The seven artifact categories
The library recognises exactly seven. The set is closed — nothing the practice produces falls outside it — and non-overlapping: nothing belongs in two. Each earns its place by a completeness test no other category’s test would catch.
Each category below carries four things: where it comes from in the two source methods, what it contains and why it earns its place, the place those two sources genuinely disagree, and a completeness test written so that you can apply it to this library rather than take our word for the result.
The divergences are the useful part. Four of the seven categories sit on a real disagreement between the BABOK Guide and the in-house method, and in two of them the external standard is the stronger of the two. Those are marked as such. A methodology page that resolves its sources into quiet agreement has removed the only content a practitioner could push back on.
Category 1 — Charter and Goals
Lineage — BABOK v3: 6.1 Analyze Current State (6.1.5 current-state performance), 6.2 Define Future State (6.2.4.1 business objectives, 6.2.4.3 constraints, 6.2.4.10 assumptions), 6.4 Define Change Strategy (6.4.4.1 solution scope), 2.3 requirements classification, 7.3.4.2 evaluation criteria and baselines · House Method: charter tier, operator goal and valuation projection, fixed-location root.
The charter states why the practice exists and how anyone would know it was working. It carries the need in the enterprise’s terms rather than any one party’s, the current state with its baseline measures, four to eight goals each decomposed into measure-bearing objectives, the declared scope in and out, the constraints, the assumptions marked as assumptions with what would falsify each, and the halt conditions — the specific observations that would stop the programme rather than adjust it.
It earns its place as the only legitimate anchor for every upward trace in the corpus. Without it, traceability terminates in nothing.
The library follows BABOK. BABOK requires a baseline captured from the current state before the change (6.1.5, 7.3.4.2, 10.28.3.3). The House Method has no current-state analysis step at all, so a “percent realised” figure carries a numerator and no defined denominator. The library imports BABOK’s baselining wholesale.
For each goal, find one measure. For each measure, find three things: a baseline value, the date it was captured, and the named instrument that captured it. Then find the OUT list in the scope statement. An empty OUT list fails. A measure missing any of the three baseline elements fails. A goal whose baseline cannot yet be captured must appear in the risk register as a declared measurement gap, never as a decorative number.
Category 2 — Decision Record
Lineage — BABOK v3: 3.3 Plan Business Analysis Governance (3.3.4.2 change control, 3.3.4.4 approvals), 5.4 Assess Requirements Changes (5.4.4.2 impact analysis), 5.5 Approve Requirements (5.5.4.1 authority), 7.5 Define Design Options, 7.6 Analyze Potential Value (7.6.4.2 costs, 7.6.4.3 estimate uncertainty), 6.3 Assess Risks (6.3.4.5 recommendation), 6.4.4.4 rejected options · House Method: decision record, governance record, governance brief.
A decision record carries the context that forced a choice, the options considered with each one costed and risk-assessed, the option taken, the value of the best option not taken, the declared uncertainty in the estimates, the consequences accepted, the named authority who decided, and a dissenting view.
It is the only category in which authority is recorded. Everything else can be authored; only this can be decided. It is also the corpus’s memory of paths not taken, which BABOK 6.4.4.4 treats as part of the strategy rather than an appendix to it.
This is where BABOK is stronger than the House Method, and the library says so plainly. BABOK 3.3.4.2 requires costed alternatives including ones the requester did not recommend; 7.6.4.2 requires opportunity cost, defined as the value of the best alternative forgone; 7.6.4.3 requires the estimate to carry its own uncertainty, and warns about the asymmetry that describes benefits as intangible while costs are hard numbers. The House Method’s decision schema requires alternatives with rejection rationale and stops there. The library adopts BABOK’s itemisation.
Count the alternatives carrying a cost figure or a stated cost basis. Fewer than two fails. Then look for the forgone value of the best rejected option, the uncertainty on each estimate, a dissenting view present even where the recommendation is strong, and a decider named separately from the author. A record whose alternatives were all obviously worse is a straw man and fails whatever it claims.
Category 3 — Need Record
Lineage — BABOK v3: 2.3 requirements classification including transition requirements, 2.4 stakeholders, 4.1–4.3 Prepare / Conduct / Confirm Elicitation, 7.1 Specify and Model Requirements (7.1.4.3 attributes, 7.1.4.4 levels of abstraction), 3.2.4.1 stakeholder analysis, 5.3 Prioritize Requirements · House Method: need intake, atomicity sizing, scan-before-create.
One unit of demand: who needs what, why, how the need was obtained and confirmed, its functional inclusions, its non-functional inclusions, its data-governance declaration, its transition obligations, its acceptance signal stated before any implementation exists, its priority with the basis declared, and its links up and down.
It earns its place as the only artifact where a person’s need enters the corpus in that person’s own terms. Everything downstream is a translation, and mistranslation is unrecoverable if the original was never preserved.
The sources diverge in both directions, and the library takes both. BABOK carries a full elicitation cycle — preparation, conduct across collaborative, research and experimental modes, and confirmation by cross-comparison against independent sources (4.2.2, 4.3.4.2) — and a per-stakeholder analysis of attitude, decision authority and influence (3.2.4.1). The House Method has neither; it models producers, not affected parties, and presumes the need is already known. Conversely, the House Method requires every need to carry non-functional inclusions and a data-governance declaration regardless of size, explicitly marked not-applicable rather than omitted. BABOK treats non-functional requirements as a sub-class with a supporting technique, which does not prevent an NFR-free set from looking complete.
Every slot filled or marked N/A with a reason — silence fails. The acceptance signal predates any implementation. Non-functional and data-governance blocks present regardless of the need’s size. Transition obligations considered and their absence stated. The need scoped to one unit. The prior register searched for overlap before creation, with the search recorded. Provenance naming the elicitation mode and the source the result was checked against.
Category 4 — Rules and Definitions Register
Lineage — BABOK v3: 10.9 Business Rules Analysis, 10.11 Concept Modelling, 10.12 Data Dictionary, 10.23 Glossary, 7.1.4.1 data and information models, 3.4.4.4 requirements reuse, 5.2.4.3 maintenance for reuse · House Method: rules registry and its disjointness rule, data dictionary, rule / entity / reference schemas.
Normative statements about behaviour, and the shared vocabulary they operate on. Each rule appears as a conditions-and-outcomes statement in a constrained grammar; each entity carries attributes, relationships, cardinalities and the rules that govern it; controlled vocabularies form a distinct sub-type; a glossary binds the whole library.
Rules and definitions are kept as one category rather than two on purpose — a rule and the thing it governs are useless apart, and the joint test below is stronger than either alone. This is also the only content genuinely reusable across initiatives (BABOK 3.4.4.4), and therefore the only content whose value compounds.
Here the two sources hold opposite positions, and the library does not blend them. BABOK 10.9 keeps a business rule’s statement independent of how it is enforced, so the rule survives a change of mechanism. The House Method wants the enforcing mechanism named, on the grounds that a rule nobody enforces is a preference. The resolution is structural: the rule is stated declaratively, with no mechanism inside the statement, and the enforcement point is recorded as separate metadata alongside it. Both properties are then obtainable — the statement stays portable, and the unenforced rule stays visible, because its enforcement field is empty and an empty field can be counted.
Bidirectional closure: every rule governs at least one defined entity, and every entity is governed by at least one rule. Every term used in a specific sense appears in the glossary and is used consistently everywhere else. Every rule carries an enforcement-point entry, populated or explicitly empty. Identifiers stable and never reused on supersession. Domain rules and method rules held in separate registers, with no content migrating between them — a statement that would fit both is misclassified.
Category 5 — Quality and Acceptance Standard
Lineage — BABOK v3: 10.1 Acceptance and Evaluation Criteria (10.1.3.1 value attributes, 10.1.3.2 acceptance versus evaluation), 10.30 Non-Functional Requirements Analysis (10.30.3.2 quantification, 10.30.3.3 context), 7.2 Verify Requirements (7.2.4.1 quality characteristics, 7.2.4.3 checklists), 7.3 Validate Requirements, 10.37 Reviews · House Method: gate tier, gate ordering and bypass, mandatory non-functional inclusions, proof by refusal.
What “good enough” means, expressed so that a check can fail. The category holds the value attributes agreed before any criterion is written, acceptance criteria kept structurally apart from evaluation criteria, quantified non-functional thresholds with their populations and conditions, the checks themselves ordered cheapest-first, and the bypass procedure with what a bypass must produce.
Without this category the rest of the library is assertion.
The divergence is stark and the library follows the House Method. BABOK’s strongest verification instrument is a review checklist (7.2.4.3), which prompts a reviewer but cannot block anything, and BABOK nowhere asks whether the check itself is capable of failing. The House Method requires a blocking check to be witnessed correctly refusing a disqualified input before it counts as present at all. In the other direction the library takes a BABOK discipline the House Method lacks: value attributes (10.1.3.1) are agreed before criteria are derived from them, which is what stops acceptance criteria measuring the convenient thing instead of the valuable one.
Every blocking check has a recorded refusal — a check observed only passing is unproven and does not count. Acceptance (pass/fail, one candidate) and evaluation (scaled, ranking candidates) are not conflated. Every non-functional statement has a number and a population, or names the external authority supplying them. Value attributes predate the criteria. Checks are ordered by cost. Every bypass produced a record, an owner, and a remediation date.
Category 6 — Traceability and Structure Map
Lineage — BABOK v3: 5.1 Trace Requirements (5.1.4.1 formality and restraint, 5.1.4.2 relationship types), 7.4 Define Requirements Architecture (7.4.2 purpose, 7.4.4.1 viewpoints and views, 7.4.4.3 completeness, 7.4.4.4 relationship quality), 3.4 Plan Business Analysis Information Management (3.4.4.3 traceability approach, 3.4.4.6 attributes), 5.2 Maintain Requirements · House Method: trace matrix and coverage-gap reading, cross-reference integrity, two guards on any generated artifact.
The links, and the separate demonstration that the linked parts form a working whole. It holds the typed link table across goals, needs, rules, entities, implementing artifacts and checks; a small set of views each shaped for one audience’s concerns and none carrying everything; a coverage-gap section distinguishing structurally-correct empties from real holes; and the identifier scheme with its no-reuse rule.
BABOK 7.4.2 supplies the reason this is one category and not merely a matrix: traceability proves each requirement links to an objective, and it does not prove the set is coherent. Those are different properties needing different artifacts, and a library with perfect links can still be incoherent.
The House Method contributes what BABOK does not have — the recognition that this is the only derived category, and that a derived artifact silently diverging from its source is a defect class in its own right. So the map is regenerated on source change and a stale map fails the build. A hand-maintained trace matrix is a projection with a single guard, and it will rot without announcing it.
Every link typed, necessary, factually correct, unambiguous and described consistently with the declared conventions. No orphans in either direction: nothing without upstream authority, no goal without something serving it. Every item stated once and only once. Coverage gaps enumerated, each classified correct-by-design or real. And the map regenerates from source, with a stale map failing rather than publishing.
Category 7 — Evidence and Performance Record
Lineage — BABOK v3: 8.1 Measure Solution Performance (8.1.4.1 define, 8.1.4.2 validate, 8.1.4.3 collect), 8.2 Analyze Performance Measures (8.2.4.1 performance versus value, 8.2.4.3 trends, 8.2.4.4 accuracy, 8.2.4.5 variance), 8.3 and 8.4 limitation analysis, 10.28 Metrics and KPIs, 10.27 Lessons Learned · House Method: signal artifacts, observability, metrics catalogue, verdict discipline.
What actually happened, measured. It holds the measures collected with their sample sizes and currency, observed against target with the variance analysed rather than reported, the performance reading and the value reading kept separate, limitations split between those inside the solution and those in the surrounding organisation, and the evidence itself — captures, run outputs, query results — each stamped with a capture time and the artifact version it describes.
It earns its place as the only category produced by reality rather than by an author, and it is the only defence against the failure both sources name independently: a corpus that is internally immaculate and describes something that does not work.
The library takes the distinctions from BABOK and the reading disciplines from the House Method. BABOK 8.2.4.1 insists performance and value are different readings, and that a high-performing solution can under-deliver value; 8.3 and 8.4 separate “the thing does not work” from “the thing works and the organisation cannot absorb it.” The House Method has neither distinction, and its metrics catalogue is almost entirely process metrics — a library built on it alone would measure the machine and not the product. What it contributes back is two reading disciplines: absolute-zero events are reported as integers and never averaged, and a perfect score is read as a warning that the check may not be looking.
Every measure traces to a Charter goal and names the baseline it is compared against. Variance is explained, not merely reported. Performance and value appear as separate readings. Every evidence item carries a capture time and the artifact version it describes. Every status claim rests on the primary artifact, never on a secondary record’s assertion about it. And where the available measures are insufficient for a stakeholder to judge value, the insufficiency is recorded as a risk rather than omitted — unmeasurability is a finding, not a hole to paper over.
Where the artifacts themselves are
One artifact from Category 4 is published complete, worked through all fourteen sections of the standard template. The goals every category serves are set out with their measures, and what the corpus does not cover is listed in full. The remaining parts are indexed on the library page, which also states which are available on request.