View all Blog

How does WGSN use product data to predict fashion trends?

Pink shoes and cardigan on display
WGSN original image
Nov 02, 2022
Updates from WGSN

Samantha Tira, Data Analytics Manager at WGSN, shares how WGSN uses product data to analyse market signals and determine successful fashion trends.

Q: What do you do at WGSN?

My role at WGSN is to work with data related to our AI-driven predictive analytics and fashion planning platform, TrendCurve+, and my job can vary greatly from day to day. A typical day could include analysing our proprietary data to identify key trends and draw insights for Category Outlook, Trend Spotlight or bespoke TrendCurve+ reports, or working with the broader data team on projects to constantly improve how we utilise data and machine learning. 

Q: What role does product data play in the trend forecasting process?

As one of the five S’s in WGSN’s data sources, “Shelf” – or product data – plays a key role in the trend forecasting process and is one of the most useful data sets due to how multifaceted it is. We have details for over 400 million SKUs (stock keeping units) that are updated on a daily basis and include everything from colour to price history. When you have such a robust data set, you can slice and aggregate it in many different ways to reveal common threads across the broader market or smaller niche trends. It’s a powerful component of our forecast model and also the basis for market signals.

Q: What are market signals, and how can you use them to indicate future fashion trends?

Market signals are key retail metrics, such as out of stock, markdown and newness, which signal the health of a trend. The relationship between market signals and their rate of change over time can often be used as an indicator to determine when and to what extent a trend will move in or out of fashion. For example, padded outerwear in UK markets began to increase rapidly last year, which is reflected in the surging out of stock metric and increasing newness. However, markdown has been increasing as well recently, which suggests that the trend may begin to level out in the future as it nears its peak.

Rains / Flying Solo / Max Mara
Rains / Flying Solo / Max Mara

Q: How far ahead can you predict trends using product data?

Combining WGSN’s wealth of proprietary data, including product data, with the latest in machine learning allows us to predict trends up to two years out with over 90% accuracy.

Q: What’s an interesting/unexpected trend you’ve uncovered from product data?

It has been interesting to watch the diversification of denim fits through data points. Since skinny jeans had dominated the market for so many years, retailers began to test other fits in fairly even proportions. Watching the market signals change over time to gain that first insight into which fit(s) will be the new #1 has felt a bit like watching a race unfold – it has been really fun. Wide-leg jeans are a clear winner for both UK and US markets based on percentage point growth year over year. However, the second-place fit is different depending on the market. In the US, bootcut/flare is the next fastest-growing fit, while in the UK straight-leg jeans are forecast to grow faster.

Nili Lotan / Etro / Martine Rose / Tory Burch
Nili Lotan / Etro / Martine Rose / Tory Burch

Q: How can fashion brands use WGSN product data to plan ahead for their next product launch?

WGSN combines product data with industry experts to highlight the most relevant and impactful changes for fashion brands. TrendCurve+ in particular is ideal for product line management, including launches, because of quantified longer-range forecasts. Fashion brands will know how much product to assort into and when for their launch, leading to improved margins and less waste.

For more information on TrendCurve+ and our data-driven forecasting, please visit our website or request a demo here.

Share article: