In a supply chain, we understand the need to carefully balance supply and demand. Knowing and predicting your potential sales is critical to your success. Successfully launching a new product is central to this. However, how do you – with reasonable accuracy – predict demand for a new product? How can you be sure that your next product is going to be a success that improves profits, and builds your brand?
Predicting Demand for a New Product – The Challenges
Forecasting demand for a new product, or product line, is always going to be one of the hardest parts of supply chain management. At the end of the day, the consumer is a fickle character whose behaviour it isn’t always possible to predict. You only need to look at a child’s latest playground craze to understand this. What makes a diablo the product of 2009 and loom bands the craze of 2014? How in situations such as these can you ensure you don’t fall victim of costly under or over-stocking?
The problems that come, if you get your predictions wrong, can come in many forms. You might be sat with a full and costly warehouse, and then obsolete stock. On the other hand, you might find yourself with hefty bills for expedited delivery, relying on air travel instead of ship, or a competitor who can step in and meet demand instead.
Insight into the likely demand of a new product is therefore invaluable.
Predicting Demand for a New Product – The Solutions
In a world that seems increasingly to sever ties with its roots, it can be easy to think that the only tools you have in your arsenal – when it comes to predicting demand – is a savvy market research department with their finger on the pulse. However, in a study partnered with Dell, it appears that historical data is far more important than we’ve been giving it credit for. The study shows that Dell, using this historical data, could improve forecasts for new products by up to 9%. Figures like this can translate to millions, if not billions, of pounds. This means a new partnership needs to be fostered between traditional market researchers, and data analysts drawing on the past. These can be brought together under the umbrella of executives drawing on their own experience. The reason this is important is that most products are not actually truly new. They are usually upgrades, adaptations, a re-model. Only rarely does something completely new come in to the market place. This makes historical data important.
The Dell experience shows us that it is possible to create product ‘clusters’ and compare these to demand ‘curves’ from the past. You can then “find the product lifestyle curve that fits it best and use this curve to forecast demand for the new product” (see Researchgate).
The Dell Experience
The researchers took the data from 133 Dell computer products and began to establish which lifestyle curves were the ‘best fit’. Some were more typical, others more convoluted. However, fundamentally they consist of an introduction, growth in popularity, a stable period, and then a subsequent decline. In reality, the term ‘curve’ is a misnomer when it comes to product demand. It more accurately resembles a triangle, with a peak. For the vast majority of products there is never really a stable period. Instead, demand rises, peaks, and then drops off.
This is intensely valuable information for predicting the demand for a new product. It means that you are working with three important variables: how long do you think interest will last in the product, when do you think the peaks will happen, and how high will that peak be?
Going back to the original clusters, the researchers were then able to identify optimal lifecycle curves for each cluster. This forms the basis of predictive data. This can then be used, along with all of the other predictive information the company has about the new product (from market research, for example) to draw a more accurate prediction about the product lifecycle and anticipated demand. You are therefore forewarned and forearmed going into launch; meeting demand; and knowing when the end of the product life will come. The element of surprise is dampened down.
Applying the Lessons of Dell to Other New Products
The researchers suppose that this method of demand forecasting could be suitable for a range of different products from electronics, technology, clothes, and accessories: The products which tend to become obsolete quite quickly.
New Product Demand Forecasting
Whilst using this historical data information could prove critical for predicting the demand for a new product, it can’t be used in isolation and must be used in conjunction with other demand forecasting techniques. It would be a dangerous mistake to think we can completely predict the future based on the experiences of the past – things do change which will ultimately affect the end of a product, or a market’s saturation point.
Using a range of techniques, it becomes more possible to predict the demand for a new product and therefore accurately make investments, accurately pitch marketing, and accurately prime logistics and resource strategies. This can help to put the business, rather than whimsical market forces, in control.
Strategies to embrace alongside historical analysis include more traditional market research techniques such as opinion poll forecasting and utilising a sample market. These techniques are more in tune with the intricacies of the current marketplace. Opinion polls will make direct enquiries with the consumers in order to probe their intentions and use this to feedback in to product design, and prediction of demand. Providing samples of product to a sample market does a similar thing – it gauges opinion on the ground, in current circumstances. These aren’t perfect sciences, more of an art form, that all come together to predict demand.