Learn how Halo supports a comprehensive Demand Sensing implementation, and what types of businesses will benefit most.
In our post How Does Demand Sensing differ from Demand Forecasting, we stated that the accuracy of conventional time-series forecasts eventually hits a ceiling. Why? Because historical data is a flawed predictor of supply chain conditions over a short period of time. Even high-volume consumer packaged products with well-understood seasonality established over decades continue to experience high near-term forecast error rates.
Studies have shown that, on average, Demand Sensing reduces forecast error by up to 50%, increases inventory accuracy by up to 20% and optimally deploys downstream inventory. The typical performance of demand sensing systems reduces near-term forecast error by 30% or more compared to traditional time-series forecasting techniques.
Here are some examples of appropriate data sources for building a demand-sensitive forecast:
- Sales orders (historical and future) and shipments
- Final forecast together with forecast accuracy and forecast bias
- Campaigns and promotions
- New product introductions
- POS data from the field
- Economic trends
- Social network sentiment
- Competitive tactics
- Changes in consumer behavior
- Impacts of weather; natural disasters
A Demand Sensing system can be thought of as residing inside the demand planning process. In fact, Demand Sensing uses the consensus demand plan as an input. First, Halo uses Machine Learning forecasting tools with pattern recognition algorithms, automated workflow, and scenario testing to monitor short-term demand. Then, using Machine Learning to interpret a range of demand inputs, the system senses changes to the plan and reacts by making forecast and inventory allocation adjustments.
Some of the benefits companies can experience through a successful Demand Sensing implementation such as this include a 5-10% improvement in cycle-time and on-time delivery, a 15-40% improvement in short-term forecast accuracy, a 50% reduction in Days Inventory On Hand, and improved market responsiveness, higher service levels and less obsolescence through more effective allocation decisions.
As you assess the relevance of Demand Sensing for your supply chain, ponder these points:
- Think of Demand Sensing as daily inventory allocations. Switching from a long-term planning forecast to a daily demand forecast enables better near-term inventory allocation decisions. It may also apply to those organizations that control manufacturing and are able to quickly change the production schedule. Think food and beverage, but not offshore manufacturing.
- Data availability may be more challenging than data blending and analysis. Demand Sensing can incorporate a broad range of demand signals including real-world events such as retail sales, order patterns, promotions, market shifts and weather changes. If necessary, start modestly by looking at POS data and daily orders.
- Don’t be fooled into thinking forecast accuracy is the end-game. Companies with supply chain agility and quick response times are in the best position to leverage a Demand Sensing investment. Said another way, the business value of forecast improvements comes from the ability to use that forecast to reduce downstream costs. If your company can’t link Demand Sensing to a short-term response capability, don’t bother.
Many studies have shown that Demand Sensing can improve short-term forecasts versus traditional time-series forecasting methods. The accuracy of the latter is capped based on the fundamentals of information theory, and the 'perfect fit' forecast is not a function of computing power. Remember, long-term production planning is strategic and necessary. Consider Demand Sensing for short-term tactical gains.
Companies with supply chain agility and quick-response capabilities are in the best position to leverage a Demand Sensing investment. Because Demand Sensing typically operates well inside production lead times, look first for improvements in inventory allocation.