Demonstrate Engineering Progress

The quarterly review is due and the only data is ticket counts. Ticket counts single out individuals. They do not describe the system. This guide shows how to prepare a quarterly presentation with Landmark. The presentation uses system-level trends, marker evidence, and engineer voice. These show direction and do not name individuals.

Prerequisites

• Getting Started: Map for Leaders -- install Map, migrate the activity schema, load your roster, and sync GetDX data.
• Getting Started: Landmark for Leaders -- install Landmark and confirm you can run npx fit-landmark org show.
• Authoring Agent-Aligned Engineering Standards -- define drivers and markers in your standard data. Landmark's health, evidence, and readiness views require drivers in drivers.yaml and markers in your capability YAML files.

The rest of this guide assumes Map's activity layer runs and holds data. If you want to explore with synthetic data first, see Trying the activity layer with synthetic data in the Map guide.

Confirm your data is ready

Before you build views for a quarterly review, confirm that the standard data, roster, and snapshots are in place.

Validate your standard data against the schema:

npx fit-map validate

Expected output (your counts will reflect your installation's standard):

Validation passed

Data Summary
  Skills       — 12
  Behaviours   — 6
  Disciplines  — 3
  Tracks       — 2
  Levels       — 5
  Drivers      — 4

If any errors appear, resolve them with the guidance in Authoring Agent-Aligned Engineering Standards.

Confirm that you loaded the roster and that the team hierarchy is visible:

npx fit-landmark org team --manager alice@example.com

  Team under alice@example.com

    Alice Smith           alice@example.com    Software Engineering / J080 (manager)
    Bob Chen              bob@example.com      Software Engineering / J060
    Carol Davis           carol@example.com    Software Engineering / J070
    Dan Park              dan@example.com      Data Engineering / J060

If the output is empty, re-run npx fit-map people push roster.csv with your current roster file.

Confirm that snapshot data is available:

npx fit-landmark snapshot list

  GetDX Snapshots

    MjUyNbaY                        2025-03-15   completed
    NzE4MmRk                        2025-06-14   completed

The third column is the snapshot status (completed or pending). The date is the snapshot's scheduled_for value.

If the output is empty, run npx fit-map getdx sync. Then run npx fit-map activity transform to ingest the latest GetDX data.

See system-level trends across snapshots

Quarterly reviews need context. The context says whether a score improves, declines, or stays flat. Before you open the health view, check how a specific driver moved over time.

Track a driver's trend across snapshots, scoped to your team:

npx fit-landmark snapshot trend --item code-review --manager alice@example.com

  Trend for code-review

    2025-03-15       72
    2025-06-14       78
    2025-09-13       81

The output shows the driver's score at each snapshot date. That makes the direction visible. Replace code-review with any driver ID from your drivers.yaml. The starter data includes code-review, incident-response, and deep-work.

Compare the latest snapshot against organizational benchmarks:

npx fit-landmark snapshot compare --snapshot MjUyNbaY --manager alice@example.com  # ID from 'snapshot list'

  Snapshot snap_2025_Q3

    Code Review        78    vs_prev: +6, vs_org: +8, vs_50th: +8, vs_75th: -2, vs_90th: -10
    Incident Response  65    vs_prev: -3, vs_org: -3, vs_50th: -3, vs_75th: -11, vs_90th: -19
    Deep Work          82    vs_prev: +1, vs_org: +10, vs_50th: +10, vs_75th: +1, vs_90th: -8

Each row shows the team's score. Signed deltas follow it. The deltas compare against the previous snapshot (vs_prev), the organization median (vs_org), and the 50th/75th/90th percentiles. Use the snapshot ID from npx fit-landmark snapshot list.

Build the health view

The health view is the core of Landmark's quarterly presentation. It joins driver scores, contributing-skill evidence, engineer voice comments, and growth recommendations into a single picture. It scopes that picture to a manager's team.

Run the health view for your team:

npx fit-landmark health --manager alice@example.com

  alice@example.com team — health view

  Drivers (2)
  ────────────────────────────────────────────────────────────
  #  Driver             Percentile  vs_org  More
  1  code-review        72nd        +5      -
  2  incident-response  48th        -3      -

  Recommendations (1 unique)
  ────────────────────────────────────────────────────────────
  - Carol Davis (working) could develop planning — for code-review (high)

The default output is a compact table organized by driver, followed by deduped growth recommendations. Each row shows:

• Driver name -- the driver ID from your drivers.yaml.
• Percentile -- the team's GetDX score position relative to the organization (e.g. 72nd).
• vs_org -- the signed delta against the org median (e.g. +5).
• More -- a hint when additional per-driver anchors are available through --verbose.

Landmark populates the Recommendations table at the end of the output when the installation includes Summit. It dedupes the table per (candidate, skill). Each line names the individual who could develop the skill. It also names their current proficiency and the driver the development serves.

Pass --verbose to switch to a per-driver paragraph layout. That layout discloses all percentile anchors, contributing skills, evidence counts, and the two most recent GetDX comments per driver:

npx fit-landmark health --manager alice@example.com --verbose

  alice@example.com team — health view

    Driver: code-review (72nd percentile)
      Anchors: percentile=72, vs_org=+5
      Contributing skills: task-completion, planning
      Evidence: 12 artifacts for task-completion, 8 artifacts for planning
      GetDX comments: "We've been catching more issues in review lately"
                      "Design docs are getting better but still inconsistent"

      ⮕ Recommendation: Carol Davis (working) could develop planning.
        (Summit growth alignment: high)

What the health view shows

The health view supports conversations about the system. It does not support conversations about individuals.

Driver scores are team-level aggregates from GetDX. Evidence counts show how many artifacts across the team match a skill's markers. They do not show which individual produced them. Landmark surfaces comments by keyword relevance to the driver. It does not attribute them to specific respondents.

When you present health data in a quarterly review, the narrative names where the system is strong. It names the trend. It reports what engineers say about the system. The data supports that narrative. Nobody has to name names.

Hear what engineers say

GetDX snapshot comments contain direct engineer feedback. Landmark surfaces these comments in two modes, both useful for quarterly preparation.

See comments themed by topic across your team:

npx fit-landmark voice --manager alice@example.com

  alice@example.com team — engineer voice

    Most discussed themes:
      incident              3 comments   "On-call handoffs are still rough", "Runbook coverage is improving but gaps remain"
      planning              2 comments   "Sprint planning feels more realistic this quarter", "Design docs are getting better but still inconsistent"
      testing               1 comments   "Integration tests saved us twice this month"

    Below-50th driver alignment:
      incident-response (48th percentile) — 3 incident comments

The manager view buckets comments by theme. It counts how many comments mention each theme. It shows the two most recent snippets inline per theme. It also highlights drivers that score below the 50th percentile where engineer comments align. There, sentiment matches the quantitative data.

This view helps a quarterly review because it grounds numerical scores in the team's own words. A low incident-response score paired with three incident comments is clearer than the score alone.

Check where evidence supports the standard

Evidence coverage shows whether the team's actual work produces artifacts that match your standard's markers. Two views help here. The first shows practice patterns across the team. The second shows the gap between derived and evidenced capability.

See practice patterns for your team:

npx fit-landmark practice --manager alice@example.com

  Practice patterns

    task-completion       matched: 12  unmatched: 4   total: 16
    planning              matched: 8   unmatched: 2   total: 10
    incident-response     matched: 4   unmatched: 6   total: 10
    sre_practices         matched: 2   unmatched: 5   total: 7

Each row shows how many marker-matched artifacts exist for the skill, how many unmatched candidates remain, and the total considered. Skills with high matched: counts have strong evidence. Rows with low matched and high unmatched signal where the evidence pipeline is light. Filter to a specific skill for detail:

npx fit-landmark practice --skill task-completion --manager alice@example.com

Compare what the standard predicts the team should be capable of against what evidence actually shows:

npx fit-landmark practiced --manager alice@example.com

  Practiced capability — alice@example.com (4 members)

    Task Completion       derived: practitioner   evidenced: 18 evidence rows
    Planning              derived: working        evidenced: 7 evidence rows
    Incident Response     derived: working        evidenced: 0 ← on paper only
    SRE Practices         derived: working        evidenced: 0 ← on paper only
    Architecture Design   derived: practitioner   evidenced: 0 ← on paper only

Each row aggregates across the team. The derived: value is the highest proficiency the team's role definitions imply for the skill. The evidenced: value counts the marker-matched evidence rows behind it.

Rows that trail ← on paper only flag skills that the standard predicts but that evidence does not yet cover. The inverse marker ← evidenced beyond role marks skills whose evidence outruns the derived role profile. This can mean the evidence pipeline has a gap. It can also highlight an opportunity to coach. Either way, the information is worth a mention in a quarterly review. It shows where the organization's definitions and actual practice diverge.

Verify

You demonstrate engineering progress without surveillance when:

1. Health view renders with data. npx fit-landmark health --manager alice@example.com shows at least one driver with a score, contributing skills, and evidence counts. No "No GetDX snapshot data available" messages.

2. Trends show direction. npx fit-landmark snapshot trend --item code-review --manager alice@example.com shows scores across multiple snapshots. That makes the trajectory visible.

3. Landmark surfaces engineer voice. npx fit-landmark voice --manager alice@example.com shows themed comments with counts. Comments align to drivers. Landmark does not attribute them to specific individuals.

4. Evidence backs the story. npx fit-landmark practiced --manager alice@example.com shows where the team's actual work matches the standard and where it does not. The result is system-level insight. It is not individual performance data.

All commands accept --format text, --format json, or --format markdown. Use --format markdown to produce output you can share in documents and presentations.

What's next