BBM-STD-83-METRICS_AND_BENEFITS-V0.2 | Office standard | 4 November 2026 | V0.2 | Private controlled WIP draft
1 Decisions, metric dictionary and ownership#
This standard addresses. The practice operations lead owns operational decisions and benefits reviews; each metric has one accountable definition owner; the information analyst/data steward controls source and transformation lineage; an independent checker reconciles the actual snapshot; authorised release and receiving functions approve its specific distribution/use. Confirm actual role holders and scope under BBM-STD-55. A dashboard colour, software certification badge or generated report does not grant technical, statutory or project release approval.
Retain OIR, office register, EIR/LOIN, quality/corrective action, Manual 9, MIDP and TIDP owners. Generic TBP D15/D19 inform the method; they are not a second office decision register. The editable metric/dashboard/benefit forms contain empty templates and synthetic worked records, not a deployed dashboard or measured office performance.
Start with the business question and OIR decision, not a graph. Dictionary fields are MetricID/revision, definition and type (leading/outcome/estimate), OIR/PIR/EIR/objective derivation, decision consumer, owner/steward/checker, source dataset/keys/grain, formula/numerator/denominator/units, scope/time window/cut-off/time zone, exclusions/null rules, refresh/deadline, quality/uncertainty, planned target/source, threshold/action/owner, access/retention and adoption/review evidence. Preserve issued definitions and invalidate affected comparisons when scope or formula changes; never silently make a favourable trend by changing a denominator. Metric review is the office QMS monitoring, measurement, analysis and evaluation activity under ISO 9001:2026 (edition 6, 16 Sep 2026, replacing ISO 9001:2015); the clause numbers are taken from the licensed text and recorded in BBM-STD-68, which owns the QMS interface.
A survey of perception or self-reported practice (for example the share of respondents who say they follow ISO 19650) is context only: it is never used as an adoption, compliance or benefit metric, and no BenefitID or KPI target may cite it as its source. Benefit and adoption metrics come only from dated source records under §2.
Retain office cadence: Manual 9.1 weekly active-design quality dashboards; OIR 5.3 monthly requirement exceptions within 5 working days of cut-off and before the decision, quarterly usefulness/benefits review and annual full OIR review, with urgent triggers earlier. OIR-004 serves funding/forecast/benefit decisions; OIR-005 supports capacity and delivery decisions. Each actual reporting profile maps its metric cadence to those obligations, not an invented universal service level. Synthetic K83-01/02/03 are draft OIR-005 resource-prioritisation allocations via SYN-EIR83 R01; BEN83 is draft OIR-004 linkage. These new example allocations/targets require real owner adoption before use.
2 Populations, lineage, quality and calculation rules#
Each displayed value traces through SnapshotID/MetricID+revision/filter context -> query/transform/version -> raw dataset ID/revision/hash and approval/effective time -> stable source record IDs and event timestamps. Store source URI/owner, extraction method/time, reporting cut-off/time zone, code/dictionary/build revisions, join keys and cardinality, deduplication/aggregation rules, refresh dependency/results and independent comparison. A lineage diagram alone may omit manually joined sources; record actual row-level transformations and reconcile them. Report rendering time differs from source-effective time and last successful refresh.
Freeze source grain: one scoped element/requirement value for completeness; one first-submission identity for first-pass acceptance; one mandatory record for failure counts. Preserve resubmissions as separate events that cannot improve the original first-pass numerator. Reconcile duplicates, missing/invalid/excluded/not-applicable/withheld records explicitly. Only an approved applicability reason removes a record from a denominator; missing data stays applicable. Unknown is not zero. A zero denominator displays Not applicable with the scoped reason, never 100% or division error; an unexplained empty source displays Hold.
SYN-DATA83 R01 has 10 distinct optional records R01..R10 plus one duplicated raw R01 row. Keep one authoritative R01 by SourceEventID/revision, retain duplicate disposition; 11 raw -> 10 distinct. R01..R06 valid, R07/R08 invalid, R09 missing, R10 approved N/A under SYN-EXC83 R01. Thus applicable 9, valid 6, invalid 2, missing 1, N/A 1; 6/9=66.666...%, display 66.7%. Do not report 6/10 or omit the missing row. Preserve unrounded calculation and counts next to the display; rounded percentages are not release evidence.
SYN-SUB83 R01 has five eligible first submissions S01..S05; S01/S02/S03 pass, S04/S05 fail. A later successful S04 resubmission S04-R2 is excluded from first-pass numerator/denominator but retained in history. First-pass rate 3/5=60%. This is receiving-event quality, not an assertion that all delivered information is complete. SYN-MAND83 R01 has eight applicable mandatory records M01..M08; M01..M07 valid, M08 missing. Required failure count 1 triggers the existing mandatory-scope Hold regardless of optional completion or first-pass rate. No portfolio average can replace the source-specific technical gate.
Quality checks compare unique keys/joins, expected scope, count balances, datatype/unit/domain, source revision and cut-off, late/missing data, duplicate disposition and independent query/hand count. Record test-rule revision and actual failed evidence; do not trim outliers or redefine late delivery to improve a KPI. Aggregate ratios from matching summed numerators/denominators across disjoint scopes, not the arithmetic mean of percentages. Counts across periods may be stocks or events; do not sum repeated open-issue snapshots as new issues. Source absence or a many-to-many join that multiplies records holds affected results.
3 Targets, actuals, estimates and stale snapshots#
Separate planned target, measured result and estimate, each with scope, period, approval/source and uncertainty. A target is not a measured actual; an extrapolation is not realised benefit. Historical snapshots remain immutable; corrections have ChangeID, old/new source/definition/values, reason, owner/review and notification. Baseline, forecast and actual commitments from MIDP/TIDP remain distinct. Record late evidence after cut-off in the next snapshot or controlled restatement, not an undocumented overwrite.
The synthetic reporting window is [2026-10-09 17:00, 2026-10-16 17:00) Asia/Dubai; events at the end boundary enter the next period. SYN-SNAP83 R01 is a hypothetical weekly review snapshot with cut-off 2026-10-16 17:00 Asia/Dubai, source-effective/extracted 16:00, proposed approval reference SYN-SRCAPP83 R01 at 16:30 and planned rendering 17:00. Source approval is a fictional record only, not a signed office decision. Dictionary/transform SYN-TR83 R01 records the above raw-key rules; no native query is executed here. K83-01 has fictional target 80% for optional-data improvement, K83-02 target 70% first-pass, and K83-03 required failure count 0 under existing mandatory-scope rule. Optional targets are scenario assumptions, not ISO or office-wide acceptance criteria.
| Metric / dated synthetic source | n / N and worked result | Decision / action |
|---|---|---|
| K83-01 / SYN-DATA83 R01, 16 Oct 16:00 | 6/9 valid optional = 66.7%; 2 invalid, 1 missing, 1 N/A | Below fictional 80%; assign correction; mandatory acceptance separately checked |
| K83-02 / SYN-SUB83 R01, 16 Oct 16:00 | 3/5 first submissions pass = 60% | Below fictional 70%; review S04/S05 causes; do not count resubmission as first-pass |
| K83-03 / SYN-MAND83 R01, 16 Oct 16:00 | 1 missing mandatory of 8 applicable; 7 valid | Hold affected exchange until M08 corrected and independently retested |
Dashboard layout specification: header with SnapshotID, period/cut-off, source-effective/refresh times and scope; one value with n/N, units, definition/target/source link and exception status per metric; visible missing/invalid/stale/held counts and action owner; drill-through to source IDs and correction history. Use text labels plus colour, never colour alone; show units and uncertainty without misleading precision or truncated axes. Filter context is visible and versioned. No green roll-up is allowed while K83-03 has an unresolved mandatory failure.
SYN-REFRESH83 profile requires this review's effective source snapshot and successful refresh after 16:00 before 17:00. If the last successful source remains 2026-10-09, show Stale/Hold and its dated historical value, never relabel it current. If source approval is absent, show Unverified/Hold. If refresh fails, retain last good snapshot with its old timestamp and failure notice; do not display empty results as zero. Actual staleness thresholds, cadence and escalation derive from the adopted decision deadline. Verify source-to-query-to-semantic-model-to-report refresh and receiving values; successful UI reload alone is not source refresh.
4 Benefits baseline, costs, attribution and sensitivity#
BenefitID joins OIR/business decision, enabled BIM use, beneficiary, accountable owner, baseline/counterfactual, methodology, measurement window, comparable scope/complexity/quality, source observations, planned target, expected/realised classification, attribution limits, overlap exclusions, investment/recurring/transition/exit resources, risk, sensitivity and review. Compare feasible alternatives, including do-minimum; do not assign every improvement to BIM or automation when staffing, brief complexity, tools, learning or checking effort also changed. Record disbenefits and failed pilots. Keep output volume/quality/rework and effort comparable; faster incomplete output is no benefit.
SYN-TIME83 R01 is the synthetic source for Effort observations EFF83-01..03, baseline window 12-13 October 2026 and pilot window 14-15 October 2026; source-effective/extracted 16 October 16:00 Asia/Dubai, fictional approval SYN-SRCAPP83 R01 at 16:30. SYN-ASSUME83 R01 separately records training 12 h, QA/setup 6 h, future volumes 10/5 tasks and retained saving 4/2 h per task; effective/extracted 16 October 16:00, fictional assumption approval SYN-CASEAPP83 R01 at 16:30. Source approval and assumption/case approval are distinct references. BEN83 R01 and all its cases join both sources through SYN-TR83B R01 to SYN-SNAP83 R01; the transform computes 36/3=12 and 24/3=8, subtracts means for 4 h, then volume*saving-(12+6). The 2 h saving scenario is an explicit assumption, not another measured result. These dated references are invented training records, not signed approvals or actual time observations. SYN-BEN83 is a wholly fictional capacity example: matched-scope baseline task hours 12,10,14 (mean 12); pilot 8,9,7 (mean 8), all assumed equally checked. Difference 4 h/task or 33.333...% relative to 12 is a descriptive synthetic comparison, not causal proof or measured staff performance. Pilot sample 3 is too small/uncontrolled for an office-wide savings claim. For a future 10-task forecast, assume 4 h/task retained: 40 h gross capacity. Subtract 12 h training and 6 h QA/setup = 18 h incremental resource, net 22 h forecast capacity. These are hours only, no prices/rates/cost totals. Three observed fictional pilot tasks imply 12 h gross difference, which alone does not recover 18 h setup; do not report 22 h as pilot realised savings.
Sensitivity: 5 future tasks at 4 h yield 20 - 18 = 2 h net; 10 tasks at 2 h yield 20 - 18 = 2 h; 5 tasks at 2 h yield 10 - 18 = -8 h. The independent check must retain all cases, not only the favourable 22 h. If avoided rework is already inside task-hour differences, do not add it again as a second benefit. Released hours require evidence of productive reuse before claiming capacity realised; reduced expenditure needs separate actual finance ledger evidence. No hours become cash savings by multiplying an unverified rate. Actual commercial costing remains restricted to its authorised owner outside this no-price pack.
Real reviews include software/hardware/deployment, training/competence, production/checking, rework, maintenance, security, storage, transition and exit resource/cost categories, with explicit justified exclusions and source evidence. Quantified monetary analysis, discount/inflation/currency/price-base assumptions and uncertainty need their proper finance owner and adopted methodology; this standard sets method only, not financial advice. Do not discount or aggregate unlike time horizons silently. Qualitative safety/quality/user benefits remain described and evidenced where monetisation is unsupported.
5 Restricted access, corrective action and management decisions#
Apply Manual 9.4, Directory and BBM-STD-76 to data/aggregates, especially personal productivity/security/commercial information. Define audience/purpose, minimum fields, row/object restrictions, export/share/retention and actual allowed/denied tests. An aggregate may still identify a person in a small group. Staff-level productivity, time and competence metrics are personal data under the UAE PDPL (Federal Decree-Law No. 45 of 2021, effective 2 Jan 2022, plus any Executive Regulations listed on uaelegislation.gov.ae at project start): each such metric records its purpose, lawful basis and retention class per BBM-STD-90 §4, is reported only at the minimum aggregation the decision needs (team or task level by default; a group too small to protect identity is suppressed), and the access tests include a denied-viewer case for every staff-level view. Use synthetic identifiers in examples; do not publish staff rankings or public links simply because a dashboard platform supports them. Credentials stay in secure storage, never source files or forms.
Below-target or stale/unverified results create an ActionID linked to MetricID/SnapshotID, requirement/source/issue IDs, consequence, owner/due date, correction/retest/effectiveness and separate closure decision under BBM-STD-68. ACT83-01 corrects optional R07/R08/R09 before next review; ACT83-02 investigates first-submission defects S04/S05; ACT83-03 holds affected mandatory scope M08 until independent current-source retest. Targets prioritise actions; they never downgrade mandatory acceptance, safety, design or authority criteria. A lower issue count may reflect reduced detection; show checked scope and effort alongside trends.
Management review records continue/change/extend-measurement/stop decisions as an input to the ISO 9001:2026 management review held under BBM-STD-68; KPI review outputs are filed against that review record with source snapshot, baseline/case revision, benefit/quality/resource actuals, variance causes/attribution, risks, sensitivity and owner/date. Review at pilot completion, quarterly usefulness/benefits cycle, project close and the agreed operational follow-up interval with the asset owner. Actual next dates/holders are assigned by the project; none is fabricated here. Preserve original case and actuals separately, document lessons and approved definition/target changes, then review corrective-action effectiveness rather than closing when the chart turns green.
6 Editable forms, reconciliation and receiving tests#
Workbook Dictionary, Sources, Optional data, Submissions, Mandatory, Dashboard, Effort, Benefits, Actions, Tests, Acceptance have matching Blank forms. Dictionary and Sources together define grain, period/denominator, formula and dated source/approval lineage. Example results are static manually maintained specifications; recompute after editing, not a working BI refresh engine. Every displayed result joins an explicit snapshot/metric/source/transform and a dated synthetic approval reference; actual authorised source approval and installed receiving checks remain Not executed.
Reset tests independently. T83-01 counts duplicate R01: Fail; deduplicate by defined source event and verify 11 raw->10 distinct->9 applicable. T83-02 drops missing R09: 6/8=75% misstates 66.7%; Fail, restore missing denominator. T83-03 includes S04-R2 as first-pass: Fail, restore 3/5. T83-04 empty applicable scope: show N/A only with approved reason; never 100%. T83-05 source dated 9 Oct shown current 16 Oct: Hold stale; refresh or labelled historic view. T83-06 green optional average hides M08: Fail mandatory gate; show Hold and required retest. T83-07 claims 22 h realised from 3 pilot tasks: Fail forecast/actual boundary; label forecast and show pilot 12 gross versus 18 setup. T83-08 sums avoided rework twice: Fail, retain one benefit ownership. T83-09 source approval missing: Unverified/Hold, no fabricated approver.
Independent checker reproduces each value from approved dated raw records and transform/filter rules, checks exceptions/count balances, compares actual final dashboard/export and receiving drill-through/access, then records hashes/builds/refresh receipt/measurements and discrepancies. Definition owner reviews meaning; delivery authoriser and recipient separately approve scoped use. No displayed claim is usable as current until source/date/approval/reconciliation requirements pass. Office adoption, real datasets/permissions, installed refresh/dashboard tests, actual benefits/finance and management/receiving decisions remain activation holds; draft acceptance does not establish them.
References#
- Matrix, plan 36; office OIR 3.4/5.3, Manual 9.1, BBM-STD-51/54/68 and retained MIDP/TIDP. Generic D15/D19 are method prior art only; no text copied as office authority.
- Microsoft Power BI lineage and refresh inform dependency/refresh questions; no Power BI configuration or working dashboard is delivered.
- NISTA benefits assurance v2.1 August 2026 and official publication page, compared with office OIR decision/cadence. CDBB 2018 BIM benefits methodology is a prior-art lead confirmed by search; direct PDF 403 prevented renewed detailed review. No new detailed assertion relies on inaccessible content.
- ISO 19650-1 clause 5 and ISO 19650-2 clauses 5.6/5.7, IMI Framework (formerly UK BIM Framework, renamed May 2025) archived Guidance Part D ed2 are requirements/production/delivery reading pointers. ISO 19650-3, -5, ISO 7817-1 support lifecycle/security/LOIN context; BS EN 17412-1 is legacy predecessor where appointed. Licensed clause conformity not audited.
- NIBS v4 BEP provides use/goal context. UAE reporting/authority applicability remains with BBM-STD-57 and BBM-STD-85: DM building permit circulars index, ADM BIM Code of Practice, October 2020, DDA.
- ISO 9001:2026 and UAE PDPL, Federal Decree-Law No. 45 of 2021, checked 2026-11-04. NBS Digital Construction Report 2025 (research R6): one in three respondents define BIM as following ISO 19650, a perception figure cited here only to show why perception is not a metric. KPI results cannot establish statutory compliance. Sources checked 16 October 2026; no fees, prices or client names introduced.
Document history
| Version | Date | Change |
|---|---|---|
| V0.1 | 16 October 2026 | First controlled draft. |
| V0.2 | 4 November 2026 | closure: UAE PDPL rules for staff-level metrics via BBM-STD-90 in §5; ISO 9001:2026 monitoring/measurement and management-review link to BBM-STD-68 in §1/§5; reference labels corrected; survey perception excluded as a benefit metric (research R6); IMI Framework naming; links to OIR V0.3/Manual V0.4/MIDP/TIDP V0.3; spacing fixed. |