Demand Planning

Limitations of the traditional approach to forecasting

Demand planning starts with a fully automatically generated forecast of future demand based on historic demand. This forecast is dramatically more accurate than alternatives due to a radically different approach which avoids the fundamental flaws of traditional forecasting approaches. Most notable flaw is that they assume future demand can be expressed as series of exact numbers. Soon they run into the limits of achievable forecast accuracy, especially for items that have slow, intermittent, or lumpy demand; even branding some items “unforecastable”.

Embrace uncertainty

Wahupa SCM on the other hand embraces the inherent uncertainty and expresses forecasts in terms of probability distributions. This is fundamentally the only way accuracy can be obtained for intermittent demand patterns, but it is also more accurate for the fastest, smoothest demand patterns, and anything in between. No item is “unforecastable”; it simply has a different probability distribution, which might be more costly to provide customer service for.

Automatic baseline

Wahupa SCM will fully automatically find certain patterns - such as trends, level shifts, and seasonality – and utilize these to improve baseline forecast accuracy. Often however, other factors impact demand. These could be actively shaped by the company through promotions, pricing changes, or advertising, but could also be external factors such as the weather or competitor behaviors. If data exists in ERP, planning systems, spreadsheets, or elsewhere it can be loaded into Wahupa SCM to allow it to further improve accuracy.

Scientifically sound human input

If desired, planners can then further enrich the forecast to improve accuracy. This is done in a scientifically sound manner that does not introduce bias, allowing not just consensus but buy-in from all parties involved. Ideally, enrichment occurs by providing information that allows the system to determine impact on demand, rather than simply overriding the demand quantities themselves. If on the other hand the planner wishes to explicitly override demand quantities, past performance of the planner is taken into consideration to determine the effect on uncertainty, which in turn will impact safety buffers.