Most employee recognition dashboards are two things: a leaderboard and a big total. "1,240 props given this quarter, and here are the top ten." It looks like data. It is actually a highlight reel, and a highlight reel hides every problem you'd want a dashboard to catch: the people nobody recognized, the team that never thanks another team, the company value with zero mentions. A total going up feels great while forty percent of the company sits at zero.
This post is about what an employee recognition dashboard should show instead: eight metrics, each with a definition, how to compute it from the raw log, what a healthy versus unhealthy reading looks like with real numbers, and the action a bad reading triggers. Then how to build it in a spreadsheet, and what to leave off. The recognition data playbook makes the case for why these signals matter; this is the build sheet.
Everything below assumes one input: a log of recognition events with four columns — giver, receiver, timestamp, tag — plus a people table with each person's team and start date. If your tool exports that, you can build all eight.
1. Participation Rate
Definition: the percentage of active employees who gave at least one recognition this month.
Compute: count distinct givers in the month, divide by headcount. Healthy: 55% or more, holding steady. Unhealthy: 22% — a dozen enthusiasts while everyone else forgot the program exists. Action: a low reading is a friction problem before it's a culture problem. Cut steps out of giving (a Slack command beats a portal), and get managers giving first — Gallup attributes about 70% of the variance in team engagement to the manager.
2. Reach (Receiver Coverage)
Definition: the percentage of employees who received at least one recognition this month. Participation tells you who's giving; reach tells you who's being seen. The people at zero are the story.
Compute: distinct receivers divided by headcount, and — the important part — the list of names with a zero. Healthy: 85%+ over a quarter, with the zero list short enough to read aloud. Unhealthy: 70% participation paired with 45% reach: a lot of people giving, all to the same handful. Gallup and Workhuman found inadequately recognized employees are about twice as likely to say they'll quit within a year, so the zero list is a retention risk register. Action: route each zero to their manager as a prompt, and watch for whole teams on the list — usually invisible infrastructure work, not a performance issue.
3. Concentration
Definition: the share of all recognition going to the top 10% of recipients versus the remaining 90%.
Compute: rank people by props received, take the top decile's sum over the total. Healthy: the top 10% receiving 25–30% — stars do earn more, and that's fine. Unhealthy: the top 10% receiving 55%: a tenth of the company absorbing more than half the appreciation, month after month. This is a simple share split, not a Gini coefficient — cruder, but readable at a glance, which is what a dashboard is for. Action: don't ration praise for stars. Widen what counts as recognizable — glue work, mentorship, reliability — and prompt the people who never give. The full diagnosis is in fixing recognition inequality.
4. Cross-Team Share
Definition: the percentage of gives where the giver and receiver belong to different teams.
Compute: join both sides of each event to the people table, flag rows where giver_team ≠ receiver_team, divide by total events. Then break it out as a team-by-team matrix. Healthy: 30–40% of gives cross a boundary, and every team pair that actually works together shows a nonzero cell. Unhealthy: 8% overall, or an engineering-to-support cell that reads zero for three straight months despite shared incidents. Action: a boundary that goes quiet after a reorg or a rough joint project is where a relationship got damaged; pair it with a deliberate shout-out ritual. Breaking down silos with recognition data walks through reading the matrix.
5. Reciprocity
Definition: of all pairs of people where A recognized B this quarter, the share where B also recognized A.
Compute: build the set of (giver, receiver) pairs, count how many have their mirror pair present, divide. Healthy: 15–25%. Some mutual appreciation is what collaboration looks like. Unhealthy: 60%: that's a clique trading praise in a closed loop, and it's often the same people who dominate the concentration metric. Action: check whether the reciprocal pairs sit inside one team, then nudge that team toward recognizing the people they depend on. High reciprocity plus low reach is one problem, not two.
6. Tag and Value Mix
Definition: the distribution of tags (or company values) attached to recognition — and which ones never appear.
Compute: group events by tag, count, and sort. Add a row for each stated value with zero hits — the absences are the finding. Healthy: a spread where every core value appears at least monthly, with #teamwork or #customer-love leading. Unhealthy: #hustle at 41% of tags while #quality got two mentions all quarter, or a surge in #firefighting. Action: a values gap is a leadership conversation about what's actually rewarded; a firefighting spike is an operations problem wearing a thank-you costume. Celebrate the responders, then fix the fire.
7. Trend and Streaks
Definition: month-over-month change in participation and reach, plus the number of people currently on a giving streak (gave in each of the last N consecutive weeks).
Compute: the trend is this month's rate minus last month's; streaks come from grouping events by giver and ISO week and counting unbroken runs back from the current week. Healthy: participation flat or up three months running, with 20% of people on a streak of four weeks or more. Unhealthy: participation sliding 58% → 49% → 41% — the classic post-launch fade — with the streak count collapsing to near zero: giving only happens after all-hands reminders. Action: the streak count is your durability indicator, so reward it explicitly. Individual trend drops — someone who gave weekly for a year and has gone quiet — are the earliest attrition signal you own; that's the premise of detecting quiet quitting with recognition data.
8. Time to First Recognition for New Hires
Definition: for each person hired in the last six months, the number of days between their start date and the first recognition they received. Report the median and the count still waiting.
Compute: for every new hire, take the earliest event where they are the receiver, subtract start date. Anyone with no event yet shows as "still waiting" with days elapsed. Healthy: a median of 12 days, nobody past 30. Unhealthy: a median of 47 days with three people past 60. New hires have no track record and no allies yet; recognition is how they learn they're on track. Action: every name past 30 days goes to the hiring manager with a one-line ask. We cover the playbook in recognizing new hires in the first 90 days.
Build It in a Spreadsheet
Export the event log and people table onto two sheets, then build one pivot table with rows = people and four columns:
- Given — count of events where the person is the giver
- Received — count of events where the person is the receiver
- Distinct givers — count of unique givers among their received events (the popularity-versus-breadth check)
- Cross-team received — received events where the giver's team differs from theirs
That single table produces most of the dashboard. Participation is the share of rows with Given > 0; reach is the share with Received > 0; the zero list is a filter; concentration is the top decile of Received over the column total; cross-team share is the last column's total over all events. Add a second pivot with rows = tags for metric 6, and a third with rows = giver and columns = ISO week for streaks. Reciprocity and new-hire timing need a helper column each. Budget an hour the first time, fifteen minutes monthly after.
The limitation isn't the math; it's that a spreadsheet only refreshes when someone remembers, and "someone" is usually one person who eventually changes roles.
Or Get It Built In
Disclosure first: Propsly is ours. Every /props give in Slack records the giver, receivers, timestamp, and a required hashtag — exactly the log above. The free tier gives you the raw feed, leaderboards, giving streaks, and basic engagement stats. The Pro tier ($50/month flat for the workspace) adds the analytics views that map to this list: a top-10%-versus-bottom-90% concentration split, reciprocity, team breakdown, engagement gaps (the people at zero), top contributors, and tag stats, filterable by this month, last month, year-to-date, or all time. There's no dedicated new-hire time-to-first-recognition view — you'd still build that one from the data. Whatever tool you use, the non-negotiable is a log with all four columns.
What NOT to Put on the Dashboard
A dashboard shapes behavior as much as it reports it, so a few things belong on an admin's screen and nowhere else:
- The zero list, publicly. The most useful view here, and the most damaging if anyone but managers can see it. Publishing who received nothing turns a prompt into a shaming ritual.
- Per-person rankings on the office TV. A live wall leaderboard makes recognition a competition, and competitions get gamed — reciprocal pairs and volume-over-meaning follow within weeks. Leaderboards are fine as an opt-in page, not as ambient pressure.
- Vanity totals. "12,000 props all-time" is not actionable and crowds out the numbers that are. Trend and rate, not cumulative count.
- Recognition counts as performance inputs. The moment "gives" show up in a review, every metric above turns to mush. This data informs conversations; it is never a quota.
One last check: managers are the biggest blind spot in most recognition data, because their own gives are the ones most likely to be missing. If the manager rows of your participation pivot read low, fix that first — manager blind spots in recognition data explains why. And to connect these metrics to the retention outcomes that justify the program, measuring recognition program success covers the leading-versus-lagging framing this dashboard feeds.