Does the QBR metric use the right denominator?

Research question and decision boundary

This study asks whether a QBR percentage remains decision-useful when every eligible, excluded, missing, and late record is made visible. The unit is one metric claim tied to its population, inclusion rule, period, numerator, denominator, and source extract. That boundary prevents an account label from replacing the actual evidence. The intended decision is QBR metric definition and presentation; the study does not grade an employee, infer client sentiment, establish legal compliance, or promise retention, revenue, response time, or service quality.

Facts, interpretations, and decisions are recorded separately. A Philippines-based account specialist may gather permitted sources, normalize fields, prepare a neutral brief, flag exceptions, and coordinate an approved follow-up. Contract, pricing, refund, legal, privacy, security, access, scope, and sensitive client-commitment decisions remain with the accountable owner.

Why the apparent signal can mislead

A completion rate can rise when incomplete records disappear from the denominator. Mixing opened, due, closed, canceled, and carried-over items can produce a precise number that answers no stable question.

The practical safeguard is a claim ledger: source, observed fact, interpretation, uncertainty, authority, affected account work, owner, and next check. “Sent,” “discussed,” “green,” “urgent,” and “complete” are not self-proving states. Each needs a written definition and evidence appropriate to the decision consequence.

Methodology and sampling plan

Recompute each sampled claim from a frozen source extract. Record eligibility before outcome, retain exclusions with reasons, separate unknown from zero, and test how the result changes under disclosed alternatives. Reviewers should reproduce both counts independently.

Use the comparison: reported results under the approved denominator, all eligible records, complete-record-only filtering, and plausible missing-data scenarios. Include routine, incomplete, corrected, disputed, missing, and reopened examples. Set the observation window, population, inclusion and exclusion rules, source hierarchy, and stopping rule before reviewing outcomes. Keep unavailable evidence unavailable rather than reconstructing it from memory.

Have two reviewers independently code a meaningful subset with the same rubric. Preserve disagreements by field. Agreement supports reproducibility under this protocol, not objective truth; disagreement may expose an unclear definition, inaccessible source, or authority boundary. Retain negative cases and disclose changes before a later replication.

Evidence, privacy, and authority controls

Rank sources by authority, specificity, relevance, and freshness for the particular claim. A newer informal note may not override an approved record, while a direct authorized correction may supersede an older field. Preserve both the former and current state, who changed it, the effective date, and why. Do not average incompatible evidence into certainty.

Apply data minimization and least privilege throughout. Review only approved systems and fields, avoid copying personal information into working papers, restrict exports, and use named accounts. When evidence is inaccessible, record the limitation and owner. The Data Privacy Act and FTC guidance provide context, but qualified owners must determine obligations and approved handling for the actual client.

Analysis and reporting

Report counts, denominators, exclusions, missing fields, state transitions, contradictory cases, and reviewer disagreements before summarizing a pattern. Distinguish description from inference. If one state appears more often in the sample, say exactly that; do not claim it caused delay, satisfaction, churn, or a commercial result.

A client-safe brief should state the strongest supported observation, the material uncertainty, the accountable decision, the permitted next action, and the review trigger. Comparisons across accounts require stable definitions and comparable visibility. Where those conditions fail, publish separate case findings rather than a misleading pooled rate.

Framework interpretation

NIST CSF 2.0 offers governance and risk-management outcomes; SP 800-53 supplies adaptable control language for access, audit, assessment, and information integrity. GAO’s 2025 Green Book emphasizes preventive controls, quality information, monitoring, and documented change assessment. FTC guidance supports minimization, sensible access, retention discipline, and service-provider oversight. ISO principles support process and evidence-based decisions.

None of these sources sets an outsourced-account-management staffing ratio, service level, workflow duration, or performance benchmark. They do not prove that this method improves a client outcome. The buyer’s contracts, policies, system rules, approved procedures, and named decision rights govern. Framework language should clarify a control, never manufacture authority.

Limitations and replication

Material limits include small or convenience samples, inaccessible channels, informal decisions, inconsistent clocks, different contracts, holidays, geography, tool configuration, stakeholder availability, survivorship in retained records, and reviewer knowledge of outcomes. Public findings must not reveal client-confidential or personal data.

A later team can replicate the study by freezing the same definitions, selecting a new period without inspecting results, recoding a subset independently, reporting missingness, and documenting every method change. The narrow conclusion is that QBR metric definition and presentation becomes more reviewable when source, definition, authority, uncertainty, owner, and verification remain connected. That is an operating-control finding, not a guarantee.