# Forecast Override Log Worksheet > Source: https://netray.co/tools/forecast-override-log-worksheet > Publisher: Netray (netray.co), AI and integration engineering for ERP and manufacturing systems. > Section: Free Tools ## Short answer Log every manual forecast override and measure whether it improved accuracy against the statistical baseline. ## Category supply-chain ## Audience - Demand planner - Sales operations - S&OP lead ## Tool **Kind:** worksheet **Mode:** editable **Columns:** #### SKU **Key:** sku **Type:** text #### Requested by **Key:** requester **Type:** select **Options:** - Sales - Marketing - Finance - Management - Demand planner #### Reason **Key:** reason **Type:** select **Options:** - Promotion - Customer order - New customer - Market intelligence - Correct model - Other #### Statistical forecast **Key:** stat **Type:** number #### Adjusted forecast **Key:** adj **Type:** number #### Override size **Key:** delta **Type:** number **Expr:** adj-stat **Format:** number #### Actual **Key:** actual **Type:** number #### Statistical abs error **Key:** stat_err **Type:** number **Expr:** abs(actual-stat) **Format:** number #### Adjusted abs error **Key:** adj_err **Type:** number **Expr:** abs(actual-adj) **Format:** number #### Error reduction **Key:** gain **Type:** number **Expr:** stat_err-adj_err **Format:** number **Help:** Positive means the override helped. **Rows:** **Sku:** GR-2210 **Requester:** Sales **Reason:** Customer order **Stat:** 1200 **Adj:** 1500 **Actual:** 1480 **Sku:** GR-2214 **Requester:** Marketing **Reason:** Promotion **Stat:** 600 **Adj:** 900 **Actual:** 700 **Sku:** VX-0415 **Requester:** Management **Reason:** Market intelligence **Stat:** 3400 **Adj:** 3000 **Actual:** 3450 **Sku:** VX-0420 **Requester:** Demand planner **Reason:** Correct model **Stat:** 900 **Adj:** 820 **Actual:** 815 **Sku:** PT-7001 **Requester:** Sales **Reason:** New customer **Stat:** 250 **Adj:** 400 **Actual:** 260 **Totals:** #### Total statistical error **Key:** stat_err **Agg:** sum **Format:** number #### Total adjusted error **Key:** adj_err **Agg:** sum **Format:** number #### Net error reduction **Key:** gain **Agg:** sum **Format:** number #### Net override volume **Key:** delta **Agg:** sum **Format:** number **Summary:** #### Forecast value added **Key:** fva_pct **Expr:** if(gt(sum_stat_err,0),sum_gain/sum_stat_err*100,0) **Format:** percent **Primary:** true **Description:** Share of statistical error removed by overrides. Negative means overrides made things worse. **Notes:** Fill actuals after the period closes. Summarise by requester in Excel to see whose adjustments add value. Positive FVA means overrides help, negative means they hurt. ## Content **Intro:** Manual overrides are sometimes the smartest part of a forecast and sometimes the biggest source of bias. This log records each override with its requester and reason, then compares the error of the adjusted forecast with the statistical baseline once actuals are in. **Sections:** #### How to use the worksheet **Body:** Record each override when it is agreed: SKU, who asked, why, the statistical forecast and the adjusted one. After the period closes, add the actuals. The worksheet computes absolute error for both versions, the error reduction per row, and the total forecast value added across the log. Export to Excel and pivot by requester or reason. #### Reading forecast value added **Body:** Forecast value added is the share of statistical error removed by the adjustments. A positive result means that human input improved the forecast. A negative result means overrides, in total, increased error. Break it down by reason: promotions with evidence often add value, while optimistic sales uplifts with no evidence often subtract it. **Bullets:** - Positive FVA: overrides helped - Negative FVA: overrides hurt - Slice by requester and by reason - Look at net override volume for bias #### Using the findings **Body:** Share the results with each requester in the demand review in a neutral way. If one group consistently pushes volumes up without result, require evidence for their overrides, such as signed orders. If the planner correcting the model adds value, keep that route open. The aim is a fair process, not blame. #### Common mistakes **Body:** Logging only large overrides hides the many small ones that add up to bias. Recording actuals against a version of the forecast that was changed after the fact makes the comparison meaningless. Freeze both versions at the time of the review and compare only against those. **Cta:** Netray can capture overrides from your planning system and ERP automatically and report forecast value added by requester every cycle. **Faq:** #### What is forecast value added? **Answer:** Forecast value added measures whether a step in the forecast process, such as a manual override, improved accuracy compared with the previous step. It is calculated as the reduction in error versus the baseline. #### Should I allow manual overrides at all? **Answer:** Yes, when they carry information the model cannot see, such as signed orders or planned promotions. Track them, require a reason and review the results so poor overrides are discouraged. #### Which error measure should I use? **Answer:** Absolute error in units is simple and clear in this worksheet. Many teams also report percentage or weighted error. Whatever you choose, use it consistently for both versions. #### How many cycles of data do I need? **Answer:** At least three to six cycles before drawing conclusions about a group, because a few large items can dominate any single cycle. ## Search terms forecast override log, manual forecast adjustment tracking, forecast value add override, demand planning override analysis, judgmental forecast adjustment log, forecast overrides excel, sales override forecast accuracy --- Talk to an engineer: https://netray.co/contact?intent=erp_ai | Machine catalog: https://netray.co/agent/catalog.json