Analytics and Fashion: 3 Ways Predictive Analytics Is Improving the Apparel Retail Industry

Spec Work: White Paper

Nowhere has the phenomenon of instant gratification been more evident than in the retail industry. Conveniences of the digital age, like Amazon’s two- and one-day deliveries, have conditioned consumers to expect an almost instant fulfillment of their orders. Fashion industry in particular has had its own challenges to overcome keeping up with the demands of their customers.

Consumers in fashion have always desired novelty and variety. However, while in the past the brands had months to plan for a collection and manufacture it, today’s companies are keeping up with the runway trends that are traveling at breakneck speed through the Internet and social media. The brands that lag behind customers’ expectations get left behind among the never-ending streams of new products.

This demand for fast and affordable runway styles has left the fashion brands to walk the fine line between quickly satisfying customer demand and doing so at the right quantity and the right price.

The most successful players in fashion have demonstrated their ability to adopt to the latest trends, quicken the pace of manufacturing and reduce their lead times. Brands like Topshop and Zara have built their businesses on agility and speed.

Planning a collection with speed and agility can be difficult employing traditional time-series forecast as this method of forecasting can be quite rigid.  Time-series forecasting relies completely on historical data and provides answers to clearly defined business questions like what to order.

As such, these forecasts are done in isolation, with an assumption that past patterns will continue in the future. Assuming that trends will remain stable in a dynamic retail market often affected by seasonality, promotions and new products could result in forecasting errors. These errors can lead to adverse business effects like excess inventory, stock-outs or excessive discounts.

On the other hand, predictive analytics uses statistical modeling and machine learning to sift through current and historical data to detect patterns and find opportunities and risks that might not have been previously considered. As such, predictive analytics transforms data into insights.

Whereas traditional time-series forecast focuses on addressing internal assumptions, predictive analytics adopts an approach that helps discover economic drivers and understand customers, business and market. Equipped with this knowledge, the brands are then better able to make smart business decisions.

Here are some of the ways predictive analytics is improving the fashion industry:

Driving Sustainability with Better Design and Procurement Decisions

Better purchasing and design practices bring about much more than just cost reductions and increased ROI. They can become a source of sustainability and social responsibility by optimizing purchasing and design processes and eliminating or minimizing last-minute decisions. Last-minute design and procurement decisions travel down the supply chain and can adversely affect factories working on the contract, resulting in overtime, worker burn-out and sometimes even compliance violations as suppliers scramble to fulfill the orders. These rushed decisions also affect sustainability: just in 2017 an estimated 60 billion meters of fabric ended up dumped into landfills, wasted during the production phase. Predictive analytics helps brands gain speed and clarity by taking the guesswork out of how best to reach consumers and enabling them to get to know their customer better, thus minimizing these last-minute decisions.

Designer’s creative process is enhanced by analytics software that forecasts color trends and future best-sellers while purchasing optimization provides to procurement team data-driven insights on how many units to manufacture and when to order them as well. Furthermore, predictive analytics software also helps procurement team perform spend analysis, assess suppliers, do compliance checks and develop reports. It lets brands collect and analyze internal as well as external data to benchmark their performance against the competitors in real time. Suppliers are assessed based on, not the lowest price or an existing relationship, but on their past performance data and market pricing and risk evaluation.

Predictive analytics allows procurement team to juggle multiple variables to arrive at the optimal cost-saving solution. For example, MOQs (minimal order quantities) can often result in extra inventory and tied-up capital, as extra unnecessary items are ordered just to fulfill the contract. The ordering process becomes even more challenging when there are multiple MOQs at different levels like minimum orders of fabric or minimum units of product per purchase order. Predictive analytics enables procurement to quickly find the most profitable combinations of all these MOQs while keeping inventory under control.

Rue 21 is one of the brands that has found success using customer-driven predictive analytics. They have partnered with First Insight to make buying decisions. New products are selected by collecting customer data though social platforms, email and third-party panels and then analyzing it through predictive analytic models to determine the offerings that would appeal to potential consumers most.

“Rue21 is re-energizing the brand, and that starts with having the right product at the right price,” said Michael Appel, CEO of rue21. “First Insight is helping us ensure we have differentiated products that our customers will value. We are already seeing results through eliminating under-performing products early in the selection process, while re-investing our inventory dollars into higher performers.”

Cutting Production Lead Times

In the digital era, trends in the fashion industry are more often established, not by editors and retailers, but by consumers. Previously, production was based on forecasts made by designers and buyers and products were marketed in “seasons” with the entire design and production process stretching up to a year.

With the shift towards the “see-now, buy-now” model, inventory is replenished once depleted. The brands that succeed in reducing lead times and streamlining operations are able to capitalize on the fast-changing preferences of consumers.

Predictive analytics gives companies the opportunity to postpone final design decisions and enable just-in-time manufacturing to rapidly respond to trends and demand. Product assortment is optimized by composing the range of colors, shapes, and sizes with the initial designs to fit closely to the wishes of the customers. This leads to higher turnover and fewer markdowns down the road.

Furthermore, analytics assists brands with identifying high-demand products and slow-movers. Every week, sales forecasts are generated based on open orders, in-transit orders, and inventory positions for each SKU at the store and the distribution center, and replenishment plan is generated for managers to review. Best-sellers trigger early replenishments and slow movers are removed to make space for better performing items.

These slow-moving items might be moved to stores with higher demand for them. The system proactively analyzes aspects like combination of colors and sizes, mileage between stores, pricing, and shipping costs to recommend the optimal inter-store transfers.

Spanish retailer Zara is known for agile supply chain system and ability to leverage big data to streamline their operations. They collect data from sales, PDA devices, RFID tags and social media to discover customer’s fashion preferences and manage inventory. This data is then analyzed and findings are shared with the design team, who then creates new releases that incorporate these trends. Designs are offered in small batches that satisfy the demand at that location. Zara generates weekly predictions for each item in the collection and delivers shipments to their stores twice a week.

By leveraging big data and analytics, Zara was able to cut their lead times to three weeks and avoid large advertising costs.

Boosting Revenue with Smart Pricing Decisions

Fast fashion has also had a substantial impact on a perception of value from consumer’s point of view and subsequent intensified use of promotions in the apparel industry. However, a study done by First Insight found that consumers are starting to prefer quality over price when making purchasing decisions. To optimize pricing, retailers need to understand where consumers perceive value, have the ability to respond to competition, and recognize the impact of promotions on the financial performance.

Analytics gives brands an opportunity to understand values of the key customer segments. Instead of competing on price in every category, retailers can use analytics to align the price with what the customer is willing to pay by grouping items that respond similarly to price fluctuations. For example, the sales of novelty items are typically less elastic than those of the essentials. Elasticity modeling is used to capture the nuances of apparel and also to isolate the incremental impact of each promotion.

Along with markdown optimization, predictive analytics models also address complexities of omnichannel pricing and fulfillment by accounting for factors like logistics costs.

Integrating analytics when making pricing decisions has shown to increase sales by 3 to 6 percent.

Embrace Predictive Analytics to Drive Revenue

Fashion industry has changed considerably over the past 30 years. As evolving customer expectations push retailers towards flexibility in design and speed to market, remaining relevant and competitive becomes harder to achieve.

Traditionally used processes like time-series forecasting fail to keep up with the increasingly dynamic fashion landscape and complexities of omnichannel retail.

Predictive analytics software enables brands to integrate internal sales data with behavioral profiling to achieve demand-driven supply chain.

Sources

1.      “How See-Now Buy-Now is Rewiring Retail,”, Forbes, https://www.forbes.com/sites/gregpetro/2018/01/31/how-see-now-buy-now-is-rewiring-retail/

2.      “Demand Forecasting at Zara: A look at Seasonality, Product Lifecycle and Cannibalization,” MIT, https://dspace.mit.edu/bitstream/handle/1721.1/90163/890199174-MIT.pdf?sequence=2&isAllowed=y

3.      “Demand Forecasting in the Fashion Industry: A Review,” International Journal of Engineering Business Management, https://www.researchgate.net/publication/268386280_Demand_Forecasting_in_the_Fashion_Industry_A_Review

4.      “Application of Predictive Analytics to Sales Forecasting in Fashion Business,” Research Gate, https://www.researchgate.net/publication/325100494_Application_of_predictive_analytics_to_sales_forecasting_in_fashion_business

5.      “Geek meets chic: Four actions to jump-start advanced analytics in apparel,” McKinsey, https://www.mckinsey.com/industries/retail/our-insights/geek-meets-chic-four-actions-to-jump-start-advanced-analytics-in-apparel

6.      “How Rue21 Uses Predictive Analytics to Make Buying Decisions,” RIS, https://risnews.com/how-rue21-uses-predictive-analytics-make-buying-decisions