Preseason planning is one of the most complicated problems in forecasting. Each year, fashion retailers face the challenge of accurately predicting future demand for the upcoming season.
What will be the basic demand for a new article that will be on the market in six months? This is a billion dollar question.
Fashion retailers need to recognize and accept that uncertainty is a fact in forecasting demand. The first step for retailers to manage this problem is to segment products using advanced prescriptive and predictive analytics such as clustering algorithms to segment products and define a supply chain strategy for each segment.
When planning items with a high forecast error, very little information is available on what the prevailing fashion will be in the future.
Forecast for basic elements such as a White T-shirt it is relatively simpler than fashion items, as forecasts can be based on the sales history of similar items.
But consider the forecast for a new fashion item as a floral printed fluo dress. That’s when things get complicated.
Fashion items have short life cycles, long lead times, and no historical data to draw on. Rapidly changing customer preferences, new competition, macro influences and “see now buy now” trends make it incredibly difficult to accurately predict long-term demand. That’s why judging how many units a fashion retailer will have to order from the supplier becomes more of a guess.
By guessing wrong, you will either run out of inventory, which is a problem for many consumers, or stock up on too much inventory that will need to be flagged later.
To our knowledge, there is no “right way” to accurately predict the demand for new fashion items. But nowadays, the data is plentiful and there are different approaches that retailers apply.
Here are 6 commonly used methods.
1. Rely on the opinion of designers, buyers and merchandisers
Despite all the developments in AI-based demand forecasting, many fashion retailers still use an instinct-based approach and trust their shoppers, merchandisers and designers to make preseason predictions.
Merchants read the market, shoppers visit production and design houses, and designers use their personal observations of what people will buy. In this method, long-term predictions are limited by intuitions. This is more of an art and method based on creativity rather than something scientific.
In addition, any designer or buyer can work on a narrow segment of the merchandise. For example, one can work on scarves, while the other can work on crop tops. Therefore, by using this method alone, fashion retailers cannot accurately predict effects such as cannibalization or product substitution.
2. Find similar items in the past and project from there
Fashion retailers may have similar products close enough to compare. Think of a retailer who wants to predict demand for a “never out of stock” as a black dress for next season.
Typically, the retailer has access to historical data of existing or previously sold black dresses in recent years. Looking at data from previous years can help predict demand at levels sufficient for existing black clothing. But they can’t be 100% efficient at predicting demand for a new item. Due to the rapidly changing nature of the fashion industry, it is completely impossible to meet tomorrow’s consumer demand if predictions are based solely on yesterday’s data for similar products.
3. Work with a trend forecasting agency
Unlike other retail sectors, fashion is strongly trend-oriented. Fashion retailers can partner with data-driven trend forecasting companies that offer predictive analytics on upcoming trends and products.