Good demand planning is essential for an organisation's operational and financial performance. How do you avoid the most common pitfalls?
When we talk about demand planning, we refer to all functionalities within the organisation that contribute to streamlining operations versus market demand.
Digitisation in recent years seemed to deliver on its promise of breaking down silo formation and organisational barriers by reducing human interaction. Yet many companies struggle with these very issues to get their demand planning right.
Pitfall 1: Lack of commitment
In many organisations, a forecast is still viewed with scepticism. 'The planning isn't right anyway, because....' is a common view within companies. As a result, human intervention adjusts actual production rather than improving the forecast, creating a self-fulfilling prophecy. At management and executive level, there is then often too little guidance to avoid this pitfall.
There is also often a lack of commitment to break organisational barriers. Often, demand planning gets stuck at the level of middle management within a specific silo. At this level, people then try to align different metrics, which causes the next problem: the bullwhip. By viewing the supply chain as a sequence of metrics, variations are automatically amplified and your demand planning quickly gets out of sync. And not infrequently, this is done within individual spreadsheets rather than an integrated demand planning system.
Pitfall 2: Past results are no guarantee for the future
Many statistical forecasts, regardless of their method, are based on historical trends. This goes wrong for two reasons: the aforementioned internal bullwhip effect but also due to the fact that historical data ages quickly. Especially in a market that is constantly subject to rapid changes in demand.
In S&OP, the challenge is precisely to make predictions on current developments in the market through smart use of Big Data. Statistical data can be used to examine market segmentations and to create baseline forecasts. But at, say, SKU level, it is totally inappropriate.
Pitfall 3: Lack of cooperation
Developments that directly affect demand are not properly discussed or communicated within the organisation beforehand. Think, for example, of a new marketing campaign or the introduction of new products in the market. For the operation, this creates an entirely new situation that must be responded to ad hoc. The operation cannot follow the change in demand and, as a result, the desired result of the marketing campaign is not realised. The result: mutual recriminations within the organisation.
The organisational structure is often the problem. This is mostly vertically organised with many functional silos. Within each silo, new ideas and concepts are developed without involving the rest of the organisation from the start. People prefer to have a ready-made story on the table before consulting with the organisation.
Pitfall 4: Lack of mature forecasting methods
As stated earlier, historical data are a poor predictor of future developments in a rapidly changing market. Many organisations lack an incremental change in their forecasting process. From historical to prescriptive analyses.
Historical data can be used for an initial baseline of information and segregation of markets. Within this subdivision, organisations should take the following steps to avoid the pitfall.
- Descriptive analysis of data management. This is perhaps the most essential step. Not just the collection and management of the data, but especially the method of capture and how the data is accessed to the organisation. Providing data with the right characteristics and parameters is fundamental to its usability.
- Monitoring analysis. Once you have determined which data you want to capture and manage, constant monitoring and updating of information is essential. Not just the data itself, but especially changes in the data sets.
- Diagnostic analysis. The collected data is analysed for variations compared to baseline data and compared with expected developments and which deviations and to what extent can be identified.
- Predictive analysis. Here, diagnostic information is used to determine the new expected outcome. In short, what is the effect of a change in planning or a change in demand.
- Prescriptive analysis. This then answers the question of which corrective actions to take. This uses scenario comparisons and other process control algorithms to arrive at the best possible choice.
Pitfall 5: Lack of innovation
Innovation is a sudden discovery. Innovation is an incremental process where an organisation takes a small step every day to improve itself. Similar to ecosystems that continuously try 'new' things to see if it works. This is precisely also the strength of startups that make many attempts to serve the market in a relatively short time with a mimimum of process controland hardly any overhead. This allows for very quick switching and adjustments if one has gambled on the wrong horse.
In this process, one often uses 6S approach to determine the maximum movement space. Relative to the optimal flow, a UCL and LCL are established within which functions within the organisations are allowed to move freely. Responsibility, both operational and financial, is fully delegated to the lowest possible level. If one exceeds the upper or lower limit, horizontal cooperation is required to get back within the limits or leave the chosen path.
It is important that the parties involved use the same range within these frameworks in terms of performance, planning horizon, the (shared) risks and the exception conditions used.
