See decisions, quality and risk in context.

Nardux reporting is designed to connect queue activity, quality review, appeals, escalations and improvement actions using definitions agreed for the engagement.

A practical view of the operation.

Every value below is fictional and exists only to demonstrate a possible reporting structure.

SYNTHETIC EXAMPLE · NOT CLIENT DATA

Weekly operations summary

Illustrative period: Week 04 · Example scope: one text-report queue

FICTIONAL DATA
COMPLETED1,840Example items reviewed
ENDING BACKLOG126Example open items at cut-off
QA AGREEMENT91.7%110 agreements / 120 sampled
APPEAL OUTCOMES1814 upheld · 4 reversed

Issue categories

Spam / promotion720Harassment460Impersonation310Other policy areas220Escalated / pending130

Action log

  1. POLICY

    Clarify repeat-spam example.

    Client Policy · due W05
  2. QUALITY

    Calibrate impersonation evidence.

    Nardux QA · due W05
  3. OPERATIONS

    Review launch-volume forecast.

    Shared · due W06
Illustrative sample only. It is not a live dashboard, client portal, performance claim or service commitment.

Inspect the fields behind the presentation.

Each file contains fictional examples only. The structure is illustrative and must be adapted to approved policy, authority, privacy, security, retention and reporting requirements.

QUEUE RECORD

Moderation decision sample

Shows policy version, reason, outcome, escalation reference and quality selection.

Download CSV
QUALITY RECORD

Independent review sample

Shows comparison outcomes, disagreement category, adjudication and action status.

Download CSV
DECISION CONTROL

Escalation log sample

Shows the trigger, authority, risk, decision needed, owner and follow-up.

Download CSV
REPORT EXPORT

Weekly operations sample

Shows definitions, denominators, limitations, fictional values and action owners.

Download CSV

Not production-ready by default. These files are not client data, customer results, contractual deliverables or a substitute for secure case-management tooling. Read the usage and data-handling notes.

Define the measure before interpreting the movement.

VOLUME

Intake and completed

Count rules identify the queue, included case types, duplicates, reopened work, period and cut-off time.

SERVICE HEALTH

Backlog and age

Open work is grouped by agreed age bands and priority, with outages, holds and missing client decisions identified.

QUALITY

Agreement and error

The report states the sample source, size, rubric, denominator, critical-error rules and whether cases were independently reviewed.

REMEDY

Appeals and reversals

Outcomes distinguish upheld, reversed, modified and escalated decisions and explain whether new evidence or policy clarification affected the result.

RISK

Escalations and incidents

Counts are separated by reason and status. Sensitive detail is limited to authorized recipients and appropriate channels.

IMPROVEMENT

Actions and decisions

Every follow-up records an owner, target date, status and link to the policy, training, tooling or workflow issue it addresses.

One score cannot explain the whole system.

Random samples can describe routine work but miss rare severe errors. A useful quality program can combine random, risk-based, new-reviewer, changed-policy, appealed and anomaly-driven samples.

Disagreement is classified before it becomes a performance conclusion: reviewer error, policy ambiguity, training gap, tooling issue or legitimate judgment range can require different action.

A changed rubric, denominator, sample or policy version is disclosed. Historical results are not silently made comparable.

Use outcomes to find operating causes.

APPEAL OUTCOMES

Was the original decision supportable?

  • Uphold, reverse, modify or escalate
  • Original and appeal policy versions
  • New evidence or context received
  • Independent review where feasible
  • User communication owned by client
TREND ANALYSIS

What is changing over time?

  • Issue and severity movement
  • Surface, source and reason patterns
  • Repeated ambiguity or error clusters
  • Launch and seasonality context
  • Known data completeness limits
INTERPRETATION LIMITS

What the report does not prove

  • Total platform prevalence
  • Absence of unseen harm
  • Universal reviewer accuracy
  • Legal or regulatory compliance
  • Cause from correlation alone

Questions about measures and delivery.

Do clients receive a live dashboard?

No custom client portal or live reporting dashboard is currently offered. The page sample is synthetic. The approved engagement defines the report format and secure delivery method.

Can you guarantee an accuracy percentage?

No universal percentage is promised. Quality depends on the policy, sample, content, reviewer readiness and method. Results are reported with the relevant definition and limitations.

Will reports include individual moderator rankings?

Not as a simplistic leaderboard. Individual feedback may be necessary for quality and training, but interpretation must consider complexity, policy conditions, escalation and wellbeing and follow the agreed privacy and employment approach.

Can we use our existing metrics?

Yes, after confirming definitions, data quality and decision purpose. If a denominator, sample or method changes, the report records the change rather than implying direct comparability.

Build reporting around the decisions your team must make.

Tell us what you review today, which definitions exist and where visibility is missing.

Discuss Reporting Needs