A sales SPIFF or contest is one of the fastest levers a sales leader can pull to shift rep behavior inside a quarter. Done well, it concentrates effort on a specific motion, lifts the metric you care about, and leaves the team wanting the next one. Done poorly, it pays out to the reps who were going to win anyway, breeds resentment with the ones who were not, and teaches the field that contests are theater. This guide walks the full cycle, from the behavior you are trying to buy to the retro that tells you whether to run it again.
Time: 1-2 weeks to design, 30-90 days to run
Most failed contests start with the words sell more. Sell more is a wish, not a design brief. A working SPIFF names a specific behavior the field is not doing enough of today and that you believe moves pipeline or revenue. New-product pipeline created, multi-threaded opportunities, discovery calls per rep, demos run on a specific segment, renewals closed 30 days early, outbound meetings booked into a target vertical. The sharper the behavior, the easier it is to measure, the harder it is to game, and the clearer the field is on what to do differently on Monday. If you cannot write the behavior on a sticky note in one sentence, the contest is not ready to design.
Contest length should match the behavior cycle, not the calendar. Short motions like outbound meetings or discovery calls run well on a 30-day window because the behavior repeats fast enough to show daily movement on the scoreboard. Mid-length motions like new-product pipeline or multi-thread rate want 45-60 days so reps have time to work accounts and see the result land. Full-cycle motions like closed-won in a target vertical need the full 90 days to let deals move through stages. Running a contest shorter than the behavior cycle rewards reps who already had the opportunity in motion. Running it longer than the cycle lets energy decay and the scoreboard goes stale by week six.
A prize structure with only a top-performer award pays out to the reps who were going to lead the board anyway and tells the middle of the team to sit this one out. The structure that actually changes behavior has three tiers. First, a top-performer prize that rewards the number-one finisher and gives the stretch reps something to chase. Second, a threshold reward that any rep can earn by hitting a defined bar, which pulls the middle of the team into the contest and is where most of the behavior lift comes from. Third, a raffle or random draw tied to qualifying activity, which keeps the back half of the team participating because any qualifying rep can win. Fund the three tiers from one budget and weight the spend toward the threshold tier, not the top prize.
Every contest that ends badly ends badly for the same reason: the rules were loose and leadership changed them midstream to patch something. Write the rules down before kickoff and publish them. Cover eligibility, the exact metric being measured and how it is calculated, the start and end times down to the time zone, how ties are broken, what disqualifies an entry, how prizes are paid out, and the tax handling. Call out what counts and what does not: canceled deals that recognized, overturned opportunities, split credits, team sells, SDR-sourced versus rep-sourced. Then freeze the rules. If a loophole appears mid-contest, document it, honor the current rules, and fix the design in the next one. Changing rules retroactively costs more trust than any single payout saves.
A SPIFF that launches by email alone does not land. Open the contest with a short, high-energy session, 15 to 30 minutes live, that walks the behavior, the three prize tiers, the rules doc, and the scoreboard reps will check every day. The scoreboard is the single most important asset in the whole program. It needs to show every rep their rank, their distance from the threshold, and the top of the leaderboard, and it needs to update at least daily without a RevOps ping. Record the kickoff so new hires and anyone who missed it can watch asynchronously. Then open an ongoing channel, Slack or Teams, where results, hype, and clarifying questions all live in one place for the duration of the contest.
Energy on a contest peaks in the first week and the final week, and dies in the middle unless someone actively feeds it. Build a weekly rhythm. Every Monday, post a scoreboard update with the top five, the threshold progress by rep, and a callout for the biggest climber of the previous week. Mid-contest, run a leaderboard shake-up: a 48-hour mini-SPIFF layered on top of the main one, a double-points day on a specific activity, or a manager shout-out thread for the behavior you want more of. First-line managers should name the contest in every one-on-one so it stays top of mind during pipeline work. If the field goes quiet in week three, the contest is already drifting. Fix it that week, not at the retro.
Private payouts kill half the ROI of a SPIFF. The point of a contest is as much about the public moment as the prize itself. Announce winners live, in the all-hands or sales standup, with the sales leader on camera. Name every threshold earner, not just the top finisher, so the middle of the team sees that hitting the bar got recognized too. Draw the raffle live on the same call. Share the aggregate numbers: how much new-product pipeline the team built, how many extra discovery calls got run, what the net lift was versus baseline. Tie the win language to the behavior you were buying, not to the prize value, so the lesson the field takes away is about the motion, not the money.
Two weeks after the contest closes, run a formal retro on three questions. First, did the behavior move versus the baseline you captured in step three of the design, and did it move for the middle of the team or only for the top. Second, did the behavior shift drive measurable pipeline or revenue in the window, not just activity for its own sake. Third, what did the full program cost, including prize payout, enablement time, and leadership attention, and was the lift worth that spend. Document the answers in a one-page retro and store it next to the rules doc. If the behavior moved and the lift paid for the program, keep the design and change only the behavior you target next time. If it did not, redesign the structure before running another one.
A SPIFF (Sales Performance Incentive Fund) is a short-term cash or near-cash incentive layered on top of commission, usually tied to a very specific product, segment, or motion. A sales contest is the broader category and often uses non-cash prizes, tiered winners, and a leaderboard dynamic. In practice teams use the terms interchangeably, and the design principles in this guide apply to both.
Match contest length to the behavior cycle, not the calendar. Short motions like outbound meetings or discovery calls run well on 30 days. Mid-length motions like new-product pipeline want 45-60 days. Full-cycle motions like closed-won in a target vertical need the full 90 days. Running shorter than the behavior cycle rewards reps who already had the pipeline. Running longer lets energy decay by week six.
Both work, and the right answer depends on the behavior and the field. Cash is simplest for threshold rewards and raffle tiers because it is easy to administer and reps always value it. Non-cash prizes, especially experiences and trips, create a stronger public celebration moment and often get remembered longer. Most teams run a mix: cash for the threshold tier where volume matters, non-cash for the top-performer tier where the story matters.
Three defenses. First, pick a behavior metric with a clear calculation and a clean audit trail in the CRM. Second, spell out credit rules in writing for splits, team sells, SDR assists, overturned deals, and canceled bookings, before kickoff. Third, hold payouts until after the standard reversal or clawback window has closed so a cancelled deal cannot pay out. If you cannot audit the metric from the CRM in under five minutes per rep, the contest is too complex.
First-line managers should not compete head to head with their reps because it muddies the incentive and discourages coaching mid-contest. A clean pattern is to run a parallel manager contest tied to team-level behavior lift, so managers are rewarded for pulling their whole team into the threshold tier rather than for their own individual activity. Everyone above the first line should be out of scope entirely.
There is no universal ratio, but the right test is at the retro. Add up prize payouts, enablement time, and leadership attention, and compare that total to the measurable pipeline or revenue lift versus baseline. A healthy contest returns multiples of its full cost in the window it runs. If the program only breaks even, the structure is likely paying out to deals that would have closed anyway and the design needs a tighter behavior focus before running another one.
Strkr makes it easy to define the behavior metric, publish a live scoreboard, inspect threshold progress in pipeline reviews, and measure lift versus baseline so every contest earns its spend.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
