Skip to main content
itzseo
Sign inSign up
Phase 38 · Anti-stack-ranking design

Recognize team wins
without burning anyone out.

Performance leaderboards usually destroy team culture. We rebuilt the pattern from the ground up — bands not exact scores, top half public only, opt-out by design, formula locked behind governance.

Start 14-day trial See pricing

Personal scoreboard unlocks on Growth. Department leaderboard + champions + badges on Agency.

Why most leaderboards fail

The Tesla-stack-ranking problem.

Naive leaderboards turn into status anxiety machines. The bottom 10% leaves. The top 10% sandbags to protect their spot. Everyone else either burns out or games the metric.

What we deliberately don't do:

  • • Publish exact scores publicly
  • • Force-rank the entire department
  • • Make ranks mandatory (every employee can opt out)
  • • Let the formula change without notice or governance
  • • Show bottom-half names — only your manager sees those
How it works

Three surfaces. Carefully scoped.

Private to you

Personal scoreboard

4-axis score (Output, Consistency, Depth, Flags), 28-day trajectory chart, weak-axis coaching tips from Claude. Only you see your exact number. Opt-out toggle hides you from the dept leaderboard entirely.

Top-50% only · Bands

Department leaderboard

Top half of the department publicly named. Bottom half hidden entirely. Bands (Top performer / Strong / Above average / Developing / Building) — not exact scores. Avoids weaponized rankings.

Weekly · Celebratory

Workspace champions

Every Sunday, 4 winners: Output / Consistency / Most Improved / Streak. Achievement badges fire for milestones (streaks, no-late, early-bird, no-blockers). Multi-winner — many people can hit the same badge.

The 4 axes

What we measure.

We don't score lines of code or tickets closed. We score the things that actually predict good agency work.

Output

/25 pts

Volume of meaningful work — tasks completed, articles written, audits run, keywords added. Anti-gamed by requiring task-level evidence (a closed task with no log entry doesn't count).

Consistency

/25 pts

Daily activity rhythm. Standups submitted on time, no quiet days, even pace across the week. A 10-task Wednesday doesn't offset 4 quiet days.

Depth

/25 pts

Quality + cross-project breadth. Were articles approved without major rewrite? Did the engineer touch 1 project or 5? Solo specialists and generalists both score well — for different reasons.

Flags

/25 pts

Reverse-scored. Late punches, blockers carried week-over-week, audit issues raised by a teammate that should have been caught earlier. Flags subtract — fewer is better.

Score = sum across 4 axes = 0-100. Full formula in the help center.

5 bands

Bands, not numbers.

Public-facing rankings show your band, not your exact score. A 78 and an 82 are both “Strong” — they shouldn't be competing for one rank.

Top performer
Top ~20%

Consistently delivers above expectation across multiple axes

Strong
~21–40%

Reliable, multi-project, low-flag

Above average
~41–60%

Meets expectations on most axes

Developing
~61–80%

Manager-only visibility — coaching opportunity

Building
Bottom ~20%

Manager-only visibility — likely new joiner or recovering from a tough month

Top 50% of each band is publicly visible on the department leaderboard. Bottom 50% (Developing + Building) is visible only to the department manager and HR, not to peers.

Phase 38.3

Champions + badges.

Multi-winner recognition. Most badges aren't scarce — many people can hit the milestone at once.

4 weekly champions
  • Output Champion — Highest cumulative output across the week
  • Consistency Champion — Highest cumulative consistency
  • Most Improved — Biggest delta in 28d rolling average vs prior week
  • Streak Champion — Longest active standup streak (reserved, v1.1)
7 achievement badges (multi-winner)
  • Streak 30 / 90 / 180 — Consecutive standups submitted
  • No Late Punch (4w) — Zero late marks across 4 weeks
  • Early Bird Week — Clocked in before shift start, 5 days in a row
  • Zero Blockers (4w) — Clean standups, no carried blockers
  • Cross-Project 5+ — Active in 5+ projects in 28 days
  • Perfect Flags (4w) — No flags raised across 4 weeks
Built-in safeguards

Things we did so this doesn't go wrong.

Opt-out toggle

Every employee can hide themselves from the public department leaderboard, including champions. Opt-out users still see their personal scoreboard privately. No questions asked.

Workspace-owner exclusion

The owner of the workspace is not snapshotted (they're not an employee). Naive systems include the founder in the ranking — we don't. Safe by construction.

Locked-formula governance

The scoring formula can't be changed unilaterally. Phase 38.5 requires 2-approver sign-off + 14-day advance notice to employees before any formula change. No surprise re-rankings.

No bottom-half name leaks

Department leaderboard explicitly hides bottom-half names from peers — visible only to managers + HR. Phase 32E.4 closed the cosmetic-leak path on the /hrms hub.

FAQ

Leaderboard questions.

Won't ranking employees damage culture?
Naive ranking does. The design here was deliberately built to avoid the Tesla failure mode: bands not exact scores, only top half public, opt-out toggle, formula locked behind governance with 14-day notice. We use this internally at Blogrator — it's been net-positive on retention so far. We also know it can still go wrong if a manager uses it as a stick instead of a coaching surface; that's a management problem, not a tool problem.
Can a manager fire someone based on their band?
ItzSEO doesn't make hiring or firing decisions. The leaderboard is a coaching signal — a way to see who's doing well, who needs support, who's improving. Firing decisions involve a lot more than 4 axes of activity data. If a manager uses bands as a sole performance review input, that's a process problem on their end.
What happens if I opt out?
You disappear from the public department leaderboard, the workspace champions list, and the achievement badges feed. Your personal scoreboard still works — you can still see your own trajectory privately. Opt-out is one click, no questions asked, no notification to your manager.
Can the formula be gamed?
Partially. Output rewards activity, so someone could pad task closures. But Consistency penalizes burst-then-quiet patterns. Depth requires cross-project breadth that's hard to fake. Flags subtract — gaming usually correlates with raising flags elsewhere. The formula isn't perfect, but it's harder to game than ‘hours logged’ or ‘tickets closed’.
Who picks the weekly champions?
Nobody. A Sunday 23:00 IST cron runs the picks deterministically from the week's snapshot data. No human judgment in the loop. If a winner opts out mid-week, the runner-up gets the badge. Audit trail in the LeaderboardBadge table.
What about new hires? Do they get scored on day 1?
The scoring window is rolling 28 days. New hires see partial scores until day 28, with a banner explaining the data isn't fully populated yet. Most Improved (which requires both this-week and prior-week data) won't fire for them until day 14.
Can I see the exact formula?
Yes — it's in the help center, plus the source code at lib/leaderboard/scoring.ts. We don't hide the formula. Hidden formulas would be a black box; published formulas mean employees can understand exactly what's being measured (and what isn't).

Recognize wins.
Without weaponizing them.

14-day trial of the Agency tier. Or unlock personal scoreboards on Growth.

Start free trial The full Agency OS