Session-based exploratory testing per the Bachs' SBTM: authoring charters (Explore X with Y to discover Z), running time-boxed sessions (60-90 min), logging session sheets with TBS metrics, and closing with the PROOF session debrief (Past, Results, Outlook, Obstacles, Feelings). Bundles the classic exploration heuristics as references: Whittaker's seven test tours (Feature, Money, Landmark, Intellectual, Bad-data, Configuration, Garbage collector's), Kelly's FCC CUTS VIDS recon tours, Bach's SFDPOT what-to-vary catalog, Bolton's HICCUPPS-F oracle heuristic, and Bach's CRUSSPIC STMPL quality criteria - plus a ready-to-fill charter-card template and a session-sheet review checklist. Use when planning, chartering, running, debriefing, or reviewing an exploratory testing session, or when picking a test tour, heuristic, or oracle mid-session. For scripted manual test cases, use manual-test-script-author instead.
86
91%
Does it follow best practices?
Impact
86%
1.01xAverage score across 10 eval scenarios
Passed
No findings from the security scan
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.
Exploratory testing is not "clicking around". ISTQB defines it as testing where "tests are simultaneously designed, executed, and evaluated while the tester learns about the test object" (ISTQB CTFL Syllabus v4.0.1 §4.4.2, p.44). Human-driven testing has two modes, and this skill owns one of them:
| Scripted manual testing | Exploratory testing | |
|---|---|---|
| Designed | Before execution, by someone else | During execution, by the person running it |
| Fixed in advance | Steps, data, expected result | A mission and a time box only |
| Output | Pass/fail per case, signed and dated | Session notes, bugs, issues, new charters |
| Finds | Deviations from what was anticipated | What nobody thought to anticipate |
For the scripted side (step-tables, Gherkin, UAT scripts, checklists), use
manual-test-script-author. This skill covers the exploratory side end to
end: charter, session, heuristics, debrief, review.
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:
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; a full charter-card template (mission + areas + oracles + dimensions + tours + deliverables) is in references/charter-template.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).
Each reference is a distinct lens; a strong session composes 2-3, never all of them. Route by what the session needs:
| You need | Lens | Reference |
|---|---|---|
| A themed bug-hunting mission on a product you already know | Whittaker's seven tours (Feature, Money, Landmark, Intellectual, Bad-data, Configuration, Garbage collector's) | references/tours.md |
| Recon on an unfamiliar product, before any charter exists | Kelly's FCC CUTS VIDS eleven touring questions | references/fcc-cuts-vids.md |
| Ideas for what to vary when stuck mid-session | Bach's SFDPOT (Structure, Function, Data, Platform, Operations, Time) | references/sfdpot.md |
| To decide whether an observation is a bug | Bolton's HICCUPPS-F oracle catalog | references/hiccupps-f.md |
| Which quality criteria the session evaluates | Bach's CRUSSPIC STMPL thirteen criteria | references/crusspic-stmpl.md |
| A ready-to-fill charter card | Charter template + quality rules | references/charter-template.md |
| To close and file the session | PROOF debrief template + aggregation | references/debrief.md |
| To review someone else's completed session sheet | Lead's review checklist + verdict rules | references/session-review-checklist.md |
| The session-sheet template, TBS metrics, dashboard | SBTM artifacts in full | references/session-sheet-and-metrics.md |
The catalogs complement each other: tours give the session a theme, SFDPOT gives variations, HICCUPPS-F interprets observations, and CRUSSPIC STMPL frames the criteria. Bach + Bolton's Rapid Software Testing curriculum teaches them as one toolkit.
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.
Every session closes with a PROOF debrief - the five-section report from the original Bach & Bach SBTM paper:
| Letter | Stands for |
|---|---|
| P | Past: what happened in this session |
| R | Results: what was found |
| O | Outlook: what's left to test; suggested follow-up charters |
| O | Obstacles: what blocked the tester |
| F | Feelings: tester's qualitative read on quality |
The debrief flow: the tester fills the PROOF template within 30 min of session end (memory fades fast); every Result links a bug ID or a confirmed-working item; Outlook names the recommended next charter (the chain to the next session); the lead reviews the debrief with the tester briefly (5-10 min) within 24h before the session sheet is filed, using the checklist in references/session-review-checklist.md. Feelings is the load-bearing field teams skip - the tester's qualitative judgment is signal no metric captures. Aggregated debriefs drive the next round of charters: which areas are well-covered, which are stale, which were never touched.
Full blank template, a fully worked debrief, the quarterly rollup table, and the attention tracker: references/debrief.md.
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.
| Confusion | Reality |
|---|---|
| "SBTM = ad-hoc testing." | SBTM is structured: chartered, time-boxed, reviewed. Ad-hoc has none of those. |
| "Session sheet = test report." | Session sheet captures the journey; test report aggregates outcomes. |
| "Charters = test cases." | Charters state the mission; test cases prescribe steps. |
| "More TBS time = better tester." | Wrong: T% is environment + charter health, not tester skill. |
| "Sessions need to find bugs to be valuable." | Wrong: confirming coverage in a known-clean area is also valuable. |
| "Exploratory testing = unscripted." | Distinct concepts: SBTM is management; unscripted execution is technique. SBTM sessions can be more or less scripted. |
| Anti-pattern | Why it fails | Fix |
|---|---|---|
| Sessions > 2 hours | Focus collapses; quality drops | Hard 2-hour cap |
| Multi-charter sessions | Loses focus; metrics ambiguous | One charter per session |
| Sessions tracked in a spreadsheet that no one reviews | Effort wasted; learning lost | Lead reviews each session within 24h |
| Charter as "test X" | Doesn't direct exploration | Use Explore X with Y to discover Z |
| TBS metrics never used | Throwaway data | Aggregate weekly; drive process change |
| Bug-finding rate as KPI | Incentivises shallow bug-hunting | Charter throughput + dashboard health, not bugs-per-session |
| Session sheets without PROOF | Lose tester's qualitative signal | Always close with PROOF debrief |
| All heuristics in one session | Eleven shallow passes crowd out one useful one | Compose 2-3 lenses from the routing table |
manual-test-script-author - the scripted-manual sibling (step-table,
Gherkin, UAT, and checklist formats).bug-bash-facilitator - multi-cohort bug bash; inherits the PROOF
debrief format.