Forecasting and optimisation

Forecast and optimise without leaving Orthantic.

Orthantic brings forecasting and optimisation into the planning workflow. Teams can use these capabilities in the same platform as the plan instead of moving data between separate tools.

Stop moving forecasts between tools.

In many companies, the forecast is created in a separate tool and then pasted into the plan. That adds delay and makes the process harder to manage.

Planning teams then have to work with results they did not create in a system they do not control. If assumptions change, the process starts again.

Orthantic keeps the forecast close to the plan. Teams can run it, review it and continue planning in one environment.

That means less manual handoff, less waiting and fewer chances for the data to drift out of step.

Use forecasting methods that fit your data.

Choose the series, horizon and frequency. Orthantic can evaluate different forecasting methods against your history and support selection at the series level.

Different products and markets behave differently. A method that works well for one series may not work well for another, so the platform supports a more tailored approach.

  • AutoARIMA
  • AutoETS
  • AutoTheta
  • Dynamic Optimised Theta
  • Holt-Winters
  • TBATS
  • Croston
  • Seasonal naïve

Statistical baseline, consensus plan and actual

Thousands of units. Actual stops after July; the interval band opens from there.

    Illustrative figures. The band widens with horizon, which is the honest shape for a statistical forecast — the further out it reaches, the less it should pretend to know.

    Keep the statistical forecast and planner judgement separate.

    A statistical forecast is a starting point. Planners still need to apply judgement for things like promotions, customer changes or market events.

    Orthantic keeps overrides separate from the baseline, so teams can review what changed and why.

    That gives teams a clearer planning process and a better way to learn which changes improved the plan.

    Consensus Plan = "Statistical Baseline"["Units"] + "Planner Override"["Units"]

    Optimise decisions in the same planning workflow.

    Teams should not have to export the plan to another tool just to test a better allocation. Orthantic helps you review options in the same environment as the plan.

    Search the space, not two points of it.

    Compare more than one option and rank outcomes on the measures that matter, such as margin, service level, cost or cash.

    Allocate under real constraints.

    Work with real constraints such as demand, supply, capacity, service levels, budgets and business rules. Review the result and decide how to use it in the plan.

    A fixed total, allocated differently

    The same £13.0m, distributed to a different answer.

      Illustrative figures. £1.1m leaves paid search and £0.8m leaves events; trade promotion takes most of it. Brand is held flat by a constraint. The total is unchanged, so this is a reallocation rather than a cut.

      Where this stops

      Orthantic helps teams test and compare planning options quickly. It is designed to support practical planning decisions inside the workflow.

      If a problem needs a more specialised optimisation tool, that should be evaluated separately.

      What it takes off your plate.

      How forecasting and optimisation work today compared with running them inside the plan
      Today Inside the model
      A script one person maintains and can run A step in the plan any planner can run
      One method chosen for the whole catalogue Best fit selected per series, scored on your history
      Results pasted in, then immediately static Written back into the model, live from that point
      Overrides blended in, so the baseline is lost Overrides in their own layer, with reasons attached
      Forecast accuracy argued from memory Baseline and override each measurable after the fact
      Two options compared, because a third costs a day As many as the question deserves, at full detail

      Run it against your own history.

      The best way to judge the workflow is to test it against the data and planning questions you already have.