Skip to main content
The best replenishment plan is only as good as the demand signal behind it. If you are planning from last month’s shipments or a static annual budget, you are always reacting rather than anticipating. Bold generates automatic sales forecasts from your own sales history, using mathematical models that detect trends, seasonality, and growth or decline patterns — at the individual SKU level. Those forecasts feed directly into the purchasing assistant and manufacturing planning, so every proposal is grounded in what the data actually says about your future demand.

​
How Bold builds forecasts

Bold analyses your historical sales data for each SKU and fits the most appropriate statistical model from a set of advanced forecasting algorithms. The models account for:
  • Trend — a steady long-term increase or decrease in demand
  • Seasonality — recurring peaks and troughs tied to the calendar (quarterly cycles, holiday surges, summer lulls)
  • Growth and decline patterns — accelerating or decelerating demand typical of product lifecycle stages
The output is a time-series forecast expressed in sales units per day, week, or month, projected as far forward as your longest supplier lead time requires. Because the forecast is updated automatically each time new sales data arrives, it improves continuously without any manual intervention.
Forecast accuracy improves as more sales history accumulates. For new products or SKUs with fewer than three months of history, Bold blends the available data with category-level averages to produce a usable starting estimate. Review these early forecasts more frequently and adjust them manually if you have specific market intelligence.

​
Review your forecasts

Navigate to Planning → Demand Forecasting to see the forecast view. For each SKU you will find:
  • A forecast chart showing historical sales alongside the projected demand curve
  • The statistical model Bold selected and its confidence interval
  • The adjusted forecast if you have applied any manual overrides (see below)
  • The impact on purchasing: how the forecast translates into replenishment quantities over the next rolling horizon
Use the date range selector to zoom in on specific periods — for example, to verify that a seasonal spike is being captured correctly before a peak trading period.

​
Manually adjust a forecast

Automatic models are powerful, but they cannot know about a key account contract you signed yesterday or a promotional campaign planned for next quarter. Bold lets you layer manual adjustments on top of the statistical baseline without overwriting the model.
1

Open the SKU forecast

In the Demand Forecasting view, search for or filter to the SKU you want to adjust and click on it to open the detail panel.
2

Select the period to adjust

Click and drag on the forecast chart to select the specific weeks or months you want to modify, or use the period picker below the chart.
3

Enter the adjustment

You can override the forecast in two ways: enter an absolute value (e.g., you expect to sell 500 units in week 12) or enter a percentage uplift or reduction (e.g., +30 % for the duration of a campaign). Bold displays the adjusted curve alongside the original model so you can compare them visually.
4

Add a note

Document the reason for the adjustment in the notes field — for example, “Summer campaign, agreed with sales on 2024-05-10”. This gives context to any colleague who reviews the forecast later.
5

Save and propagate

Click Save adjustment. The purchasing assistant and manufacturing plan recalculate immediately using the updated forecast. No separate sync step is needed.
Use adjustments sparingly and for genuinely exceptional events. Over-adjusting the model reduces its ability to learn from real demand signals. If you find yourself adjusting the same SKU every month, review whether the underlying model settings need updating instead.

​
Common adjustment scenarios

Apply a percentage uplift to the campaign period and the two to three weeks following it (demand sometimes pulls forward and leaves a trough after a promotion ends). Set a reminder to remove the adjustment once the campaign is over so the model returns to its normal baseline.
If you have signed a supply agreement with committed weekly volumes, enter those volumes as absolute overrides for the contract duration. This ensures the purchasing plan reflects the contracted commitment, not just the statistical forecast.
For a new SKU with no history, enter a manual launch ramp — your commercial team’s sales estimate by period — for the first three to six months. As actual sales data accumulates, Bold will progressively blend it into the model and reduce dependence on the manual estimate.
Apply a downward adjustment or set the forecast to zero from the planned discontinuation date. This prevents the purchasing assistant from proposing unnecessary replenishment for a product you are winding down.

​
How forecasts drive manufacturing requests

Forecasts do not stop at the purchasing level. Bold uses each SKU’s forecast together with its bill of materials to calculate how much of each component needs to be available — and when. Where the production route involves in-house manufacturing steps, Bold automatically creates manufacturing requests and sends them to the production module, where the production manager can review, adjust, and release them as Work Orders.
Manufacturing requests appear in Production → Planning as proposed Work Orders. The production manager decides when to confirm and schedule them — MRP proposes, the production team decides. See Production (MES) for how Work Orders are managed on the shop floor.

​
Forecast accuracy over time

Bold tracks forecast accuracy per SKU using standard error metrics (MAE and MAPE) and displays them in the forecast detail panel. As your sales history grows and Bold’s models accumulate more signal, accuracy improves automatically. You can use the accuracy metrics to prioritise which SKUs deserve closer manual attention — typically, high-value or long-lead-time items where a forecasting error has the most financial impact.