Risk-based testing: check your scores against escaped defects
Take a QA lead whose team uses risk-based testing to plan each sprint. Every story gets a risk score, the team tests the high scorers first, and the low scorers…
You have read this

A Selenium to Playwright migration is the testing theme of 2026. TestGuild's survey of 40,000+ testers puts the shift at the top of the trend list, and Reddit's testing communities are openly recommending teams not start new Selenium projects in 2026. The question is no longer whether to migrate. The question is how to migrate without spending a full quarter on it.
This is the playbook. It covers when migration is worth it, when it isn't, the two approaches that work, the three hidden costs that derail teams, and a 4-week timeline you can adapt.
The honest framework first.
Migrate now if any of these are true:
Do not migrate yet if all of these are true:
The trap to avoid is migrating for fashion. Playwright is better for most teams in 2026, but "better" is not the same as "worth a quarter of your team's time right now." If your Selenium suite is paying for itself, leave it alone and migrate the next net-new test surface to Playwright instead.
There are two real approaches. There are also three fake ones (we will name them).
You wrap your Selenium WebDriver calls in a thin layer that, behind the scenes, dispatches to Playwright. Your existing test code keeps running. New tests can be written in either framework. You migrate incrementally.
Pros:
Cons:
Use when: your suite is large (500+ tests), the team is risk-averse, and you need to keep shipping while migrating.
You write a new Playwright suite from scratch, mirroring critical paths. Old Selenium tests run in parallel during the transition. When Playwright coverage is sufficient, you turn the Selenium suite off.
Pros:
Cons:
Use when: your suite is under 500 tests, the team can absorb a 4-6 week investment, and you want the full benefit, not a halfway state.
For honesty:
These are the costs that derail migrations after the kickoff energy fades.
Your Selenium tests almost certainly have a test data layer. Fixture builders, factory methods, database seeders, mock API responses. Most of this code is framework-agnostic in theory and framework-coupled in practice.
When you migrate, you discover that your fixtures depend on Selenium-specific timing assumptions, or on a custom retry helper that wraps Selenium's wait logic. Untangling this is the longest part of most migrations. Budget at least 30% of total migration time for fixture refactoring.
The mitigation: refactor fixtures first, before you touch the test code. Make the fixture layer framework-agnostic. Then migrating the test code is straightforward.
Selenium and Playwright have different CI footprints. Playwright tends to be faster per-test but heavier per-runner because it spins up browser contexts. Your existing CI parallelization strategy may not translate.
Specifically: if you are running 8 parallel Selenium jobs on small runners, you may need 4 parallel Playwright jobs on medium runners. Total wall-clock time can stay the same or improve, but your runner costs and your CI configuration both change.
The mitigation: prototype the CI configuration on a single critical-path test before you commit to the full migration. Measure runner cost and wall-clock time on real CI. Adjust your migration plan based on what you measure, not what the docs promise.
Every QA engineer on your team has muscle memory for Selenium. Every habit, every debugging instinct, every "I know exactly where to look" reflex is built on Selenium's behavior model.
Playwright's behavior model is different enough that the first two weeks of using it feel slower, not faster. Engineers will ask whether the migration was worth it. The answer becomes "yes" around week 3, but you have to get to week 3.
The mitigation: pair the most senior engineer with the most junior on the first migration sprint. Both will learn faster than either would alone, and the team builds shared muscle memory rather than depending on one person's expertise.
This timeline is for a suite of 200-500 tests using the rewrite approach. Adjust proportionally for larger suites.
If your suite is 200-300 tests and you stay disciplined about deletion, four weeks is realistic. If you keep all your tests, double the timeline. If you have more than 500 tests, plan in 4-week chunks.
The most underrated outcome of a Selenium-to-Playwright migration is the number of tests you discover were not worth migrating in the first place.
Delete during migration if a test:
A healthy migration deletes 15-25% of the original suite. If you are deleting less than 10%, you are migrating dead weight.
If your team is the one doing the migration, Nopmark's certification programs cover the test design fundamentals (ISTQB CTFL) and team-level skills assessment (TestAscend) that make migration projects succeed instead of stall.
The technical migration is the easy part. The hard part is the team-level discipline: deciding what to delete, agreeing on patterns, building shared muscle memory in a new framework. Strong fundamentals make those decisions faster.
How long does a Selenium-to-Playwright migration actually take? Four weeks is realistic for 200-500 tests with the rewrite approach, assuming you delete 15-25% of the suite and have one engineer with prior Playwright experience. Double the timeline for the shim approach because you have to do it twice.
Should I migrate everything, or only some tests? Migrate the critical paths and the test surfaces that are actively changing. Leave stable, low-maintenance test surfaces alone. Delete anything that has not earned its keep.
Will my CI bill go up or down? Probably stay roughly the same, but the shape changes. Playwright runs faster per-test but uses heavier runners. Prototype on real CI before committing.
Can I use AI to migrate the tests for me? AI is useful for the mechanical work (locator translation, syntax conversion). It cannot make the strategic calls about what to delete or what to redesign. Use it as a tool, not a lead.
What's the biggest mistake teams make? Migrating for completeness instead of for value. The point is not to end up with a Playwright copy of your Selenium suite. The point is to end up with a smaller, faster, more maintainable test suite that happens to be in Playwright.
Does this apply to Cypress migrations too? The strategic framework yes, the specifics no. Cypress and Playwright have different CI footprints and different page-object idioms. A separate playbook is coming for Cypress.
Ready to migrate a Selenium suite without losing a quarter to it?
Take a QA lead whose team uses risk-based testing to plan each sprint. Every story gets a risk score, the team tests the high scorers first, and the low scorers…
You have read this