How do you coach a rep who sandbags their forecast?

Two roads: fix the incentive or fix the conversation

Almost every manager facing a rep who sandbags their forecast lands on one of two paths, and picking the wrong one wastes a full quarter.

Road one — change the system. You treat the sandbag as a rational response to the environment you built. If a rep's accelerator kicks in at 100% and quota resets off attainment, holding deals into next quarter is arithmetic, not character. The same is true when the comp plan pays identically for a 90% accurate forecast and a 60% accurate one: accuracy is unpaid labor, and unpaid labor gets skipped. Fixing the system means auditing the plan with finance and RevOps, adding a forecast-accuracy component or an MBO tied to commit-to-close ratio, flattening the reset cliff, and rewriting the forecast category definitions so "commit" has exit criteria a third party could audit. The advantage is leverage: one plan change moves every rep on the team, including the quiet sandbaggers you haven't caught. The cost is time and politics — comp changes usually land at plan year or half-year boundaries, need CRO and finance sign-off, and carry blast radius. Change the accelerator badly and you demotivate the two reps who were forecasting honestly the whole time.

Road two — change the relationship. You treat the sandbag as a trust artifact: the rep committed a deal once, it slipped, they got publicly grilled, and they learned that honesty is expensive. The fix is a private 1:1, a named pattern, an explicit promise about what happens when an evidence-backed deal slips, and a weekly inspection cadence that makes the new behavior stick. The advantage is speed — you can start Monday, no approvals required, and it works even when the comp plan is genuinely fine. The cost is that it's per-rep, it depends on your credibility, and it evaporates the moment you break the promise. One ugly forecast call in front of the team and you're back to square one with that rep and everyone who watched.

There's a third road worth naming because managers stumble into it by accident: mandate a higher commit number. "Your commit is too low, add two deals." This never works. It doesn't change the incentive or the fear; it just teaches the rep to park deals one stage earlier, in "best case" or in an unlogged opportunity that appears the week it closes. You've moved the hiding place, not the habit, and now you've also taught them you'll override their judgment — which makes the next honest call less likely, not more.

The honest answer is that roads one and two are complements, not alternatives. The comp audit tells you whether sandbagging is rational; the coaching conversation tells you whether it's emotional. Run the audit in the background because it's slow, and run the conversation this week because it's fast.

How to decide which road you're on

The diagnosis matters more than the intervention. Four root causes produce identical symptoms — a rep whose commit lands at 60% while their close lands at 95% — and each one has a different fix.

Knowledge. The rep doesn't actually know what your categories mean. "Commit" is a word with a precise meaning in your head and a vague one in theirs. This is the most common cause on teams under two years old or after a manager change, and it's the cheapest to fix: write the definitions down, walk through five real deals against them, done in a week.

Skill. The rep genuinely cannot read deal signals. They can't tell a champion from a coach, they've never confirmed an economic buyer, and they mark deals "best case" because guessing low feels safer than guessing wrong. They aren't gaming you — they're hedging against their own uncertainty. Fix with MEDDPICC or your qualification framework plus call review, not with an accountability speech.

Will. The rep knows the definitions, reads signals fine, and hides deals anyway because it's advantageous or comfortable. This is the rarest of the four and the only one where a direct accountability conversation is the right opening move.

System. The "sandbagging" is honest conservatism. Their territory is three whales and forty minnows, so any single deal swings their number 30%. Or the stage model in the CRM was built for a transactional motion and they sell an eighteen-month enterprise cycle. Or quota was set on last year's tailwind. Coaching a system problem is how you lose a good rep.

Run the symptom through the tree before you open your mouth in the 1:1.

The tree is not academic. A manager who skips it and delivers an accountability lecture to a rep with a knowledge gap does real damage: the rep now believes you think they're dishonest, and they were just confused. Ask the diagnostic questions in the 1:1 itself and let the rep's answers route you. "Walk me through how you decide what goes in commit" separates knowledge from skill in about ninety seconds. "Is there anything about how you're paid that makes calling a deal low the smarter move?" separates system from will, and reps answer it honestly far more often than managers expect — because you asked instead of accused.

The numbers that tell you which problem you have

Sandbagging is one of the few sales behaviors with a clean quantitative fingerprint. You don't have to guess.

Commit-to-close ratio. Take committed dollars at the start of the quarter and divide closed-won by it. Healthy sits somewhere around 85–95% — reps should miss a little, because a forecast that never misses isn't a forecast, it's a floor. A rep running 130% or higher quarter after quarter is not lucky; they're parking deals. A rep at 60% has the opposite problem, happy ears, which is a different coaching conversation with the same root technique. Pull four quarters, not one — a single quarter can be one whale landing early.

Slip symmetry. Count deals that moved up into commit mid-quarter versus deals that fell out. An honest forecaster's surprises run in both directions. A sandbagger's surprises are almost entirely positive: things appear, nothing disappears. That asymmetry is the single hardest signal to explain away in a 1:1, and it's why you bring it to the conversation instead of an accusation.

Best-case conversion. What share of "best case" closes in-quarter? If best case is converting at 70%+ while commit converts at 95%, best case has become the real commit and your categories have drifted. On a healthy board, best case converts in a much lower band and the spread between the two categories is wide and stable.

Signal completeness. Of the deals in commit, what percentage have a named economic buyer you or the rep has actually spoken to, a compelling event with a date attached, and a mutual action plan the buyer has acknowledged? This is auditable in the CRM if you require the fields, and it converts "I feel good about it" into a number. A rep whose commit deals are 90% signal-complete and still under-committing is conservative, not sandbagging.

Late-stage age. How long do deals sit in the final two stages? A sandbagger often has deals aging in a late stage that are functionally closed — verbal given, paperwork moving — but not categorized as commit. Cross-reference stage age against category and the parked deals surface fast.

Self-flagged risk. Count how many times per quarter the rep proactively tells you a deal is in trouble before you ask. This is the trust metric. It starts near zero on a fearful rep and rises within two cycles once they believe an honest downgrade isn't punished. It's the leading indicator that the coaching worked — commit-to-close is the lagging one.

One caution on all six: a rep with eight deals a quarter produces noisy ratios, and a rep with eighty produces clean ones. In enterprise territories with low deal counts, weight signal completeness and slip symmetry heavily and treat the ratio as directional. In high-velocity SMB motions, the ratio stabilizes fast and you can trust it inside two quarters. Same behavior, different instruments — RevOps should be building the dashboard with segment-appropriate thresholds rather than one global band that makes every enterprise rep look like a sandbagger.

Running the conversation and the ninety days after it

Do this in a private 1:1. Never in a forecast call, never in front of peers — a rep who feels exposed will defend rather than explain, and you'll get a story instead of a cause.

Open with the pattern, not the verdict. Name the data and hand them the interpretation: "I pulled four quarters. You committed 60% of your number and closed 95% both years running. That's a great result, and it also means I can't forecast this region — which costs us both when I go ask for headcount. Walk me through how you decide what goes in commit." You've stated a fact, given credit, named your problem, and asked an open question. Nothing in there is an accusation.

Get their definition of accuracy. Ask what a perfectly accurate forecast would look like for them — not a safe one, an accurate one. Reps who have never been asked this often realize mid-sentence that they've been optimizing for something else.

Surface the driver with silence. Two questions do most of the work: "The last time you committed a deal and it slipped, what happened?" and "Is there anything about how you're paid that makes calling a deal low the smarter move?" Ask, then stop talking. If they describe getting grilled, you have a trust gap. If they describe an accelerator or reset quirk, you have a system problem to escalate — and you should thank them for it, because they just did diagnostic work for you.

Build the standard together. Replace feel with three confirmable signals: an economic buyer you've actually spoken with, a compelling event with a date, and a mutual action plan the buyer agreed to. All three present means commit. Missing one means best case, and best case is a respectable place for a deal to live. Ask directly whether they can work with that definition, and let them argue — a rep who negotiates the standard owns it.

Make the trade explicit. The line that breaks the loop is a promise with a boundary: you will never punish an honest miss on a deal that had real signals; you will absolutely address hiding. When a signal-backed deal slips, you debrief which signal was wrong — you don't relitigate their judgment. They give you accuracy; you give them safety. That exchange is the entire mechanism, and it only works if you keep it the first time it's tested, which will be within six weeks.

Then run the cadence, because one conversation doesn't undo a quarter of habit.

Days 0–30, re-baseline. Write the definitions down where both of you can see them. Re-score every open deal together against the three signals. Expect the commit number to jump as parked deals surface — that jump is the point, not a problem, and you should tell your own manager it's coming so nobody misreads it as a sudden pipeline miracle.

Days 31–60, inspect and reinforce. Every weekly 1:1, review exactly two deals: one they committed and one they held back. Ask for the evidence on both. The held-back deal matters more — that's where the habit lives. When they make an accurate call, including correctly calling a deal at risk, name it out loud in the moment. Praise for a downgrade feels strange to give and is the most valuable thing you'll say all week.

Days 61–90, reward accuracy. Put the commit-to-close ratio in how you talk about their performance, alongside attainment. Celebrate the rep whose number you can bank on in the team meeting, not only the rep who beat a soft commit. The team is watching which behavior gets applause, and they will do that one.

Two drills accelerate all of it. First, the blind forecast bet: before the quarter, the rep writes their real expected number in a sealed note; at quarter end you compare it to what they actually committed. The gap *is* the sandbag, rendered as a number neither of you can argue with. Second, evidence role-play: you play a skeptical CRO, the rep defends one commit deal, and every time they say "I feel good about it" you ask what the proof is. Feelings don't survive that drill, which is exactly the point.

Where the same failure shows up outside the rep's forecast

Sandbagging isn't a sales-rep pathology. It's what any forecasting system produces when accuracy is unrewarded and misses are punished, and once you see the pattern you'll find it upstream and downstream of the rep.

Marketing's pipeline number. A demand gen team burned once for missing an MQL-to-pipeline commitment will start reporting a conservative number and quietly bank overflow into next month's attribution. Same mechanism, different spreadsheet. The tell is identical: positive surprises only.

CS and renewals. Renewal forecasts sandbag hard, because a CSM who calls a renewal at risk invites executive attention they don't want, while one who calls it safe and loses it can point at the customer. If your gross-retention forecast is always beaten, you have a sandbagging problem in the post-sale org and you're making capacity decisions on fiction.

The manager's own roll-up. Frontline managers sandbag their number to their director for exactly the reasons their reps sandbag to them. This is the part most coaching advice skips: if you're coaching a rep to forecast honestly while you shave 15% off your own roll-up before sending it up, the rep will figure that out, and your promise about honest misses will read as theater. Fix your own call first.

Delivery, services, and finance. Professional services teams pad delivery estimates when a slipped date gets a postmortem and an early finish gets nothing. Finance builds a conservative plan when a beat is celebrated and a miss triggers an all-hands. Every one of these is the same incentive geometry, and the same three interventions apply: define the category by evidence, make the honest miss survivable, and put accuracy in the reward function.

There's a real trade-off here worth stating plainly. A board and a CFO genuinely want a number they can beat — some conservatism at the top of the house is a feature, not a bug, because guiding low and delivering high is how public companies stay credible. The distinction that keeps this coherent: conservatism belongs at the guidance layer, accuracy belongs at the operating layer. Your reps' forecasts are the operating layer. If they sandbag, and then you sandbag their number again, and then the VP sandbags that, you've compounded three layers of hedge and nobody in the building knows what's actually going to happen. Capacity, hiring, and inventory decisions get made on that number. Let the CFO apply the haircut once, at the top, deliberately — and let RevOps be the function that protects the raw signal underneath it.

The practical implication for a RevOps team: instrument the layers separately. Store the rep's raw call, the manager's adjusted roll-up, and the executive commit as distinct fields, and report the delta between them. When the delta is stable and explainable, the system is healthy. When one layer's delta balloons, you know exactly which conversation to have and with whom — which is a far better use of a forecast meeting than going deal by deal through forty opportunities.

Related questions

What if the rep keeps sandbagging after the coaching plan?

If you've rewritten definitions, ruled out comp and territory causes, kept your promise on an honest miss, and the behavior persists across two full cycles, you've moved from coaching to accountability. Document the pattern factually and handle it as a performance conversation.

Does sandbagging actually hurt the company if the rep hits quota?

Yes. Capacity planning, hiring, inventory, and cash decisions all run off the forecast. A sandbagged number understates demand, so the company under-hires and under-invests, then scrambles. The rep's attainment is fine; the company's decisions are wrong.

Can forecast tooling catch sandbagging on its own?

Partly. Platforms like Clari and Gong surface deals with strong buyer engagement sitting in low-confidence categories — the sandbag fingerprint. That gives you evidence to open a conversation with, but tools can't distinguish fear from gaming, and that distinction determines the fix.

How do you coach the opposite problem, a rep with happy ears?

Same architecture, mirrored. Define commit by the same three signals, but the drill inverts: make them argue why a deal *won't* close. Happy-ears reps usually have a skill gap in confirming an economic buyer, not a trust gap.

Should forecast accuracy be part of variable comp?

It can be, usually as a modest MBO rather than a core component — often single-digit percentages of variable pay. Weight it too heavily and you incentivize a new game: reps stop pursuing upside deals that would blow their own forecast.

FAQ

How is sandbagging different from a rep who's simply conservative?

Conservatism is calling deals low because you genuinely can't confirm the signals — an honest information limit. Sandbagging is deliberately under-committing deals you expect to win. The blind-forecast-bet drill separates them cleanly: a conservative rep's private number matches their commit, a sandbagger's doesn't. Coach the first with qualification skills, the second with incentives and trust.

Can I just tell the rep to add deals to commit?

No, and it usually backfires. Mandating a higher number without changing the fear or the incentive teaches the rep to hide deals one stage earlier — or to keep them out of the CRM entirely until the week they close. You've relocated the hiding place and signaled that you'll override their judgment, which makes the next honest call less likely.

How long before the forecast straightens out?

Plan on two full cycles. The first surfaces the parked deals and the commit number jumps, which can look alarming if you haven't warned your own manager. The second is where the new habit and the trust fix produce a stable commit-to-close ratio you can actually bank. Anything faster is usually compliance, not change.

What's the single fastest way to create a sandbagger?

Publicly grill a rep for a committed deal that slipped despite real signals. Everyone in the room learns that honesty is expensive and hedging is free. The second fastest is celebrating the rep who "always beats their commit" without ever asking how soft that commit was — that trains the whole team to lowball.

Should this conversation happen in the forecast call or the 1:1?

The 1:1, always. A forecast call has an audience, and an audience turns a diagnostic conversation into a defense. Use the forecast call for the standard — asking every rep for evidence on commit deals, uniformly — and the 1:1 for the individual pattern and the promise.

Does RevOps own any of this or is it purely a manager problem?

RevOps owns the instrumentation and the definitions: category exit criteria, the required signal fields, the accuracy dashboard by segment, and the audit of whether comp mechanics reward hiding. Managers own the conversation and the cadence. Sandbagging that survives good coaching is usually a sign the RevOps layer never got built.

Sources

• Harvard Business Review: Get More From Your Sales Coaching
• Gong: Sales Forecasting Research and Guides
• Clari: Sales Forecasting and Forecast Accuracy
• MEDDIC Academy: MEDDPICC Explained
• RAIN Group: Sales Coaching Best Practices
• Salesforce: What Is Sales Forecasting
• McKinsey: Insights on Growth, Marketing and Sales
• Winning by Design: Revenue Architecture Resources

Related on PULSE

• [How do you coach a rep to improve forecast accuracy?](/knowledge/cg0107)
• [How do you coach a rep whose forecast is always wrong?](/knowledge/cg0103)
• [How do you coach a rep to forecast a deal honestly?](/knowledge/cg0098)
• [Top 10 Coaching Techniques for Sales Forecast Hygiene](/knowledge/cg0787)
• [Top 10 Forecast Coaching Habits for Remote Reps](/knowledge/cg0697)
• [Top 10 Forecast Coaching Habits for Top Performers](/knowledge/cg0696)