Sandbagging: Why Sellers Understate Forecasts and What Fixes It

Sandbagging is understating what you expect to achieve, or holding back business that could be closed now, so that a target is easier to beat. In sales it appears in two forms: a forecast deliberately kept below what the seller privately believes, and a closable deal quietly allowed to slip into the following period. The word carries a moral tone that is usually unhelpful, because in most organisations the behaviour is a rational response to how quotas are set and how compensation is structured. Treating it as a character flaw produces a difficult conversation; treating it as a system output produces a fix.

The primary signal

A seller whose closed revenue regularly and substantially exceeds their commit, across several periods, where much of the excess was never committed at any point during the period. One strong quarter is luck; the same pattern four times is a tendency. The measurement that reveals it is forecast accuracy calculated in both directions, which most teams do not do because only over-commitment feels like an error.

Supporting signals

Deals closing in the first days of a period that had clearly been ready in the last days of the previous one. Close dates pushed repeatedly by exactly one period. Forecasts that jump sharply in the final week. Qualified opportunities entered into the system late, which is the harder-to-detect version because the deal never appears in a forecast at all.

A worked illustration

Consider a seller with a quarterly quota of ₹50,00,000. Over four quarters they commit ₹40,00,000, ₹42,00,000, ₹38,00,000 and ₹41,00,000, and they close ₹56,00,000, ₹58,00,000, ₹54,00,000 and ₹57,00,000. Every quarter is a success by attainment, and every quarter the forecast understated the result by roughly a third. The business planned four times on numbers that were materially wrong in a consistent direction. Weekly snapshots would show something more specific: that much of the delivered revenue appeared in the commit only in the final fortnight, and that a portion closed in the opening days of the following quarter. That is not a run of good fortune, and it is visible only because the forecast history was stored.

Ratcheted quotas

Where next period's target is set from this period's result, exceeding a quota is punished with a harder quota. A seller who delivers far above target once will spend subsequent periods chasing a number derived from an exceptional quarter. Landing just above the line is the strategy the system teaches, and it is difficult to unlearn once taught.

Accelerators and thresholds

Compensation plans that pay a higher rate above a threshold create a specific incentive to move revenue across period boundaries. A seller comfortably above target this quarter earns more by starting the next one with a large deal than by closing it now, particularly where the threshold resets. The maths is straightforward and the behaviour follows it.

Asymmetric consequences

Where missing a commit is treated as a serious failure and exceeding one is celebrated, the safe forecast is a low one. This is the cheapest of the three causes to address, because it lives in how reviews are conducted rather than in a plan document, and it can be changed within a quarter.

Start with the compensation structure, because it dominates everything else. Annual rather than quarterly thresholds, or smoothing that reduces the value of shifting revenue between periods, removes most of the mechanical incentive. Set quotas from territory potential and capacity rather than from the previous result, so overachievement is not self-punishing.

Then change how accuracy is discussed. Report under-forecasting as an error with the same weight as over-forecasting, and show each seller their own calibration history across several periods. Most people forecast noticeably better once they can see their own pattern, and the conversation becomes technical rather than accusatory. Finally, make early accuracy the measure that matters, since a forecast that becomes correct on the final day provided no value at all.

Be careful about attributing motive. Under-forecasting has at least three causes that look identical in a spreadsheet: deliberate timing, learned caution after being penalised for a miss, and simple poor calibration. The remedies differ substantially, and applying the wrong one, particularly treating caution as dishonesty, tends to make forecasting worse rather than better.

Watch for the opposite pattern at the same time. Teams that punish misses severely often produce deal-pulling rather than sandbagging: revenue brought forward with discounts, which costs margin and empties the following period. A cluster of unusually discounted deals in the final days of a quarter is the tell, and it comes from the same root cause, which is a period boundary that matters far more than it should.

These four patterns all distort a forecast and call for different responses.

Only one of the four is usually described as dishonest, and all four produce forecasts that mislead the business by similar amounts. Treating them as a family of measurement problems rather than as a question of integrity tends to produce faster improvement and considerably less friction.

Everything here depends on having a durable record of what was said and when. Start with what a CRM is, then look at CRM versus Excel for why spreadsheets cannot hold forecast history, AI CRM for spotting deals that stall or move, sales automation, lead management software, and sales use cases.