Synthetic sample / No candidate data

See what the reviewer receives.

Choose a role. Trace a finding to the work, inspect what the checks establish, and find a useful follow-up question. All four briefs use fictional examples; they contain no candidate data.

SeeTheWork / Evidence brief

Data & analytics

Synthetic sample

The metric is reconciled. The cause needs more evidence.

The sample separates duplicate records from the underlying revenue trend, checks the revised totals, and avoids turning a descriptive result into a causal claim. [04]

Checks data quality before interpretation

Supported in this example

[01][02]

Reconciles the metric definition

Supported in this example

[03]

Distinguishes observation from cause

Follow-up needed

[04]

All excerpts below are illustrative, including filenames and check results. They explain the review method and contain no real candidate data.

  1. [01]

    Data audit

    Illustrative

    Locate duplicate order records.

    quality_checks.sql

    SELECT order_id, COUNT(*)
    FROM orders
    GROUP BY order_id
    HAVING COUNT(*) > 1;
    What it supports
    The query identifies orders represented by multiple rows before aggregation.
    What remains open
    Repeated rows may be legitimate updates; the order lifecycle must also be understood.
    Back to competency evidence
  2. [02]

    Query excerpt

    Illustrative

    Make the counting rule explicit.

    revenue_analysis.sql

    ROW_NUMBER() OVER (
      PARTITION BY order_id
      ORDER BY updated_at DESC
    ) AS version
    -- Aggregate the latest paid record per order.
    What it supports
    The sample uses an explicit order-level rule instead of counting every update as new revenue.
    What remains open
    Refunds, currencies, and late-arriving updates need separate accounting rules.
    Back to competency evidence
  3. [03]

    Reconciliation check

    Illustrative

    Reconcile the reported total.

    reconciliation.md · illustrative result

    CHECK latest order IDs are unique: passed
    CHECK paid-order totals reconcile: passed
    CHECK null revenue rows are isolated: passed
    What it supports
    These illustrative checks connect the aggregation to the chosen definition and input records.
    What remains open
    Internal reconciliation does not establish that the business definition itself is correct.
    Back to competency evidence
  4. [04]

    Decision memo

    Illustrative

    Separate the finding from the explanation.

    findings.md

    “Duplicate updates inflated the earlier view.
    The corrected series still declines.
    We have not established why customers changed.”
    What it supports
    The memo distinguishes a reporting issue from a remaining business question.
    What remains open
    Segment, cohort, and external-event analysis would be needed before proposing a cause.
    Back to competency evidence

What would you check before attributing the remaining change to customer behavior?

Evidence for human review
Not an automated verdict

What would evidence like this help your team understand about its next hire?