Benchmark

Benchmark the user's pipeline performance against industry norms for their role, level, company size, and geography.

Workflow

1. Check data threshold

Application data uses the split layout defined in references/tracker-schema.md — read it before any read or write.

Read {user_dir}/CareerNavigator/tracker.json. Count the total number of applications (any status) from the summary rows — no detail files needed for this check. If fewer than 5:

"You need at least 5 applications to run a meaningful benchmark — you have {n} so far. Keep logging applications via /career-navigator:track-application and run this again once you have more history."

Stop here if below threshold.

If ≥5 but fewer than 10 resolved outcomes, proceed with a note that results are preliminary.

2. Load the stage history the conversion math needs

Every PIPELINE CONVERSION and TIMELINES figure below is computed from stage_history[], and tracker.json alone contains no stage history — it moved to the per-application detail files. Computing app → response, screen → interview, interview → offer, or days-to-response from tracker.json by itself silently yields zeros and reports the user as far below norm when they are not.

1. Read {user_dir}/CareerNavigator/tracker.json and take applications[].
2. Iterate every row and load its detail_file (relative to CareerNavigator/) for that application's stage_history[]. Skip a row only when its stage_count is 0 — that row genuinely has no stages to count.
3. Use the summary fields where they answer the question directly: latest_stage and latest_stage_date give the current funnel position and last movement date, and stage_count / notes_count / contact_count give volumes without a read.

3. Invoke analyst — Operation 4

Hand off to the analyst agent with:

• CareerNavigator/tracker.json (summary rows) plus the loaded applications/<application_id>.json detail files — pass both; conversion and timeline math is impossible from summary rows alone
• The full CareerNavigator/artifacts-index.json
• The full CareerNavigator/profile.md
• Instruction to run Operation 4: Market Benchmark

4. Present the benchmark report

Use the output from the analyst to render:

**Benchmark Report** — {today's date}
Data confidence: {Preliminary / Directional / Moderate / High} ({n} applications, {n} resolved outcomes)

PIPELINE CONVERSION
  App → Response:       {user%}  (norm: {low}–{high}%  ·  {above/below/at norm})
  Response → Screen:    {user%}  (norm: {low}–{high}%  ·  {above/below/at norm})
  Screen → Interview:   {user%}  (norm: {low}–{high}%  ·  {above/below/at norm})
  Interview → Offer:    {user%}  (norm: {low}–{high}%  ·  {above/below/at norm})

TIMELINES
  Avg days to response: {n} days  (norm: {low}–{high} days for {company size mix})
  Ghosting rate:        {user%}  (norm: {low}–{high}%  ·  {above/below/at norm})

ATS SCORES
  Avg artifact score:   {n}/100  (competitive threshold: 70+  ·  {above/below})
  Lowest artifact:      {score} — {filename}

MARKET CONTEXT
  {1–2 sentences on geographic competitiveness and role-level supply/demand for user's target}

GAPS TO ADDRESS
  {Numbered list of the 1–3 metrics most below norm, with a one-line interpretation each}

STRENGTHS
  {1–2 metrics above norm — acknowledge what's working}

If a metric cannot be calculated (e.g., no offers yet, so interview → offer rate is undefined), show — and note why.

5. Suggest next step

Based on the lowest-performing metric:

• App → Response below norm: > "Your resume may not be clearing ATS or targeting JDs closely enough. Run /career-navigator:resume-score against your weakest-performing artifact."
• Response → Screen or Screen → Interview below norm: > "Your materials are getting in — the issue is later in the funnel. Run /career-navigator:pattern-analysis to see which ExperienceLibrary units are in stalled applications."
• Timelines slow across the board: > "Long timelines are common for enterprise roles. Check /career-navigator:track-application to ensure follow-ups are scheduled."
• ATS scores below 70: > "Run /career-navigator:ats-optimization to surface the highest-impact fixes on your active artifacts."