management

Leaderboards That Motivate Instead of Punish

Field experiments found bare rank feedback can lower sales output. What the research says to put on the board, and what to leave off.

Every canvassing team has a leaderboard. Most of them are a list of names in descending order of sets, written on a whiteboard, updated daily.

That exact artifact is the version the research says can cost you output. Here is what the studies actually found and what to do instead.

Finding one: removing a bare ranking raised performance by 11 percent

Iwan Barankay at Wharton ran a three-year randomized experiment with 1,754 full-time furniture salespeople, published as Rank Incentives: Evidence from a Randomized Workplace Experiment. All reps had been receiving rank feedback before the study, so the experiment measured what happens when you take the leaderboard away.

The abstract reports that "removing rank feedback actually increases sales performance by 11%, or 1/10th of a standard deviation." Pay was commission on absolute performance throughout, so rank carried no money.

The explanation is specific, and it will be familiar to anyone who has managed a field team: "telling people that their rank is worse than expected leads to a demoralization effect, a drop in performance."

Barankay also reports the effect was driven by men rather than women, which is worth knowing but hard to act on.

Finding two: the benchmark is what fixes it

The same experiment's second manipulation is the actionable part, and it is the sentence most people who cite this study skip.

Reps in some conditions got not just their rank but benchmarks: what level of performance was needed to reach the top 10 percent, the top 25 percent, the top 50 percent. Barankay reports that adding benchmarks to rank feedback "significantly raises performance," because knowing what is "needed to rise in the rankings dampens this demoralization effect."

So the problem is not comparison. The problem is comparison with no visible path.

"You are 9th" tells a rep they are losing and nothing else. "You are 9th, and 4 more sets this week puts you 5th" tells them what to do with the next four hours. The first is a verdict. The second is a goal.

This lines up with Locke and Latham's 35-year review in American Psychologist, which found that "specific, difficult goals consistently led to higher performance than urging people to do their best," with meta-analytic effect sizes from .42 to .80. A rank without a benchmark is closer to "do your best" than to a goal, because, in their words, it has "no external referent."

Finding three: public ranking can make people conform downward

Anik Ashraf ran a five-month field experiment with 366 piece-rate workers, reported in CESifo Working Paper 9591 and later in Management Science. Three arms: a control, private ranks, and public ranks.

Private ranks did nothing: "intrinsic incentives from private ranks do not affect productivity." Public ranks did something, and not only upward. "Publicly-ranked workers reduce productivity to conform to their social groups in the workplace."

The magnitudes tell the story. Publicly-ranked workers with no friends in the treatment raised productivity by more than 4.5 percent. Publicly-ranked workers competing with at least one friend were more than seven percent less productive than that group, netting a 2.7 percent decrease. Cooperation also fell. Ashraf's conclusion: "inducing competition among workers may be counterproductive for firms."

This is a garment factory, not a sales floor, and the work is piece-rate rather than commission. Do not oversell the transfer. But a canvassing crew is unusually tight socially. They ride together, eat together, and work the same subdivision. If there is a workplace where "conforming to your friends' output" is a live risk, it is a canvassing truck.

Finding four: identifiable rankings beat anonymous ones, and quotas add nothing

A counterweight is worth stating, because the evidence is not one-directional.

The American Marketing Association, which publishes the Journal of Marketing, summarized research by Ahearne, Pourmasoudi, Atefi, and Lam in Do Performance Rankings Actually Motivate Salespeople?, covering more than 27,000 salespeople across 170-plus firms and 83 countries.

Their findings as the AMA describes them: anonymized rankings raised quota attainment but increased turnover. Identifiable rankings had the strongest effects, improving quota attainment and reducing turnover, attributed to reps wanting to protect their standing among peers. And when quota targets were bolted onto identifiable rankings, in the AMA's words, "we fail to see performance enhancing benefits."

Note this is the AMA's summary rather than the article itself, which is paywalled.

So Ashraf says public ranking can backfire and Ahearne says identifiable ranking works best. The thread that reconciles them is not "public good, private bad." It is that context does the work: a bare rank is corrosive, a rank with a benchmark is not, and layering a fixed quota on top of a ranking adds nothing.

What to put on the board

Translating all of that into a canvassing leaderboard:

Show the gap next to every rank. This is the single highest-value change and the one most teams can make this week. Every position on the board gets a "distance to next" figure in the units the rep controls.

Lead with input metrics. Doors logged, hours in field, callbacks honored on time. These are achievable every single day regardless of luck, which means the board is not purely a map of who got the best territory. A rep can have a zero-set day and still be top three on work, and that is the thing that keeps them in the truck in week three.

Normalize outcome metrics for conditions. Publish set rate within comparable hour bands and comparable territories, or do not publish it. A set-rate ranking that ignores turf is partly a ranking of territory assignments, and the reps can all see it.

Run multiple boards, not one. Today, this week, and all-time. A single cumulative board is permanently topped by whoever has been there longest, which makes it invisible to everyone else. A daily board resets, which means a new rep has something to win on day one.

Suppress small samples. Do not rank a rep on a ratio built from 70 doors. The error bar swamps the signal, and publishing it teaches the team that the numbers are noise, which is how you lose logging discipline.

Keep a private detail view. Give each rep their own full diagnostics privately, including the things that would be unkind in public. Ashraf's result suggests the private channel carries less social risk, and it is where coaching actually happens.

What to leave off

Bottom-of-board callouts. There is no version of naming the worst performer publicly that the evidence supports. Barankay's demoralization effect is exactly this, measured.

Comparisons across incomparable turf. Already covered, but it is the most common complaint reps have about leaderboards, and they are usually right.

A fixed quota bolted on. Per the AMA summary, adding quota targets to an identifiable ranking produced no performance benefit. It also imports the end-of-period gaming problem that quotas carry.

Anything a rep cannot verify. If a rep cannot reconstruct their own number from their own logged doors, the board is a rumor. Trust in the data is a precondition for the board doing anything at all.

A reasonable default

If you want one configuration to start from: a daily board ranked on logged doors, a weekly board ranked on sets with the gap to the next position shown next to every name, hour-band normalization on any rate, and a private per-rep view with the full diagnostics. No bottom callouts, no quota overlay, no ratios under a few hundred doors.

That is less dramatic than the whiteboard most teams run. The evidence says it is also likely to produce more knocks.

Canvass Pro tracks XP and levels per rep with today, week, and all-time boards, so the ranking can be read daily rather than as one permanent cumulative list.