Key takeaways
- Session-based test management (SBTM) organises exploratory testing into time-boxed sessions, each with a charter, a session sheet and a debrief.
- It keeps the freedom of exploration and adds accountability: what was covered, what was found and how much time went into testing itself.
- The data from many sessions gives managers real progress reports without turning exploration into scripts.
Where this picks up
You've seen the basics of exploratory testing: charters, time boxes, notes. That works well for one tester. But with a team exploring for weeks, questions come up:
- Which areas have been explored, and how deeply?
- How much time is spent actually testing, and how much is lost to setup and blockers?
- When are we done?
SBTM answers those. Jonathan Bach described the method, developed with his brother James Bach, in an article in Software Testing & Quality Engineering magazine in 2000.
The four parts
1. Charters
A charter is the mission of a session. Good charters are specific enough to focus and open enough to explore:
| Too vague | Good | Too narrow |
|---|---|---|
| Test the app | Explore discount codes with expired, reused and stacked codes to discover pricing errors | Check that code SPRING10 gives 10% off |
Keep a charter backlog: a list of planned charters, prioritised by risk. New charters are added as sessions discover new areas.
2. Time-boxed sessions
A session is an uninterrupted block of testing on one charter, typically 60 to 120 minutes. No email, no meetings.
Uninterrupted matters. A two-hour session broken by five interruptions is five short, shallow sessions.
3. The session sheet
At the end, the tester fills in a short report:
| Field | Example |
|---|---|
| Charter | Explore discount codes with expired, reused and stacked codes… |
| Tester, date, duration | Ana, 12 Nov, 90 min |
| Time breakdown | Testing 60% · Bug investigation and reporting 30% · Setup 10% |
| On charter vs opportunity | 80% on charter · 20% following something interesting nearby |
| Test notes | What was tried, with which data, what was seen |
| Bugs | BUG-412: two codes stack when applied in a second tab |
| Issues | No test account with an expired loyalty card; couldn't test that path |
The time breakdown is powerful. If testers spend half their sessions on setup, the problem isn't testing, it's the environment.
4. The debrief
After each session, or once a day, the tester and a lead spend 10–15 minutes going through it:
- What did you cover, and what did you find?
- What got in the way?
- What's still worth exploring? Does that need a new charter?
- How confident are you about this area now?
The debrief is where coaching happens, and where the lead learns things a report wouldn't show.
Reporting from sessions
With session sheets in place, you can report progress like this:
| Area | Sessions done | Bugs | Confidence | Next |
|---|---|---|---|---|
| Checkout | 6 | 4 (1 critical, fixed) | High | 1 regression session after fixes |
| Discounts | 3 | 3 | Medium | 2 charters left: stacking, currencies |
| Account settings | 1 | 0 | Low (shallow) | 2 charters planned |
That's a progress report based on what was explored, not on how many scripted tests have run.
A team of four testers explored a new mobile banking release for two weeks. The lead introduced session sheets in week one.
After 30 sessions, the time breakdown showed that only 45% of session time was actual testing. Over a third went to setup: test accounts expired every day, and each tester rebuilt their data by hand.
The lead used the numbers to justify a script that created fresh test accounts each morning. In week two, testing time rose to about 70% of each session, and the team covered the remaining charters on schedule.
Lesson: without session data, the lost time would have looked like "testing is slow". With it, the real cause was visible and fixable.
Making it work
- Start with a charter backlog for the release, ordered by risk.
- Protect session time. Agree as a team that sessions aren't interrupted.
- Keep sheets short. A template that takes 5 minutes is filled in. One that takes 20 isn't.
- Debrief every day, even briefly. Unreviewed notes lose their value fast.
- Review the numbers weekly: coverage by area, time breakdown, new charters.
Do
Write charters around risks and information you need: "to discover how payment errors are handled".
Don't
Turn charters into step-by-step scripts. Then it's scripted testing with extra paperwork, and you lose what exploration is good at.
Check yourself
Session sheets show testers spend 40% of session time investigating and reporting bugs in one area. What might that tell you?
Show answer
That area is unstable: it produces many bugs, or bugs that are hard to reproduce. It may not be ready for deep testing. Talk to the developers: maybe it needs fixes or developer testing before more exploratory sessions, so testers spend their time finding new problems instead of documenting known ones.
