How scoring works
A score nobody can explain is a score nobody should act on. Every metric below comes from a field the team already fills in as part of normal work.
Throughput
How much work actually ships, measured against the median for the same role rather than a company-wide average.
Airtable · work items reaching Done per weekTurnaround
Median time from picking work up to finishing it, plus how quickly a creator draft becomes a live post.
Airtable · In progress → Done, Draft → LiveOn-time delivery
Share of work finished before its due date. The single most reliable predictor of whether a client stays.
Airtable · completed_at vs due_atRevenue impact
Commercial weight of the work — margin on campaigns led or worked, attributed by contribution.
Airtable · campaign marginQuality
How often work clears review first time instead of bouncing back for rework.
Airtable · review cycles per itemUtilization
Hours logged against contracted hours. Scores highest at full but not over — sustained overrun is a burnout signal, not an achievement.
Harvest · hours ÷ contracted hoursWeighting by track
Heaviest weight in each row is emphasised
| Track | Thru | Turn | Time | Rev | Qual | Util |
|---|---|---|---|---|---|---|
| Leadership | 0.10 | 0.05 | 0.10 | 0.50 | 0.20 | 0.05 |
| Operations | 0.15 | 0.15 | 0.25 | 0.18 | 0.17 | 0.10 |
| Engineering | 0.18 | 0.12 | 0.15 | 0.10 | 0.35 | 0.10 |
| Delivery | 0.28 | 0.22 | 0.20 | 0.10 | 0.12 | 0.08 |
| Growth | 0.10 | 0.12 | 0.15 | 0.40 | 0.15 | 0.08 |
| Support | 0.18 | 0.12 | 0.30 | 0.05 | 0.25 | 0.10 |
Deliberate choices
Weights differ by role
One blend for everyone would be unfair. An engineer's job is quality; a delivery coordinator's is throughput and turnaround. A track is derived from department and level, never assigned by hand.
Utilization is not linear
It peaks at contracted hours and falls off on both sides. Chronic under-use is a staffing problem; chronic overrun is a burnout risk. Rewarding overrun would make the metric actively harmful.
Throughput is normalised per role
Measured against the median for the same track rather than a company-wide average, so a part-timer is not compared to a full-time coordinator.
Volume alone is gameable
Throughput on its own rewards splitting one task into five. It is paired with quality, and its curve flattens past twice the median, so volume cannot run away with a score.
Scores are not public by default
Visible to the person, their department lead and admins. Rating colleagues on a public scale changes behaviour — people optimise for the number.
History cannot be backfilled
Airtable only knows the present tense. Every trend depends on a weekly snapshot row, so the job must run from week one. Momentum needs six weeks before it says anything.