An introduction to quantitative and qualitative inventory forecasting models

Demand forecasting models

Whether you’re a manufacturer, wholesaler or retailer, forecasting future demand or customer orders is the logical starting point of all business planning activity, including inventory purchasing. It’s therefore important to choose the right demand forecasting models to aid management decision-making.

If you can accurately forecast demand, you can optimise your inventory levels, improve fulfilment rates and reduce carrying costs, all of which ultimately affect your bottom line. To improve forecasting accuracy you need to chose the right inventory forecasting model.

Types of inventory forecasting models

There are two top level inventory demand forecasting models to consider when calculating demand: the quantitative forecasting model and qualitative forecasting model. In general, qualitative forecasting is based on subjective opinions and insights, whereas quantitative forecasting is more focused on using historical demand data in statistical calculations to predict the future.

Quantitative forecasting

Quantitative forecasting takes historical demand data and combines it with mathematical formulas to determine future performance. For this reason it is also often called statistical demand forecasting. Data sets can go back decades, can be run for the last calendar year, or can be based on the previous few week’s consumption.

Quantitative forecasting models take into account factors such as demand trends and seasonality to help make the predictions more accurate. They rely on having sufficient, good quality data about the past to make a reasonable assessment about the future.

Time series analysis
Time series analysis is perhaps the most common statistical demand forecasting model. It examines patterns in past behaviour over time to forecast future behaviour. There are two main types used in quantitative forecasting:

Moving average forecasting: This takes a previous period’s demand data (e.g four week’s of sales data) and calculates the average demand over that period (average sales per week), then uses this average as the forecast amount for the coming period.

The first drawback of moving average forecasting is that it gives equal weight to each period. Secondly, it only considers data during the chosen period.

Exponential smoothing: This more advanced approach overcomes the problems above. Exponential smoothing looks at the actual demand of the current period and the forecast previously made for the current period. These observations are exponentially weighted to decrease over time to produce a forecast for the upcoming period.

Statistical forecasting using only historic consumption data works well if you sell or produce the same amount of each item in every period. But if sales/usage fluctuates over a couple of periods, then you’re going to suffer with inaccurate forecasts and therefore either stock-outs, or excess inventory.

Effective demand forecasting needs to consider demand trends, seasonality and your stock items’ stage of the product lifecycle. Unfortunately, it’s extremely challenging and time-consuming to consider these factors when using manual forecasting models. If you’re looking to be more sophisticated with your statistical forecasting, it may be time to consider demand forecasting software.

Qualitative forecasting

Qualitative forecasting models are based on opinions, past experience and, sometimes, best guesses.

With qualitative demand forecasting, predictions are based on expert knowledge of how the market works. These insights could come from one person or multiple people both internally and externally to the business.

There are a number of qualitative forecasting methods:

Panel approach: this can be a panel of experts or employees from across a business e.g sales and marketing executives, who get together and act like a focus group, reviewing data and making recommendations. Although the outcome is likely to be more balanced than one person’s opinion, even experts can get it wrong!

Delphi approach: this involves crafting a questionnaire and sending it out to relevant experts who complete it e.g customers and suppliers. The results are analysed and returned, anonymously, to the participants, who get to reconsider their original responses in light of other views and opinions, until a final consensus is found. This more formal approach helps reduce influences from face-to-face meetings but could still include inherent bias from the experts chosen.

Scenario planning: this can be used to deal with situations with greater uncertainty, or longer-range forecasts.  A panel of experts is asked to devise a range of future scenarios, likely outcomes and plans to ensure the most desirable one is achieved. For example, predicting the impact of a new sales promotion, estimating the effect a new technology may have on the marketplace or considering the influence of social trends on future buying habits.

Which demand forecasting model is best?

Smart inventory planners may choose to use both qualitative and quantitative demand forecasting techniques for a more well-rounded perspective.  For example, you may choose to use a statistical moving average calculation to set your base demand forecast (a quantitative form of forecasting). Then set up a panel to discuss future market trends into the coming year (using a qualitative approach).

If you find inventory forecasting a challenge, contact the EazyStock team today. Our demand forecasting software gives you advanced inventory management capabilities that you can utilise to improve the day-to-day running of your business – fast.