Supply chain management is getting more complicated as the data to be analyzed for the optimization of supply chains across the globe is becoming more sophisticated. Now, more than methods such as time series regression analysis are needed to plan business and supply chain operations. Demand trends today are changing at hyperspeed, and planners need accurate demand variability analysis to prepare their operations better. This is where demand sensing can be helpful for planners.
What is Demand Sensing?
Demand Sensing is the process of picking up on short-term demand trends for any product so that planners can predict short-term customer buying behavior. The method of demand sensing is generally performed for predicting short-term demand using the analysis of patterns in POS (Point-Of-Sale), social media, Weather, IoT, Internet trends, and economic data to make near-term predictions. Demand sensing helps leverage the high level of granularity in data, analyze daily demand patterns from the end customers, and detect changes in demand in real time.
How Does Demand Sensing Process Work?
The process consists of the following steps:
- Data Collection
The system collects data from different sources, such as consensus demand plans, sales orders, promotions, POS, and inventory data.
Once the system collects data, it performs pre-processing steps. It looks for outliers in the sales history to exclude them from the learning phase.
- Short-Term Forecast Optimization
After pre-processing, the algorithm uses machine learning and pattern detection techniques. It optimizes the short-term forecast using its correlation with forecast bias, extra signals, and Open Order patterns.
- Post-Processing on Weekly Sensed Demand
Weekly post-processing is performed by the algorithm on the optimized sensed demand to cap the demand according to the threshold set in the forecast models and balance promotion uplifts.
- Disaggregation of Weekly Sensed Demand
The weekly sensed demand is disaggregated based on profiles, weight calculated for each profile, and daily split of demand for profiles using the method selected in the forecast model.
- Post-Processing on Daily Sensed Demand
The algorithm post-process daily sensed demand to balance promotion uplift plans and performs backward-forward forecast consumption to balance the baseline demand.
What are the Benefits of Demand Sensing?
This methodology has numerous benefits for a business, and some of these are discussed as follows:
- More Resilient Supply Chain
When short-term demand patterns are available, it becomes easier for the supply chain to adapt to sudden disruption and market volatility. This technology can make supply chains more robust and ready for problems such as the COVID-19 Pandemic.
- Efficient Inventory Management
With the availability of daily demand data, planners can optimize inventory levels, reduce inventory levels, and make supply chain management streamlined while maintaining resilience. This way, businesses can reduce the blockage of capital due to excess inventory and reduce the excess production as well as handling costs.
- Better Predictability
A wide range of data variables and factors, including real-world events such as retail sales, promotions, order patterns, and market changes, is used for demand sensing to understand demand trends. This gives better predictability for short-term future demand, and planners can plan their operations accordingly.
Goals of Demand Sensing
The primary goal of this method is to help planners understand short-term demand trends and then make decisions since what happened some days ago instead of based on data available from the last year. This way, planners can overcome the latency issues they face with time-series statistical methods. In other words, demand sensing focuses on removing any supply chain lag by reducing the time between events and reaction to those events.
This method uses detailed short-term data for making near-future forecasts. Here, the near future can refer to hours or days, depending upon the agility of the supply chain. This forecast, built using AI/ML principles derived from daily POS data, will be more accurate than the daily forecast created from a manually entered day-weight calendar.
What Are the Limitations of Demand Sensing?
Demand sensing is a handy tool for businesses looking to create forecasts for the short term. When this technique is applied to long-term demand goals, the accuracy is limited only. It can be applied to a 4 to 12 weeks horizon only. Other demand planning methods will be used if demand trends beyond 12 weeks are required.
Another limitation of demand sensing techniques is that they are designed for something other than B2B manufacturers. They are more apt for fast-moving consumer goods (FMCG). In these techniques, the planning is done based on the data obtained from the point of sale (PoS), which is available only in B2C industries but not for B2B industries.
This technique for predicting demand trends is too much dependent on data accuracy. Even the slightest inaccuracy in data may lead to inaccuracy in prediction leading to disruption of the entire supply and business operations.
Power of Demand Sensing in Supply Chain with 3SC
Predicting the demand for the near future can be a great advantage for businesses. It can help them quickly optimize their inventory, ensure the availability of manufacturing capacity, align logistics, and plan finances to meet customer demands and avoid stock-outs.
Demand forecasting with traditional time-series analysis is not accurate for short-term prediction. For this, companies require robust, intelligent, and agile advanced analytics solutions that can leverage the capabilities of artificial intelligence and machine learning for big data analytics and provide real-time insights for businesses. 3SC offers intelligent supply chain planning solutions such as Integrated Business Planning, Risk Management & Digital Twin, Enterprise Data Management, and CarbonX, which can analyze big chunks of data and provide insights and recommendations for optimized business operation and seamless supply chain management. With the help of our intelligent business planning solutions, companies can ensure seamless, efficient, and sustainable supply chain management and business operation planning.
How does demand sensing differ from demand forecasting?
Demand sensing is used for short-term prediction of demand trends with a time horizon of 4 to 12 weeks. Demand forecasting is used for long-term forecasts of demand trends with a time horizon beyond 12 weeks.
What are some best practices for demand sensing?
Some of the best practices during sensing demand include starting with sell-in data, using all possible data sources, and considering all external forces to obtain the most accurate prediction of demand trends.
What are some examples of demand sensing?
Some prominent examples of demand sensing can be seen in the Automobile Industry, Fast-Moving Consumer Goods Industry, and Healthcare & Pharmaceuticals industry.
When to use demand sensing?
It is used when the market conditions are volatile, and demand trends change rapidly. Short-term demand prediction can help form a resilient and efficient supply chain in such a condition.