Quality assurance
Customer journey QA: what to test before a journey goes live
A practical QA checklist for automated customer journeys covering entry criteria, branching, timing, suppression, content, data and monitoring.
Tech Sierra · · 7 min read
A customer journey is software that runs on real people. It has inputs (profiles and events), logic (entry rules, branches, waits and caps) and outputs (messages across channels). Yet journeys are often tested the way a single campaign is: send a proof, check the render, launch.
That gap is where the most expensive MarTech defects live: duplicate sends, customers stuck in a loop, messages to people who unsubscribed, or a high-value journey that silently stops firing. Below is the checklist we use as a starting point when validating journeys before go-live.
1. Entry criteria
- Does the trigger event fire with the payload the journey expects? Validate against real event samples, not documentation.
- What happens when the same event arrives twice, such as on a retried webhook? Confirm re-entry rules.
- Are the audience filters evaluated at entry or at each step? Test a profile that qualifies at entry but not later.
- Are test, seed and internal profiles excluded or tagged appropriately?
2. Branching and logic
- Build at least one test profile for every path, including the default or “else” branch.
- Test boundary values: a cart worth exactly the threshold, a date exactly at the cutoff.
- Check how branches behave when the attribute they depend on is missing or null.
- Verify exit criteria. A customer who purchases should leave an abandonment journey at the next step, not after the next email.
3. Timing
- Confirm waits behave as intended across time zones and daylight-saving changes.
- Check quiet hours and send-time optimization rules against the business intent.
- Test what happens to profiles mid-journey if the journey is edited or paused.
4. Frequency, suppression and consent
- Do global frequency caps apply, and is the journey’s priority correct relative to other programs?
- Are unsubscribed, bounced and suppressed profiles excluded on every channel?
- Is SMS or push consent checked at the moment of send, not just at entry?
- Are holdout and control groups assigned and excluded consistently?
5. Content and rendering
- Render every message across the agreed client and device matrix, including dark mode and images-off.
- Test every personalization variable with real, empty and unusual values: long names, special characters, missing products.
- Confirm every dynamic block has a fallback, and that the fallback is acceptable to send.
- Validate links, UTM parameters and tracking for every variant.
- Check message weight stays comfortably under Gmail’s 102 KB clipping threshold.
6. Data and integrations
- Reconcile audience counts between the source system (CRM or CDP) and the platform.
- Confirm attributes used in logic and content are synced on the schedule the journey assumes.
- Test webhook and API steps for slow responses, errors and retries.
- If the journey writes data back, such as tags, scores or events, verify the writes and their side effects on other programs.
7. Go-live and monitoring
- Launch to a limited audience or a test cohort first, where the platform supports it.
- Agree the metrics that indicate a problem: entries, sends per step, error counts, unsubscribes. Then decide who watches them in the first 24 hours.
- Document the journey: purpose, entry rules, branches, owners and known limitations, so the next change doesn’t start from guesswork.
Making it repeatable
A checklist only helps if it’s used every time. Teams that run journeys at scale turn it into a QA gate with a test plan template, a standard set of test profiles, and evidence attached to the release ticket. That’s also where specialist QA capacity pays for itself: the people running the checks understand why each item matters, and they recognize a new failure mode when they see one.
If you’re building or tightening journey QA for your platform or your customers, talk to our QA team.
