Leaderboards

Rank the thing your team actually cares about.

Most tools ship one leaderboard and hope it is your business. If what matters this quarter is resolved escalations, filings accepted or deals closed, rank on it this afternoon without asking us.

What you choose

Six decisions, and a board is done

FieldWhat it controls
metrics[]Which event types count, and at what weight
scopeRank individuals, or rank teams
team_filter[]The whole org, or a set of teams, expanded through the hierarchy
windowWeek to date, last 7 or 30 days, last week, all time, an explicit range, or a competition
sort_by, tie_breaks[]The comparator chain, ending in the subject id so ordering is total
visibility, limitWho appears, and how deep the board goes

Weights are what stop a board rewarding the wrong thing. Two closed deals worth $2,000 should not outrank one worth $400,000, so score the revenue and show the deal count beside it without letting it count. Most bad leaderboards are one number that could have been two.

Start here

Eight boards ready to go

Two of them work for any team at all, whatever they do. The rest assume you ship software. All eight are a starting point you can change, not a menu you are stuck with.

Speed

Still instant in year three.

A board that gets slower every month is a board people stop opening. Yours opens as fast on three years of history as it did in week one.

  • An all-time board is as quick as a 30-day one
  • Adding another year of history does not slow anything down
  • Averages that would mislead you are computed properly instead

No waiting

Invent a metric at 10am. Rank on it by lunch.

Send us something we have never heard of: escalation_resolved, filing_accepted, demo_booked, and it is rankable immediately. There is no list of supported metrics to be missing from, and no request to open with us.

Which also means a name you invent can never break anything, and we test for that specifically.

Visibility

You decide who appears, and it holds

Whichever you choose, it is not a preference the product can forget. Nobody sees a name you said they could not see, including through an export.

teams_only

Nobody is individually ranked

Individual boards return nothing, and say why, so the page can explain itself rather than showing a blank table. Team boards are unaffected, so the executive question is answerable without ranking a single person.

opt_in

Only people who asked to be

A person appears once they have opted in. Absence means not opted in, so silence reads as privacy rather than as exposure.

everyone

Open ranking

The default, because the restrictive mode left every new org’s first board blank by policy, on day one, with nothing explaining it.

In every mode a viewer sees their own row. Hiding someone’s own work from them is not privacy, it is a broken page.

Competitions

Run a month-long contest that stays won.

Put dates on a board and it becomes a competition, such as a quarter-end drive or a bug-squash week. When it ends, the result is settled for good.

Modifiers

Reward the work you actually want more of.

Better odds for a ticket closed inside an hour, worse for one reopened three times. You choose what earns the good cards, so the board rewards the behaviour you wanted rather than volume.

A deliberate omission

Gratitude is scored. It is not ranked.

Thanks given in Slack shows on a feed and a profile, and no board ranks it. Ranking gratitude turns it into a currency people farm, and the farming is louder than the thanks.

  • Capped at 20 gained and 2 lost per person per day
  • Gains and losses tracked separately, because the net hides them
  • Awarded to the recipient, not the sender

Rank the thing you actually care about.

Free for your first 50 people. No card, no sales call.