Using Artificial Intelligence to Analyze All Fashion Customer Data

The apparel market is one of the largest in existence, accounting for 2 percent of the world’s GDP, and valued at roughly $3 trillion. Every year, American households spend close to $2,000 on apparel alone, and over 211 million of these shoppers make their purchases digitally. In an industry this large, and this competitive, it’s hard to keep customers engaged and active with your lines.

Globalization of Fashion

With the increase of digitalization, retail markets are growing at an unprecedented pace. 15 of the 20 cities with the largest growing apparel sales lie outside of the Western marketplace. Newer markets are already beginning to dominate sales. Markets in Asia and South America, for example, already account for one third of global revenue in female apparel, and this number is only expected to grow.

With this trend towards globalization comes the question: “Can a global brand really live up to the needs of diverse cultures from Latin America to Eastern Europe all the way to Asia without putting its identity on the line?” (McKinsey) The differences in expectations and desires from country to country are massive. It is a big challenge for global retailers to properly segment and market to their increasingly diverse international buyers.

Be a Stylist at Scale

To become a staple in someone’s wardrobe, you have to understand them and their style—now and in the future. Consumers want the brands they purchase from to provide them with promotions and messages that are tailored to their specific look. 86% of customers say that personalization has an impact on what they buy/who they buy from and 1/3 of customers feel that there isn’t enough personalization in their current shopping experience. Converting transactional sale shoppers into repeat buyers who wear multiple items from different lines is a big challenge.

In order to achieve this, you need to be flexible and dynamic in segmenting and marketing to your wearers. Trends are constantly shifting, media consumption is ever changing, channels of engagement are different from one month to the next. Today, roughly 35 percent of consumers rely on recommendations from social networks, a metric that would have been non-existent 10 years ago. Being able to get ahead of someone’s style and understand their tastes, and what SKUs suit their look, can be the difference between someone buying “those shirts on sale that one time” and a repeat customer who bases their wardrobe on your seasonal pieces over the next three years.

Big Data to Big Style

The amount of customer data currently available is huge. Browsing history, purchase data, social mentions, and spending trends are all sitting in data lakes and warehouses at your company waiting to be unlocked. When used correctly, this data can show what customers are buying, when they’re buying it, and what they’re likely to buy next. In this way big data has the power to revolutionize how the apparel industry interacts with its customers—by creating forward thinking styles, offers, and recommendations at the individual customer level.

How AI Can Help

Although most big fashion brands already have tools and data scientists that help them make sense of their data, there’s an opportunity to implement an AI + Machine Learning based solution that can double or even triple the speed and accuracy with which they act on their shoppers data and preferences.

A popular way to characterize data usage in fashion marketing is the three D’s paradigm: Data, Decision, and Delivery. When collecting data, quality and completeness is paramount in order to give effective and reliable reports & predictions. Once a reliable metric has been established, analysis and decision making needs to take place. Trends in purchase history, for example, can be used to suggest relevant promotions or styles for specific customers that can be used to convert them into a frequent and higher value shopper.

The final D is delivery to the identified segments and cohorts through outreach and promotion. This stage requires a genuine approach, which is why human to human marketing outreach is still the most effective strategy. Companies need to move through this cycle of data collection, decision making, and delivery quickly and accurately in order to resonate with their ever-changing customers.

To achieve this, there are two key things a business must do—and do well:

• De-silo, and unify your data. In this way you can create a comprehensive and integrated view of your shoppers across all touch points, even enriching this data with third party information that goes beyond what they provide you themselves
• Leverage that clean and complete data to conduct deep analysis and draw correlations between thousands of data points. This allows you to analyze and predict the behaviors that are most important to your business. Things like trend affinity, propensity to certain products (cross-sell potential), regional looks, promotions, and more

Here’s where an AI engine can help. Implementing Artificial Intelligence/Machine Learning enables the creation of systems that are both smart and adaptive enough to solve problems faster and better than a human ever could. This requires a Dynamic Data Engine (DDE) and an Embedded Analytics Engine (EAE).

To have a comprehensive view of your customers, you need to unify data across all of your sources: online, mobile, in store, etc. A DDE can identify, cleanse, unify, and enrich this data in real time, and for each individual customer. Once this has been done, an EAE uses the curated data to predict customer behavior with minimal human intervention. It can tell you, for example, that people who purchased a specific color sweater online the day it came out are less likely to respond to email discounts, but more likely to respond to your new “East Coast streetwear” look-book, and you can tailor your marketing accordingly. It can also show you differences across demographics, and can suggest higher level ways to position your brand across groups and styles.

AI has the power and the speed to take care of the first two D’s (Data and Decisioning) quickly and well, with little human oversight after the initial training. The power of these systems lies in their flexibility. With the data you already possess, they can find the signals your customers are sending, and suggest the right time and way to create meaningful engagement. This frees up your time and energy to focus on “Delivery”.

Conclusion

The fashion marketplace is diverse, and within each individual market you have a variety of consumers with specific wants and needs. If your business can act faster, with more accuracy, and in line with customer needs, you will have a technological competitive advantage in the apparel industry.

About the Author

Katie DeMatte is is a writer and Digital Marketing Associate for Zylotech. ZyloTech was launched in 2014 out of MIT and provides an award-winning AI powered customer data & insights platform for omni-channel marketing operations. Headquartered in Cambridge, MA, (with offices in Sunnyvale, CA and Bangalore, India) ZyloTech is comprised of Industry experts including Ph.D.’s in AI, data scientists, and other customer marketing experts who work alongside a high-profile board and advisory team. Enterprise clients across Retail, High Tech and Financial Industry.