Chapter 5
When to Test and What to Test
Every studio knows testing needs to happen. The real question is when and how much.
Some teams test too late and drown in bugs. Others test so randomly that they never see the full picture. The truth lies in rhythm. Testing is a heartbeat that keeps production stable, and when it beats regularly, everything flows better.
This chapter will help you find that rhythm. When to test, what to test, and how to adapt it to the realities of small and growing teams.
Why Timing Matters
Testing isn’t something you do after development. It’s something you do alongside it.
If you wait until the end, you’re testing the costliest version of your mistakes. If you start too early without structure, you’ll waste time checking the same things that you're gonna change anyway.
The goal is balance - the right tests, at the right moments, with a clear purpose behind each one.
Early tests prevent expensive surprises.
Mid-development tests keep the momentum.
Pre-launch tests protect your reputation.
Testing rhythm gives your project predictability - something that every creative team should crave.
The Testing Pyramid
Think of testing like a pyramid that narrows as your game grows closer to launch.
At the base are core systems - the foundations that must always work.
In the middle are player flows and blockers - what affects the player’s journey.
At the top are polish and edge cases - details that matter when everything else is stable.

The idea is simple: never move up the pyramid until the layer below it is solid.
You can’t polish what doesn’t work.
When to Test
Testing is a rhythm you build into development. Small checks now, deeper passes when the time is right, and constant awareness as the project grows. The earlier you start, the more predictable everything becomes. And the more consistent you stay, the fewer surprises reach your future builds.
During Development
Every new feature deserves full attention (at least) twice:
Once right after it’s pushed, and again when it’s fully formed.
Early on, do a quick sweep - push the feature, try the basic interactions, make sure nothing crashes. These early checks keep new bugs from sinking too deep into the build.
When the feature is more complete, run a more focused pass: test the main flow, the edge cases, the expected and unexpected behaviour.
Alongside feature checks, keep a flexible smoke test list that contains five or six things that must always work (FTUE , boot and loading, basic UI flow, movement, input, core loop etc).
Each project has its own quirks, but these basics stay stable across genres. A quick smoke test after every push or merge keeps the project on track.
For each Internal Milestone
Internal milestones are moments where your build needs to feel coherent, not necessarily perfect. The expectation is simple: what’s supposed to be implemented should work without major issues. This means:
The golden path should be playable.
Fully implemented features should be reliable.
General stability should hold (no crashes).
Movement, input, and core mechanics should feel solid.
Depending on the milestone, expand the checks:
FTUE flow
UI structure
Connectivity features
Full playability (if relevant)
Short internal playtests help here. Not long sessions, just enough for the team to feel the current shape of the game and catch anything that slipped through.
Note that sometimes you will be sharing milestone builds with stakeholders - investors, publishers, marketing. Having a clear build that you're fully confident in and that you're sure looks good is a godsend.
Alpha Stage
Alpha is where your build starts representing the full real game. Not perfectly, but getting there. At this point:
The golden path must work without disruptive issues.
All major features should be fully functional.
UI and FTUE should make sense.
Stability should be good. Not flawless, but reliable.
Network features (if the game has them) should behave predictably.
Minor issues are fine. Low-frequency stability problems are fine. Blockers are not. Alpha is about certainty: the experience works, and its shape is clear.
Beta Stage
Beta is the first time the game must withstand real user behaviour - creative, chaotic and unpredictable.
A player should be able to play from start to finish without major issues or placeholders. Stability must be high.
This is where you refine:
Balancing
Difficulty curves
The "joy" factor
UX/UI clarity
If possible, use analytics. Beta is one of the best opportunities to measure player behaviour, not just observe it. You’ll learn what's good, where players struggle and what needs adjustment before launch.
World-Wide Launch
This stage is all about polish and trust.
Every issue you accepted during Beta now needs resolution unless it’s truly minor. No major problems should be present.
No rough edges, broken UI, placeholders or broken loops.
Apply all the Beta feedback you can, clean up systems, and run tight QA runs.
At this stage, players will judge your game not as a test, but as a product.
Congratulations! You have launched your game!
Post-Launch
After release, QA shifts from prevention to protection.
Your job now is to maintain the stability and ensure updates never undo what you've achieved.
Focus on:
Fixing issues found by the community quickly
Running small regression checks before every patch
Use analytics whenever possible in order to track crash rates, session length, drop-off points, and progression issues
A launched game is a living system. QA should still be around.
Plan, Don’t Guess
Random testing feels busy, however it’s not efficient.
Plan small, repeatable test sessions that target specific goals.
Try this rhythm:
Daily: Quick smoke tests after new builds or pushes.
Weekly: Focused feature and regression testing.
Per Milestone: Broader stability and usability checks.
Before Launch: Full playthroughs and stress testing.
Each layer confirms the previous one still works. It’s a safety net that grows stronger over time.
How to Keep It Sustainable
Testing should never feel like a burden.
Keep it light, structured, and repeatable.
1. Create a reusable checklist.
List 15–20 core items that must be verified in every build. This prevents “I thought someone else tested that” moments.
2. Combine QA and playtests.
Don’t separate “fun testing” from “functional testing.” Playing the game naturally exposes both.
3. Track patterns.
If the same issue keeps reappearing, make it part of your regression checklist.
4. Keep communication flowing.
QA should never work in isolation. A 5-minute sync with developers every few days prevents misunderstandings that could cost hours later.
Closing Thoughts
QA timing decides how much peace you’ll have at the end of production.
When testing starts early and stays consistent, launch weeks feel calm. When testing starts late, it will feel like you're stuck on damage control.
Test what matters, when it matters.
A consistent QA rhythm is what will get you through a successful release.