Supply chain · Power BI template
Demand Forecast Accuracy Power BI template
Dark Power BI demand planning template: forecast accuracy (1 − WMAPE), bias, statistical vs final forecast, forecast value added and largest SKU errors.
By Nawaf Sharaf · Founder & Reliability Engineer · Updated
Free for commercial useLight version: same layout, colours re-solved and checked for WCAG AA contrast.
BetaOpens in Power BI Desktop with 4 tables of sample data and every visual bound. See what’s inside
Free, no sign-up. The PBIP with sample data includes the background, theme and a README. 1920×1080 fits a new page in a Fluent 2 report; 1280×720 and 2× fit Classic pages. Downloads carry a small “Made with VizForge” mark (sign in to the studio to remove it).
- Layout
- Plain header, 2×2 KPI block beside actual-vs-forecast lines, category + region columns and error table
- Page size
- 1280×720 px (16:9, Power BI default)
- Theme
- Plum Dusk
Open the sample-data PBIP in Power BI Desktop
Checked against Microsoft Learn's Power BI Desktop projects documentation (as of Nov 2026).
- 1
Unzip to a short path and open the .pbip
Extract the zip to a short local folder such as C:\PBI (Windows limits project paths to 260 characters, and Desktop can't save a project straight to OneDrive or SharePoint), then double-click the .pbip file or use File → Open in Power BI Desktop. Microsoft still lists projects and TMDL as preview features (as of Nov 2026): if Desktop won't open the project, go to File → Options and settings → Options → Preview features, tick “Power BI Project (.pbip) save option” and “Store semantic model using TMDL format”, then restart Desktop.
- 2
Select Refresh once
The project ships without the local data cache (.pbi/cache.abf), so Desktop opens the full model definition without data. Select Home → Refresh: every table is built from values written into its Power Query step, so there is no data source to sign in to.
- 3
Save as .pbix or connect your data
Use File → Save as to keep a single .pbix file, or open Transform data and replace each table's Source step with your own query, keeping the column names so the measures and visuals keep working.
What’s inside the Forecast Accuracy sample model
Rendered from the same definition the download writes as TMDL, so this list always matches the file.
Twelve weeks of weekly demand for 12 fictional grocery SKUs in three regions (Aug to Oct 2026): actual units shipped, the statistical forecast and the final forecast after planner overrides. Overrides improve most lines but push drinks forecasts consistently high. Units: Units (cases).
Tables and columns
Calendar
Date table · 2026, one row per dayDate table generated in Power Query (List.Dates), marked as the model's date table.
- Date (date)
- Year (whole number)
- Quarter (text)
- Month (text)
- Month Name (text)
- Week Start (date)
- Weekday (text)
- Fiscal Year (text)
SKU
Dimension · 12 rowsProducts with category.
- SKU (text)
- Category (text)
Region
Dimension · 3 rowsDemand-planning regions.
- Region (text)
Forecast
Fact · 432 rowsOne row per SKU, region and week.
- Week (date)
- SKU (text)
- Region (text)
- Actual Units (whole number)
- Statistical Forecast (whole number)
- Final Forecast (whole number)
Relationships
Many-to-one, single direction. Inactive relationships are switched on inside a measure with USERELATIONSHIP.
- Forecast[Week] → 'Calendar'[Date]
- Forecast[SKU] → SKU[SKU]
- Forecast[Region] → Region[Region]
DAX measures
All measures live in a dedicated Key Measures table. Copy any of them into your own model; rename the table and column references to match.
Actual UnitsUnits actually shipped.
Actual Units =
SUM ( Forecast[Actual Units] )Format string: #,0
Final ForecastForecast after planner overrides.
Final Forecast =
SUM ( Forecast[Final Forecast] )Format string: #,0
Statistical ForecastForecast from the statistical model.
Statistical Forecast =
SUM ( Forecast[Statistical Forecast] )Format string: #,0
Absolute ErrorSum of absolute errors at SKU-region-week level.
Absolute Error =
SUMX ( Forecast, ABS ( Forecast[Final Forecast] - Forecast[Actual Units] ) )Format string: #,0
WMAPEWeighted mean absolute percentage error.
WMAPE =
DIVIDE ( [Absolute Error], [Actual Units] )Format string: 0.0%
Forecast Accuracy1 − WMAPE.
Forecast Accuracy =
1 - [WMAPE]Format string: 0.0%
Statistical AccuracyAccuracy of the statistical forecast alone.
Statistical Accuracy =
1 - DIVIDE ( SUMX ( Forecast, ABS ( Forecast[Statistical Forecast] - Forecast[Actual Units] ) ), [Actual Units] )Format string: 0.0%
Forecast Value AddedAccuracy points gained (or lost) by planner overrides.
Forecast Value Added =
[Forecast Accuracy] - [Statistical Accuracy]Format string: +0.0%;-0.0%
Bias %Positive = over-forecasting.
Bias % =
DIVIDE ( [Final Forecast] - [Actual Units], [Actual Units] )Format string: +0.0%;-0.0%
What each visual is bound to
The PBIP opens with these fields already on each visual.
| Visual | Type | Fields |
|---|---|---|
| CategorySlicer | Slicer | SKU[Category] |
| RegionSlicer | Slicer | Region[Region] |
| Forecast accuracyCard | Card | [Forecast Accuracy] |
| BiasCard | Card | [Bias %] |
| Actual vs forecast by weekLine chart | Line chart | Calendar[Week Start], [Actual Units], [Final Forecast], [Statistical Forecast] |
| Statistical accuracyCard | Card | [Statistical Accuracy] |
| Forecast value addedCard | Card | [Forecast Value Added] |
| Accuracy by categoryColumn chart | Column chart | SKU[Category], [Forecast Accuracy], [Statistical Accuracy] |
| Bias by regionColumn chart | Column chart | Region[Region], [Bias %] |
| Largest errors by SKUTable | Table | SKU[SKU], SKU[Category], [Absolute Error], [WMAPE], [Bias %] |
Questions this dashboard answers
Each one is answered by a visual on the page, from the sample model's own tables and measures.
- How accurate is the final forecast (1 − weighted MAPE), and is it improving?
- Is the forecast biased (consistently over or under)?
- Do planner overrides add value over the statistical forecast (forecast value added)?
- Which SKUs carry the largest absolute errors?
About this dashboard layout
Demand planners are judged on accuracy and bias, and they need both because they hide different problems. Accuracy here is 1 − weighted MAPE: absolute errors are summed at SKU, region and week level and divided by actual units, so a large error on a fast seller counts for more than the same percentage on a slow one.
Bias is the signed error as a share of actuals. A forecast can look accurate on average while being consistently high, which fills the warehouse; the sample shows exactly that in drinks, where planner overrides push the forecast up every week.
Forecast value added compares the final forecast with the statistical forecast it started from. If overrides make accuracy worse, the planning meeting is adding work and error; the accuracy-by-category chart shows both forecasts side by side so the discussion can be specific.
The weekly line chart plots actual units against the final and statistical forecasts, and the table ranks SKUs by absolute error with WMAPE and bias. The sample is 12 weeks for 12 fictional grocery SKUs in three regions.
Visuals on the page
Positions are Power BI page pixels. Enter them in Format → General → Properties → Size and position, or use the PBIP download where they are already set.
| Visual | Type | X | Y | Width | Height |
|---|---|---|---|---|---|
| Category | Slicer | 920 | 24 | 160 | 40 |
| Region | Slicer | 1096 | 24 | 160 | 40 |
| Forecast accuracy | Card | 40 | 120 | 168 | 80 |
| Bias | Card | 296 | 120 | 160 | 80 |
| Actual vs forecast by week | Line chart | 544 | 120 | 696 | 208 |
| Statistical accuracy | Card | 40 | 248 | 168 | 80 |
| Forecast value added | Card | 296 | 248 | 160 | 80 |
| Accuracy by category | Column chart | 40 | 376 | 312 | 304 |
| Bias by region | Column chart | 400 | 376 | 312 | 304 |
| Largest errors by SKU | Table | 760 | 376 | 480 | 304 |
Custom visuals for this layout
Built-in Power BI visuals cover most slots above. These slots also suit a custom visual; each page covers the data shape, DAX and a prompt to generate it as a .pbiviz.
Already have a report?
Brand Makeover tries Demand Forecast Accuracy first on every page of your Power BI project (.pbip zip) where it fits: the page's visuals have to go into its 8 data slots without overlaps. Pages it can't seat get another catalog layout or a brand grid. Queries, filters and bookmarks are left as they are. Free, in your browser.
Opens with the Analyst / Dense preset, because the Executive preset doesn't use this layout (8 data slots; this preset uses templates with at most 7).
Restyle my report into thisSwap a slot for a custom visual
Pick any slot and VizForge AI writes a custom visual (.pbiviz, with TypeScript source) sized to it, in this template's colours and shaped for the fields it is bound to.
Generate one that fits this layout.
Card · 168×80 px · bound to [Forecast Accuracy]
1, 2 or 5 credits depending on the brief · 5 free on signup. Sign in or create a free account on the next screen; the brief is saved for 30 minutes.
A Power BI card visual titled "Forecast accuracy". It sits in a 168×80 px slot on a designed report page. Match the report's brand: data colours #E879F9, #FDBA74, #A5B4FC, #FACC15, #2DD4BF, #FB7185; background #211730; text #F3EEF8; font Segoe UI. Transparent visual background so the page design shows through; no visual border. In the sample model it is bound to [Forecast Accuracy]; keep the same data roles so the fields drop straight in.
Not sure what to build? Browse the visual catalog or open the template in the design studio and hand off any slot from there. Or start in the AI generator with a blank brief.
Colours and theme
Dark plum Power BI theme with magenta, apricot and lavender data colours; a warm alternative to navy for marketing, media and audience dashboards.
Use the background in an existing report
Prefer to keep your own report? Apply the design to it instead of opening the PBIP.
- 1
Download the background
Take the 1920×1080 PNG for a new page in a Fluent 2 report (the default size for new pages), or the 2× PNG (2560×1440) for sharp text on a 1280×720 Classic page. Power BI scales the image to the canvas.
- 2
Set the canvas background
Format page → Canvas background → Browse. Set Image fit to Fill and transparency to 0%.
- 3
Import the theme
Design ribbon → Import theme (in the Power BI service: View → Theme pane), then pick the downloaded JSON. Data colours and fonts now match the design. Older Desktop builds: View → Themes → Browse for themes.
- 4
Place your visuals
Size each visual to the positions above with background and border off, or open the PBIP project where they are already placed.
Questions about this template
Does the Demand Forecast Accuracy template come with sample data?
Yes. The “PBIP with sample data” download contains a semantic model with 4 tables and 9 DAX measures, and every visual on the page is already bound to it. The sample: Twelve weeks of weekly demand for 12 fictional grocery SKUs in three regions (Aug to Oct 2026): actual units shipped, the statistical forecast and the final forecast after planner overrides. Overrides improve most lines but push drinks forecasts consistently high. The data is written into each table's Power Query step, so there is no data source or sign-in.
Is it a .pbix file?
It is a Power BI project (PBIP): a folder of text files that Power BI Desktop opens like a .pbix. Open the .pbip file, select Refresh once to load the sample rows, then use File → Save as if you want a single .pbix.
What visuals are in the Demand Forecast Accuracy template?
The page has 10 visual areas: 2 slicers, 4 cards, 1 line chart, 2 column charts, 1 table. Their exact positions are listed on this page so you can match them in Power BI's Format → General → Properties panel, or skip that step by downloading the PBIP project.
Which theme does the Demand Forecast Accuracy template use?
It ships with the Plum Dusk theme (dark tone). Download the theme JSON here, or open the template in the design studio and switch to any other theme or your own brand colours.
Is there a light version of the Demand Forecast Accuracy template?
Yes. Switch the preview to Light and every download (background, theme JSON and PBIP) comes in that version. The layout is identical; the page, card and text colours are re-solved for a light page, and the accent, status and data colours keep their hues but are adjusted until text meets WCAG AA (4.5:1) and chart colours reach 3:1 against the cards.
How do I use the background image in Power BI Desktop?
Select the page, open Format page → Canvas background, choose Browse and pick the PNG, then set Image fit to Fill and transparency to 0%. Give your visuals a transparent background and no border so the designed panels show through.
Can I open the PBIP project in Power BI Desktop?
Yes. PBIR is the default report format in Power BI Desktop since September 2026, and the download is a PBIP project: unzip it and open the .pbip file, select Refresh once and the sample data loads into visuals that are already bound (the empty PBIP leaves the containers unbound for your own data). Microsoft still lists projects and TMDL as preview features (as of Nov 2026), so if Desktop won't open it, enable them under File → Options and settings → Options → Preview features.
Build guides for this dashboard
Guides to the same dashboard job: which custom visuals it needs, the DAX behind them and prompts to generate them.
Dashboard job
Best Power BI Visuals for Supply Chain KPIsThe best Power BI visuals for supply chain KPIs — OTIF bullets, inventory SPC, network flow, OTD…
Team
Power BI for supply chainOn-time-in-full, inventory cover and supplier spend: the flow views planners and buyers check before the daily stand-up.
KPI cards for this layout
Headline numbers drawn by one DAX SVG measure each, for the Card visual or a table row.
Want the headline numbers to look like cards rather than plain values? Browse the KPI card gallery or build your own in the SVG measure builder.
Related templates
Keep exploring
61 hand-designed Power BI page layouts by department.
80 contrast-checked theme JSON files, light and dark.
Page backgrounds at 1280×720 and 1920×1080 with visual positions.
3,034 icons in 40 categories as SVG, PNG, DAX or theme icon JSON.
Make the Demand Forecast Accuracy template yours
Swap in your brand colours and logo, move panels, add icons and generate missing visuals with AI, then export the background, theme JSON and PBIP project.
The design studio for Power BI. Free, in your browser, no sign-up.