Demand forecasting
We can help you design and deliver an exceptional demand forecasting process.
COMMON FORECASTING CHALLENGES
Common forecasting challenges we're asked to solve
Inaccurate demand forecasts
Seasonal demand fluctuations
Poor-quality historical data
New product uncertainty
Changing customer demand
Forecast bias and overrides
INTRODUCTION
The need for forecasting
Demand forecasting is a critical business requirement to support a range of business functions, including logistics, production and finance. It is only with an agreed and considered forecast that functions can adequately plan capacity, inventory, labour and cash-flow.
Without a reliable forecast, planning becomes reactive. Businesses risk carrying too much inventory in some areas while experiencing shortages in others, with knock-on effects for costs, customer service and operational efficiency.
Effective forecasting also provides a common view of expected demand across the business. This creates a stronger basis for decision-making, allowing assumptions to be challenged, changes in demand to be identified and plans to be adjusted before they become operational problems.
TYPICAL RESULTS ACHIEVED
- 10 - 20% improvement in accuracy
- Improved service levels
- Reduced stockouts & excess inventory
- More confident purchasing decisions
TRUSTED BY LEADING ORGANISATIONS
DEMAND FORECASTING
The challenges of forecasting
When forecasting demand, it is key that all elements that can impact sales are identified, assessed and incorporated. Consequently, demand forecasting is not only a statistical exercise, but also requires cross-functional input from teams including sales, supply chain, finance and production.
Many businesses fail to have a cross-functional forecasting process, and often rely only on statistics. This can lead to a lack of business confidence in the forecast and it is not uncommon to find different departments developing their own isolated forecasts; sales, supply chain, production and finance will often see the future very differently.
OUR APPROACH
A proven process for better forecasting
Defining the hypothesis for future sales demand is an important first step. The hypothesis is likely to be formed from a strategic intention alongside suppositions on the economic environment, market opportunity and competitor performance.
With the hypothesis defined, further consideration can then be given to what data will need to be collected to prove/disprove the supposition. This may take the form of tracking sales calls made, volume demanded by targets, sales executives’ perceived probability of success, conversion rates into actual sales, forecasting of future demand from new clients, or tracking how that forecast compares to quantity actually demanded.
At Step 2, our forecasting consultants will convert the business hypothesis and supporting assumptions that were generated into a mathematical model, and will design a data capture process with templates. The templates will incorporate automatic data validation as well as useful dynamic dashboards for the stakeholders to observe.
Additionally, a consolidation tool to collate the data being tracked can be created, allowing for easy assimilation of data points.
Using the Data Tracking Templates from Step 2, our demand forecasting consultants will test the validity of the hypothesis as data develops, and potentially collect alternative data in the event of early detection of variances.
Any data or useful insights that could be used to help the sales team will be passed on, and additionally, dashboard reporting of performance on a monthly basis will be generated, so that different approaches can be compared and used for performance evaluation / improving sales techniques.
A scenario analysis will be run to assign probabilities to various events that could impact sales levels on the mid-to-long term horizon. Typical event variables could include substitute products causing a change in equilibrium on the supply-demand curves or regional economic conditions resulting in a shock to aggregate demand.
The outputs from this scenario analysis will be run through a Monte Carlo simulation in order to generate a prediction interval for the baseline forecast, within which it is likely future sales will fall.
Our consulting team will, using the models and the data from the preceding four steps, design an easy to use forecasting tool to allow the client to continue tracking and forecasting sales independently.
Where deep integration of forecasting tools may be required with a clients ERP system, or integration with MRP functionality, then the consultants can support system selection and process implementation as the concluding steps.
FEATURED CASE STUDY
Supply and demand planning case study
Supply and demand planning for an industrial manufacturer
HERE TO HELP
Next level demand forecasting
A good demand forecast can only be the result of a good demand forecasting process. That process needs to be fast and dynamic for all participating functions, and it needs to have cross-functional consensus.
The first step needs to be a statistical ‘baseline’ expectation which is then passed through each department for sense-checking and input. Varying additions or constraints may need to be applied, including marketing promotions, production capacities, logistics capacities, supplier issues and inventory constraints.
Depending on client requirements we can take several approaches to developing a demand forecast.
INSIGHT
Forecasting articles & advice
5 Issues That Will Break Your S&OP Process
Sales & Operations Planning (S&OP) is a powerful decision-making process that provides a platform to balance supply and demand on a mid-to-long term horizon. The
How to Enhance Adaptability in Supply Chain Planning
The supply chain is no stranger to change. In fact, we are so used to being chameleons, adapting to the latest innovation or problem, that
The Difficulties in Applying Machine Learning to Supply Planning
The supply chain industry has been adopting machine learning and applying it to demand planning for quite some time now. We’ve seen its success and