Prescio User guide

Build the demand baseline

Forecast Workbench

Manage demand history, generate Time-series and Contact Rate Forecasts, review methods and control approved versions.

Forecast

Purpose

Forecast is used to manage historical demand, generate forecasts, review method performance, stage changes and maintain the approved forecast history.

Headline measures

The page shows:

  • forecast status;
  • actuals loaded to;
  • 13-week WAPE;
  • forecast bias;
  • forecast coverage.

Main controls

  • Department selector
  • Queue selector
  • Generate Forecast
  • Upload Historical Data
  • Upload Forecast
  • Export Forecast

Forecast tabs

Current Forecast

Shows the approved baseline forecast and the resolved forecast after active demand transformations.

It includes:

  • the forecast outlook chart;
  • approved baseline versus resolved demand;
  • forecast coverage;
  • the Forecast Health Control Centre;
  • Queue-level indications of where review or regeneration may be required.

Use this tab to review the active demand plan and understand whether the approved baseline is complete and healthy.

Historical Data & Normalisation

Shows the demand history used for time-series forecasting and how the history has been conditioned.

Use it to:

  • confirm the latest actual week;
  • inspect historical coverage;
  • upload new history;
  • review anomaly smoothing;
  • review protected events;
  • see transformation-related conditioning;
  • confirm whether the series used by forecasting is representative.

Historical conditioning is intended to produce a fair forecasting series. It does not delete the original actual history.

When a demand transformation is live but not yet sufficiently represented in the normal baseline, Prescio can prevent partial emerging actuals from being double counted. Once sufficient evidence exists, the history can be rebaselined so the new level becomes part of normal forecasting.

Contact Rate Data

Maintains reusable base data for Contact Rate Forecasting.

The demonstration supports:

  • Customer Base — point-in-time base data;
  • Weekly Sales — weighted-offset base data.

Each section contains:

  • retained historical-base versions;
  • retained future-base versions;
  • upload controls;
  • readiness status;
  • an outlook chart joining historical and forecast base values;
  • a visible forecast transition point.

Weekly Sales also provides a weighting configuration. The weighting determines how sales from several weeks contribute to the effective base for a contact rate. This reflects the fact that contacts may occur after the original sale.

Use this tab to:

  1. review available base versions;
  2. upload historical or future base data;
  3. enter useful version names and descriptions;
  4. check coverage and continuity;
  5. configure the Sales weighting where required;
  6. confirm that the selected base is ready for forecasting.

Historical and future versions remain separate and are retained for audit and Scenario use.

Forecast Versions

Shows:

  • pending Queue forecast revisions;
  • the control used to confirm pending revisions;
  • immutable approved Production forecast versions;
  • Queue-level methods and decisions used in each version;
  • model recommendation;
  • backtest WAPE and bias;
  • growth or decision assumptions where applicable.

Use this tab to understand what is waiting to be applied and how an existing approved version was created.

Generating a Time-series Forecast

  1. Select the Department and Queue.
  2. Select Generate Forecast.
  3. Choose Time-series Forecast.
  4. Review eligible methods.
  5. Run the comparison.
  6. Review WAPE, bias, RMSE and the forecast shape.
  7. Select the preferred method. The recommended method is guidance, not a

    8. forced choice.

  8. Review any available growth or pattern controls.
  9. save the result as a pending Queue revision;
  10. use Forecast Versions to confirm the required pending changes.

Generating a Contact Rate Forecast

  1. Select the Department and Queue.
  2. Select Generate Forecast.
  3. Choose Contact Rate Forecast.
  4. Select Customer Base or Weekly Sales.
  5. Select the historical base version.
  6. Select the future base version.
  7. Review aligned historical coverage and future coverage.
  8. Run the eligible forecast methods against the historical contact-rate

    9. series.

  9. Compare methods using derived contact-volume accuracy.
  10. Review the derived demand forecast.
  11. Save the result as a pending Queue revision.
  12. Confirm it through the same Forecast Versions process used by

    13. Time-series Forecasts.

The underlying calculation is:

Historical contacts ÷ effective historical base
= Historical contact rate

Forecast contact rate × effective future base
= Derived baseline demand forecast

The derived demand is persisted as a normal baseline forecast. Downstream Transformations, Requirement, Workforce and Intelligence therefore follow the same path as a Time-series Forecast.

Important forecast principle

Generating a forecast does not immediately change Production. The forecast is reviewed, staged and confirmed. This protects the agreed plan and retains a clear version history.