All forecasting is difficult!! Predicting the future correctly is an impossible task – you will always get it wrong to some degree – and so it’s impossible to forecast precisely how many shoppers want how much of your product.
The good news, however, is that you can deploy proven techniques to reduce the margin of error and move from guesswork, through estimates, to well-reasoned forecasts. Your foundation is the effective analysis of rich and detailed data.
This isn’t trivial; true forecasting requires:
- an understanding of past performance;
- explanatory variables for that performance; and,
- reasonable predictions of how those variables will change in the future.
The further out that you can predict the variables, with high certainty, the more accurate your forecast will be. Let’s consider a few of the potential variables through an example.
Acme produces a range of food products and is experiencing growing demand for these from shoppers and consumers. Acme has built a reputation for good quality, price and supply and so has won contracts to supply some of their products to major supermarket chains. The Acme commercial team wants to forecast demand for their products over the next 3, 6 and 12 months to help the business buy (ingredients), manufacture (finished product) and ship (packaged products) efficiently whilst supporting strong revenue growth.
What do they need to consider when forecasting shopper demand?
- Price
The most obvious factor is price; if Acme’s retail customers reduce the price they charge for Acme products, it’s likely that shopper demand will increase (and, conversely, demand will probably fall as price increases). Price can be changed on a permanent/ long-term basis or reduced for a temporary promotion, and in every case it’s likely to impact demand.
Analysing past price changes can help to inform a picture of the price elasticity of demand for products; this is a primary variable in forecasting future shopper demand. - Distribution
Shoppers are more likely to buy Acme products if they see them in many outlets, so product distribution is a useful predictor of demand. As distribution increases, and products are more available in more stores, demand for those products is likely to increase. However, increases in distribution rarely result in linear increases in demand; once the biggest and best stores are saturated, adding more stores is likely to have a lower incremental impact on demand.
Understanding planned distribution (the stores where the product should be present) helps to inform the likely changes in demand that occur as listings increase or decrease. - Availability
Whilst distribution tells Acme something about how many stores should carry their products, it’s possible that some stores have no stock on display. This means that availability is also a driver of demand; at the extreme, if no store has any stock then demand will appear to be zero, if measured on achieved sales alone.
When analysing sales performance – as a means of assessing shopper demand – it is important to factor in known availability issues and work to increase availability to meet, and establish a better picture of, shopper demand. - Competitor activity
When Acme’s competitors promote their products – by reducing price and/or increasing media spend – this is likely to have an impact on Acme product demand. Shoppers who normally purchase Acme product may be persuaded to try alternatives if competitor activity is effective.
Monitoring competitor products, prices and promotions helps to determine how substitutable Acme products are but this is a limited future variable unless competitor promotional plans are known in advance, which is very rare. - Range
In the same way that competitor products can cannibalise normal demand for Acme products, so new Acme product variants can reduce apparent shopper demand. For example, if Acme launches a new 500g variant of a 250g item, it’s likely that demand for the 250g variant will reduce once added to retailer product ranges. Some shoppers will prefer the larger SKU and whilst overall demand for Acme products may rise (a typical goal of additional pack options) it’s likely that demand for individual SKUs will change.
Factoring new product variants into SKU-level forecasts is important – particularly when planning promotions, which can reduce demand for similar product variants. - Weather (and other external factors)
Some product categories are very affected by external factors such as the weather. Whilst ice cream is the obvious example, it’s also likely that beer, carbonated drinks and sausages will be impacted by great weather over bank holiday weekends.
Weather forecasts can help to shape short-term demand, if demand can be modelled and measured at a local rough level.
The more of these factors – and others, specific to their products and business operations – that Acme can measure and correlate with achieved sales, the better they will be able to model demand for their products and produce meaningful, actionable demand forecasts.