Optimized Sales Optimized Marketing Target Accounts For CROs For CFOs For CMOs Blog News Glossary Compare Tools About Schedule a Demo
Pipeline & Forecasting

Forecast Sandbag Detection

ORM Technologies
Home/ Glossary/ Forecast Sandbag Detection
Definition Forecast sandbag detection is the process of finding reps who forecast below what they expect to close, by comparing each rep's commit to actuals over four or more quarters and checking for one-direction misses, deals pulled in from next quarter, and late category upgrades.

How to detect sandbagging in a sales forecast

Forecast sandbag detection is the process of finding reps who forecast below what they expect to close, using their forecast history. Compare each rep's commit to actuals over four or more quarters, then check for one-direction misses, deals pulled in from next quarter, and late category upgrades. Updated November 2026: added detection metrics, a bias worked example, and false positives. Sandbagging can look harmless, since the team beats the number. It still costs the business. It distorts pipeline coverage math, drags down forecast accuracy, and leaves leadership hiring, spending and making board commitments against a number that is lower than the real pipeline. This page covers the detection method: which metrics to run, where to set the first thresholds, how to compute rep-level bias, and what to do with a flag. For the definition, see sandbagging. For the fix playbook, see how to reduce forecast sandbagging.

What metrics detect sandbagging?

Five metrics do most of the work: rep forecast bias, tracking signal, beat rate, late commit share, and pull-in share. The first three look at the rep's history. The last two look at how deals moved inside the quarter, which needs a stored record of the pipeline. Two deal-level checks round out the table. The thresholds are starting points. Tune them to your own history before you act on them.
MetricHow to compute itWhat sandbagging looks likeStarting-point flag
Rep forecast bias(Sum of week-one commits minus sum of actuals) / sum of actuals, over 4+ quartersNegative, quarter after quarterWorse than minus 10% pooled over four quarters
Tracking signalRunning sum of (commit minus actual) / mean absolute errorGrows more negative each quarterMinus 4, the conventional control limit
Beat rateQuarters where actual beat commit / quarters measuredBeats commit almost every time, rarely misses4 of the last 4, or 5 of the last 6
Late commit shareClosed-won value that entered commit in the final two weeks / total closed-won valueDeals held in best case or upside until the endWell above the team median, two quarters running
Pull-in shareClosed-won value that was dated for a later quarter on day one / total closed-won value"Next quarter" deals that close this quarterWell above the team median, two quarters running
Score-to-category gapDeals with a high predictive score held below commitCommit-quality deals parked in best caseScore above 70% with category below commit
Amount step-upClosed amount / amount on day one, per closed-won dealAmounts that rise at close, deal after dealAverage ratio above 1.0 for the rep
No single metric proves anything. Each of these can trip for an honest reason. A rep who trips three or four of them, quarter after quarter, has a pattern worth a conversation.

The base rate makes the signal easier to see. Across ORM customers, only about 20% of the pipeline value carrying an in-quarter close date on day one closes in that quarter, according to ORM Technologies. Most forecasts that trust the CRM start too high. A rep whose quarter keeps closing above their commit is running against that tide, which is why it stands out.

How do you calculate rep-level forecast bias?

Divide the gap between what the rep committed and what they closed by what they closed, and pool it over at least four quarters. Use the commit from a fixed point, such as the first week of the quarter. A commit updated in the last week has already absorbed the result.

``` Rep forecast bias % = (Sum of commits - Sum of actuals) / Sum of actuals x 100 Tracking signal = Running sum of (commit - actual) / Mean absolute error ```

This uses the same sign as ORM's forecast bias entry: positive means over-forecasting, negative means under-forecasting. Sandbagging is negative bias.

Illustrative worked example, two reps, week-one commits against actuals ($K):

QuarterRep A commitRep A actualRep A errorRep B commitRep B actualRep B error
Q1400480minus 80500450plus 50
Q2420500minus 80480530minus 50
Q3450530minus 80520490plus 30
Q4480560minus 80500540minus 40
Total1,7502,070minus 3202,0002,010minus 10
- Rep A bias: (1,750 minus 2,070) / 2,070 = minus 320 / 2,070 = minus 15.5%. - Rep A tracking signal: the running sum is minus 320 and the mean absolute error is 320 / 4 = 80, so minus 320 / 80 = minus 4.0. That is the control limit. - Rep B bias: (2,000 minus 2,010) / 2,010 = minus 0.5%. - Rep B tracking signal: the running sum is minus 10 and the mean absolute error is (50 + 50 + 30 + 40) / 4 = 42.5, so minus 10 / 42.5 = minus 0.2.

Rep B misses in both directions, and Rep B's errors cancel. Rep A beats the commit by the same $80K every quarter, which is what a cushion looks like. When every miss points the same way, the tracking signal equals the number of periods, which is why four straight low quarters land exactly on minus 4. The tracking signal entry explains the control limit, and how to correct rep-level forecast bias covers calibrating a rep once the lean is confirmed.

Which deal-level signals confirm it?

The deal-level signals are category, close date and amount changes that do not match what the buyer is doing. If a deal shows the marks of a likely close (active multi-threading, recent executive engagement, procurement involved, short time-in-stage against history) but sits below commit, that gap is worth a question.

You can only see these signals if you keep history. Pete Furseth, ORM's COO, says the most useful CRM data for forecasting is often "the change in the data." ORM snapshots the pipeline every day, so it can see when a deal entered commit, when its close date moved and when its amount changed. A live CRM view shows where a deal sits now and hides how it got there. Store a weekly forecast snapshot at minimum.

Three patterns to look for in that history:

- Category held, then jumped. A deal sits in best case for weeks with steady buyer activity, then moves to commit and closes within days near quarter end. - Close date pushed out, then pulled back. Furseth calls a close-date change the best signal that a committed deal will slip. Sandbagging runs the move the other way: the rep dates a live deal into next quarter, then closes it this quarter once the number needs it. - Amount suppressed, then restored. A $200K deal is carried at $120K with "expansion planned after close," then books near the full amount (illustrative).

Category holds are the easiest to spot in snapshot history. Close-date moves shift bookings and revenue recognition between quarters. Amount suppression is the hardest to see without deal-level review.

What can look like sandbagging?

Some honest patterns trip the same metrics, so rule them out before you talk to the rep.

- In-quarter deals. Furseth splits a quarter into deals in the pipeline on day one, deals created and closed inside the quarter, and deals pulled forward from future quarters. A rep who beats commit with deals that did not exist at commit time has a strong in-quarter motion. Count only the overage from deals that were already in the pipeline when the commit was called. - Genuine uncertainty. A rep with a few large, uncertain deals will miss in both directions. Their bias averages out and their tracking signal stays near zero, as with Rep B above. - Subjective stages. Furseth notes that when stage entry and exit criteria are loose, stages come down to the seller's judgment. A rep may hold deals below commit because the category rules are vague. Tighten the definitions in commit versus best case before you call it a pattern. - One outlier deal. A single large early close can create most of a quarter's beat. Check how much of the overage sits in one deal. - A changed market or territory. When conditions shift, every rep's history shifts with them. Compare the rep to the team in the same quarter as well as to their own past.

What do you do with a flag?

Treat a flag as a reason to look closer, then fix the process that produced it.

1. Rule out the false positives above. Split the overage into day-one pipeline and in-quarter deals. 2. Walk the flagged deals. Bring the snapshot history (category, close date and amount over time) to a deal review with the rep and the manager. Furseth's habit for whoever owns the forecast is to talk with sales leaders every day, and always before the weekly call, because the model's assumptions have to be tested against what the field is hearing. 3. Change the process, and leave the number alone. Furseth's governance rule: if you keep missing, change the process. Pressuring people into a different number fixes nothing. Look at how deals are qualified, how categories are assigned, how close dates are managed and how the forecast rolls up in the sales forecasting process. 4. Check the comp plan. Sandbagging is usually a rational answer to incentives. A plan that pays a 98% attainment the same as a 70% one, or punishes misses hard and pays richly for overattainment, invites under-committing. The fix playbook is in how to reduce forecast sandbagging. 5. Calibrate while you fix it. Adjust the rep's commit by their measured bias in the roll-up, and keep measuring. A permanent forecast haircut hides the problem one layer up. 6. Show reps their own data. When reps see their buyer signals and their forecast history side by side, some correct on their own. Reviewing deals with both the rep's read and the signals on screen makes a cushion hard to keep.

What changes once sandbagging is under control

When sandbagging falls, three things improve together. Forecast accuracy rises because deals sit in the category they belong in. Pipeline coverage becomes reliable because the forecasted pipeline reflects what reps expect to close. Planning improves because leadership can use the field's number without guessing at a cushion. The goal is a rep forecast whose errors are random, small, and point both ways.

Frequently Asked Questions

How do you detect sandbagging in a sales forecast?

Start with each rep's forecast history. Compare the commit they called at a fixed point, such as week one, to what they closed, for at least four quarters. Sandbagging shows up as misses that all point low, a tracking signal near minus 4, a habit of closing deals that were dated for next quarter, and deals that only move into commit in the final weeks.

What is an example of sandbagging in a sales forecast?

In an illustrative case, a rep commits $400K, $420K, $450K and $480K across four quarters and closes $480K, $500K, $530K and $560K. Every quarter beats the commit by $80K. Pooled bias is minus 15.5% and the tracking signal is minus 4.0. Any single quarter looks like a good result. The four together show a commit that was never a real estimate.

How do you tell sandbagging from conservative forecasting?

Look at the direction of the misses. An honest but cautious rep with uncertain deals misses both ways, so the running sum of errors stays near zero. A sandbagger misses low almost every quarter by a similar margin. Then check where the overage came from: deals already in the pipeline at commit time are evidence, while deals created and closed inside the quarter usually are not.

What is another term for sandbagging?

Common alternatives are lowballing, under-forecasting, under-committing, padding the forecast, and conservative forecast bias. In forecasting math it is negative bias, because the forecast sits below actuals. Finance teams use the same word when a company sets guidance it expects to beat. The glossary entry on sandbagging covers the definition and why reps do it.

What should a manager do when a rep is flagged for sandbagging?

Treat the flag as a question. Rule out false positives first, then walk the flagged deals with the rep and their manager using snapshot history of category, close date and amount. If the pattern holds, fix the incentive or the category definitions that reward it and calibrate the forecast in the meantime. Pressuring the rep into a different number only moves the distortion.

What is forecast sandbagging?

Sandbagging is when a rep deliberately under-forecasts, by holding a likely close in best case or upside instead of commit, dating it into next quarter, or understating the amount. It builds a cushion the rep can beat, and it leaves leadership planning hiring and spend on a number that is lower than the real pipeline.

Put these metrics to work

ORM builds custom revenue forecast models that turn concepts like forecast sandbag detection into prescriptive action for your team.

Schedule a Demo