adopt a new workflow follows a rollout from a representative task into everyday use, with responses to common difficulties.
- Reliable DORA and value-stream data walks through the records and reconciliation checks for the analytics example.
- DevOps Adoption by group explains how to read adoption indicators and what their scope and update cycle leave out.
- Vulnerability management from finding to closure helps you establish a usable triage and remediation path alongside a scanner rollout.