session-based-test-management-reference

Overview

Session-Based Test Management (SBTM) is the dominant framework for managing exploratory testing as a measurable, accountable activity. It was developed by Jonathan Bach + James Bach (1999-2000) and is documented at satisfice.com/sbtm.

The unit of work is a session - a time-boxed (60-90 min) chunk of focused exploratory testing against a mission stated as a charter. Sessions produce session sheets that capture what happened in a structured-enough format for management to aggregate, while leaving room for the tester to learn freely.

This skill is a pure reference consumed by manual-test-debrief and the tour-catalogues (exploratory-tours-reference).

For the canonical heuristic catalogs the tester applies during a session, see hiccupps-f-heuristic, sfdpot-exploratory-heuristic, fcc-cuts-vids-heuristic, crusspic-stmpl-heuristic.

When to use

• Onboarding a tester to SBTM vocabulary.
• Authoring a charter - what fields belong, what mission framing to use.
• Reviewing a session sheet for completeness.
• Building a session-tracking dashboard.

How to use

1. Write the charter - "Explore with to discover ". Keep it a mission, not a test-case list.
2. Time-box the session - 60-90 min, one charter, uninterrupted; ~2-hour hard cap before focus degrades.
3. Test and record - explore per the charter (applying tours + heuristics) and log Areas, Bugs, and Issues in the session sheet as you go. Full template: references/session-sheet-and-metrics.md.
4. Track TBS - note the Test / Bug / Setup split (plus Opportunity / Idle). Verify: T should land in the healthy 60-80% range before you trust the session; if T < 50%, stop and fix the cause (stabilise the environment or narrow the charter), then re-run the session before aggregating.
5. Close with PROOF - debrief Past, Results, Outlook, Obstacles, Feelings with the lead within 24h.
6. Aggregate on the dashboard - roll sessions up weekly (throughput, avg T%, charter progress) to drive strategy, not a bugs-per-session KPI.

The unit: a session

Per Jonathan and James Bach's SBTM paper (satisfice.com):

"A session is an uninterrupted block of reviewable, chartered test effort... A session may be short (around 60 minutes) or long (up to about 2 hours). Two hours is generally considered the upper limit because tester focus degrades."

Properties of a session:

• Time-boxed: 60-90 minutes typical; absolute upper bound ~2 hours
• Chartered: has a stated mission (what to explore, what to learn)
• Uninterrupted: no context-switching to other work mid-session
• Reviewable: produces a session sheet that the lead can review

The charter

A charter states the mission for one session. Per Bach:

Explore <area>
With <tools / resources>
To discover <information>

Example: "Explore the cart promo-stacking flow with manual sample inputs to discover discount-application bugs." More charter examples are in references/session-sheet-and-metrics.md.

Charters are not test cases. They state what to investigate and leave the how to the tester's judgment in the session.

A backlog of charters drives multi-session campaigns. The lead prioritises charters per the test strategy (test-strategy-author) and the risk register (risk-matrix).

The session sheet and metrics

Each session produces a session sheet whose sections are: Charter, Areas, session timings, TBS metrics, Data files, Test notes, Bugs, Issues (meta process problems), and the PROOF debrief. The TBS time-breakdown splits each session into T (test design + execution), B (bug investigation + reporting), and S (setup / overhead), plus Opportunity and Idle. Healthy ranges: T 60-80%, B 10-20%, S 10-15%; T under 50% signals trouble.

Full sheet template, the TBS definitions, and the cross-session dashboard metrics: references/session-sheet-and-metrics.md.

PROOF debrief

Per manual-test-debrief - the post-session debrief structure:

The lead reviews PROOFs with the tester briefly (5-10 min) before the session sheet is filed.

Worked example

A full 90-minute session (charter, timings, bugs, issues, TBS, PROOF outlook, and how the lead acts on it) is worked through in references/session-sheet-and-metrics.md.

Common confusions

Anti-patterns

Limitations

• Requires tester skill. The framework's value rises with tester expertise; it doesn't make bad testers good.
• Cadence overhead. Sessions + sheets + reviews add coordination cost - worth it for medium-+ teams, overhead for solo testers.
• Charter authoring is hard. Vague charters produce vague sessions; charter discipline pays off.
• Dashboard interpretation requires context. Raw metrics (sessions/week, bugs/session) can mislead without charter framing.
• Doesn't replace automated coverage. SBTM is for exploration, not regression - pair with automated suites.

References

• Bach J., Bach J. Session-Based Test Management (2000) - satisfice.com/download/session-based-test-management.
• Bach J. Exploratory Testing Explained - satisfice.com/exploratory-testing.
• Bolton M. Exploratory testing series - developsense.com.
• Hendrickson E. Explore It! (Pragmatic Bookshelf, 2013) - book.
• Full session-sheet template + TBS + dashboard metrics: references/session-sheet-and-metrics.md.
• Sibling references (heuristic catalogues the session applies): hiccupps-f-heuristic, sfdpot-exploratory-heuristic, fcc-cuts-vids-heuristic, crusspic-stmpl-heuristic, exploratory-tours-reference.
• Consumed by: manual-test-debrief.