-
1
Name the one gap the contest is supposed to close
Every good contest starts with a specific funnel gap, not a general wish for more revenue. Walk the pipeline from the last 90 days and name the one place where volume or conversion is below where it needs to be: new logo meetings held, pipeline-sourced opportunities, demos set with a decision maker, multi-threaded contacts per deal, outbound sequences completed. Pick the gap that most constrains the quarter and that a behavior change over 1 to 4 weeks can plausibly move. If the gap is win rate on late-stage deals, a contest is the wrong tool, because win rate changes slowly and the behaviors behind it do not respond to a leaderboard. If the gap is a top-of-funnel count, a contest can concentrate attention exactly where you need it.
- Pull a 90-day funnel waterfall and circle the single stage with the biggest drop versus plan.
- Write the gap as a specific metric with a unit, not a theme (not 'more pipeline' but 'new logo demos set per rep per week').
- Confirm the behavior that moves the metric is something a rep controls in a one-week window, not a buyer-dependent outcome.
- Reject contests aimed at win rate, cycle time, or expansion revenue; those belong in coaching, not a leaderboard.
Tip: If three sales leaders cannot agree in one meeting which gap the contest is for, the contest is premature. Fix the funnel disagreement first; a contest will not resolve it.
-
2
Pick a single metric and define it in writing before you announce
A contest with two metrics has none. Pick one countable metric, write a one-paragraph definition that spells out what counts, what does not, and what the edge cases are, and get the enablement partner, RevOps, and the sales leader to co-sign it before the contest is announced. For a new logo demo contest, the definition should state what qualifies as a demo (held, not scheduled), what counts as a new logo (no activity in the last 12 months), who verifies attendance, and how a demo gets disqualified if the account is already in another rep's pipeline. Ambiguous rules are the single biggest source of contest disputes, and the dispute always lands on the sales leader during the awards call.
- Write the metric definition as a single paragraph that any new hire could apply without asking a question.
- List at least three edge cases up front (no-show demos, reschedules, dual-coverage accounts) and the ruling for each.
- Have RevOps confirm the metric is reportable from the CRM without a manual pull so the leaderboard is auditable.
- Freeze the definition for the full contest window; mid-contest rule changes destroy trust for a full quarter.
Tip: If the metric cannot be reported from the CRM without a spreadsheet, the contest will drift. Fix the reporting first or pick a different metric.
-
3
Set a fixed window short enough to feel urgent and long enough to be fair
A contest is defined by its window. One week feels urgent but favors reps already running hot that week. A full quarter blurs into the forecast and loses its own identity. For most behavior-change contests, two to four weeks is the right range: long enough that a rep who starts behind can still catch up, short enough that the focus does not drag or cannibalize other work. Announce the exact start date, end date, and timezone cutoff in the kickoff message, publish the leaderboard refresh cadence (daily is standard, hourly is overkill), and set an awards date within one week of the close so the energy does not fade.
- Default to a two-week or four-week window; justify any shorter or longer window in writing to the sales leader.
- Publish start, end, and cutoff times in a single timezone and never change them once announced.
- Set a leaderboard refresh cadence that is predictable (daily by 9 a.m. local, for example) rather than continuous.
- Book the awards announcement within one week of close so the recognition still lands while the context is fresh.
Tip: Never extend a contest mid-flight. If the metric is tracking low, run the retro and ship a different contest next cycle; extensions reward the reps who paced themselves and punish the ones who sprinted.
-
4
Choose a prize that motivates without distorting the forecast
Prize design is where most contests go sideways. A prize so large that reps will abandon late-stage deals to chase it warps the forecast for a full quarter. A prize so small that it feels dismissive produces no behavior change and signals the gap was not important. The right range is a prize meaningful enough that a rep will reorder their week around it, but not so large that it competes with on-target commission on real deals. Mix experiential and material prizes (a dinner with the CEO, a premium headset, a Friday off, a charity donation in the rep's name) rather than piling on cash, because experiential prizes build culture and cash prizes blur into commission math. Always include a prize for second and third place so a late-start rep still has a reason to engage in week two.
- Benchmark the prize value against roughly one to three days of on-target commission for the contest window, not a full quarter.
- Mix prize categories: top prize experiential, runner-up material, team prize for a shared target hit.
- Avoid cash-equivalent prizes that read as a side commission and warp the compensation conversation.
- Publish second and third place prizes clearly up front so late-start reps still see a path to recognition.
Tip: If finance flinches at the prize budget, that is a signal the prize is too close to real commission. Shrink the prize or pick a different reward category rather than negotiating the metric down.
-
5
Write guardrails that keep tenured reps from running away with it
The default winner of an open leaderboard contest is the most tenured rep in the biggest territory, every time. If the contest only rewards raw count, you have built a contest for your top quartile and demotivated the rest of the team. Build in guardrails before you announce: run the leaderboard by percent lift over the rep's trailing 90-day baseline rather than raw count, segment the leaderboard by tenure band (new hires in ramp, mid-tenure, tenured) so there are multiple winners, or cap each rep's countable units per day so volume alone cannot carry a win. The guardrail you pick depends on team shape, but you need at least one or the contest will reward territory and tenure instead of behavior change.
- Decide the leaderboard shape up front: raw count, percent lift over baseline, or tenure-segmented bands.
- If raw count is the shape, publish segment leaders alongside the overall leaderboard so new hires have a path.
- Cap daily countable units per rep if the metric is gameable through quantity over quality.
- Add a quality gate (a demo must be scored X on the rubric to count, for example) if the metric is easy to game.
Tip: A contest where the same rep wins three cycles in a row is not a contest, it is a recognition program for one person. Rotate the leaderboard shape every quarter so different reps have a chance to appear.
-
6
Communicate the contest with a one-page brief, not a hallway chat
Reps need the rules in writing where they can refer back without asking the manager. Write a one-page brief that names the gap, the metric with its full definition, the window, the leaderboard shape, the prizes for first through third, the award date, and the two owners (sales leader and enablement partner) reps can ask when a rule is unclear. Send the brief in the kickoff message, pin it in the team channel, and make it the first slide of the sales kickoff meeting that opens the contest. Managers run the kickoff with their team; the sales leader does not run the contest personally for every rep. A one-page brief protects you from the hallway rule interpretations that always come up in week two.
- Draft the brief as a one-page document with headings for Gap, Metric, Window, Leaderboard, Prizes, Owners.
- Have each sales manager host the kickoff for their own team the same day the brief is published.
- Pin the brief in a durable channel (not a message that scrolls away) and link it in the leaderboard.
- Open an async thread for rule questions so answers are visible to the whole team, not whispered in DMs.
Tip: If you catch yourself answering the same question from three reps on different threads, publish an addendum to the brief the same day. The second-worst thing you can do mid-contest is let two reps operate on different rulebooks.
-
7
Run a daily leaderboard that focuses attention without becoming performance management
The leaderboard is the engine of the contest and the single biggest source of its risks. Publish it daily, celebrate movement not just position (the biggest gain of the day, the first new hire into the top five, the rep who hit their personal baseline for the first time), and keep performance-management conversations entirely off the leaderboard thread. If a rep is on a PIP, a public contest leaderboard is not where their numbers should surface. Keep the leaderboard thread for recognition and keep the one-to-one performance conversation in the private 1:1 where it belongs. The leaderboard is a motivation tool, not an evaluation tool, and conflating the two is how contests breed resentment on the second-tier reps who need the behavior change the most.
- Publish the leaderboard on a predictable daily schedule in a dedicated channel, not DMs.
- Call out movement and milestones daily, not just rankings, so mid-pack reps stay engaged.
- Keep PIP conversations and formal performance reviews entirely separate from the contest thread.
- Rotate the daily call-out between managers so the enablement partner does not become the sole cheerleader.
Tip: If the leaderboard thread starts attracting complaints or inside jokes at the bottom-ranked rep's expense, lock it down the same hour. One toxic thread will cost you more than two contests are worth.
-
8
Flag contest-period bookings in the forecast so RevOps and finance can see the lift
A contest that lifts the pipeline-sourced metric but is invisible to the forecast will convince finance it had no effect. Before the contest starts, agree with RevOps on how contest-period activity will be tagged in the CRM (a campaign source, a custom field, or a date-bounded filter) so the pipeline the contest creates can be reported separately from the baseline. During the contest, confirm the tag is being applied so you are not reconstructing attribution after the fact. In the forecast call during and after the contest, call out the contest-sourced pipeline explicitly so the leadership team sees both the win and any cannibalization of other sources. If the contest produced 20 new demos but new outbound dropped by 30 for the same window, the forecast reality is net negative and you need to see it.
- Agree with RevOps on the attribution tag before the contest opens; do not rely on reconstruction after.
- Spot-check tag application once a week during the contest so late corrections do not pile up.
- In the forecast call, report contest-sourced metrics alongside the baseline metrics they might cannibalize.
- Flag any suspicious pacing (a cliff on the last day, a cluster on one account) for review before counting it.
Tip: The question finance will ask after the contest is 'what did we not do while we were running it'. Prepare the cannibalization answer up front; you will need it either way.
-
9
Run a written retro before you ship the next contest
A contest that is not reviewed in writing becomes institutional habit regardless of whether it worked. Within one week of close, run a 45-minute retro with the sales leader, enablement partner, two managers, and two reps (one who won, one who did not). Review the lift against baseline, the cannibalization against adjacent metrics, the quality of what got produced (did the demos convert, did the pipeline hold, did the new logos stay engaged), and the rep sentiment captured in a short post-contest survey. Write a one-page retro doc with three decisions: ship the same contest, change one variable and ship again, or retire this contest entirely. File the retro doc with the original brief so the next contest designer three quarters from now can see what you learned, not just what you ran.
- Collect baseline, lift, and cannibalization numbers in a one-page scorecard before the retro meeting opens.
- Survey all reps with three questions (did it motivate you, was the metric fair, would you run it again) within 72 hours of close.
- Interview the winner and one non-winner on video or in writing so sentiment is captured, not inferred.
- Make an explicit ship, iterate, or retire decision in writing and file it with the original brief.
Tip: If no one is willing to retire a contest that did not move the metric, you have a culture problem, not a contest problem. Retiring a bad contest is the single clearest signal that leadership takes the design seriously.
