Help

Help›The engines›Forecast accuracy

Forecast accuracy

Reviewed Oct 29, 2026

View as Markdown
Words

What forecast accuracy measures ​

Forecast accuracy is the forecast's track record: how far Tightly's forecast missed what actually sold, week by week, set beside the misses of doing it by hand. It is measured across the lines on view. Confidence grades individual lines.

When the forecast is scored ​

The accuracy pass runs every Monday at 06:30 UTC and scores the seven completed days ending yesterday in your company's timezone. The current local day is not scored. It compares the forecast for each managed variant with what that variant sold, on your own direct-to-consumer channels, and the drawer names that population as your direct-to-consumer sales. A new tenant has no record until its first full week has been scored. Recorded customer scores update when the scoring pass lands. Changing the drawer selection changes the period shown, not the recorded forecast.

Forecast error and how it is worked out ​

The headline figure is Forecast error (the weighted absolute percentage error, WMAPE): the units the forecast missed by, over or under, added across every selling line, divided by the units sold. It reads as units of error per 100 units sold, and lower is better. Because it adds misses in both directions, a line forecast high and another forecast low do not cancel. Over / under is the other figure: the total forecast against total sales, as a percentage above or below, where opposite misses can cancel.

The by-hand methods Tightly is compared with ​

Every week the same lines are also scored by four methods a planner could run in a spreadsheet, on the same weeks and the same sales:

MethodWhat it predicts
Last week carried forwardlast week's sales again
4-week moving averagethe average of the 28 days before the week
8-week moving averagethe average of the 56 days before the week
Same week last yearthe same week 52 weeks earlier

The comparison in the head of Demand forecasting draws two bars, Tightly and the by-hand method, and says how much less or more error Tightly had, over Lower is better. The drawer shows actual sales and Tightly forecast alongside up to two benchmarks. It opens with the 4-week and 8-week moving averages; Compare changes those selections.

Backtested and Actual results ​

The comparison carries one of two words. Actual results is the forecast as it was served, scored week after week. Backtested appears after the engine itself has changed: today's forecasting engine is re-run over the same closed weeks and scored against the same by-hand methods, so the record describes the engine you are on rather than one that no longer runs. Either way the weeks and lines are the ones sold.

The Forecast performance drawer ​

Pressing the comparison opens the Forecast performance drawer. Actual sales, Tightly forecast and Forecast error % summarize the selected period. The chart and Sales & forecasts table show units for actual sales, Tightly forecast and the selected benchmarks.

The chart starts with monthly totals. Click a month to show only that month's weeks, then click a week to show only its recorded days. Weeks belong to the month in which they start. Period selects a month; Previous and Next move between weeks, and All weeks returns to the monthly overview. The table follows the same selection as the chart.

Daily values appear only where daily comparisons were recorded. An older week without them says Daily recorded comparisons are not available for this week. Weekly totals are not divided into invented daily values.

Choose Forecast accuracy for the method comparison: Actual sales, Forecast · units, Difference · units, Forecast error % and Over / under %. Monthly error is calculated from summed errors and sales, not an average of weekly percentages. The weekly and monthly comparisons use the forecasts recorded for those periods. Information buttons explain the figures.

The record follows the page's window and supported filters. Missing daily values stay unavailable.

When there is no record ​

The comparison in the head is labelled Against doing it by hand and opens the Forecast performance drawer.

Without a comparison, the head shows Not measured, Awaiting first score, No scored weeks or Not read. The drawer gives the reason:

No week has been scored yet.

No scored weeks in this date range. Choose a different date range.

No full week measured yet.

Where the read failed it says Performance unavailable, with a retry button beside it.

Accuracy is not confidence ​

Accuracy measures forecast error weekly. Confidence grades individual lines on each sync. A good accuracy record can coexist with a Watch closely line. See The forecasting engine, Forecast methods and the Monday re-score and Demand forecasting.

For a category selection carried from a plan, the head and drawer read Accuracy unavailable at this scope. No whole-book accuracy score is substituted.

Glossary line ​

Forecast accuracy is the forecast's weekly track record, its forecast error against what sold, beside the error of four by-hand methods.

Commitment forecast selections require eligible DTC channels. Customer orders are not scored as consumer forecasts. Supported account sell-out evidence stays separate in Connect; absent commitment, variant or account scores stay unavailable.