How Brands Can Start Making Use of Data

The amount of data available to retailers today can feel like too much of a good thing.

Data is invaluable for making smarter decisions on inventory planning, pricing strategies, marketing, logistics and more. But the torrent of information can also leave companies feeling paralysed trying to work out what to do with it all. Every customer visit to a store or website, every click or interaction, every purchase and return leaves a trail. Social media, the supply chain and other sources have their own data as well.

A survey of 10,000 business leaders by Salesforce recently found 30 percent were overwhelmed by the volume of data and 33 percent lacked the ability to derive insights from it. A sizable share of brand executives and consultants also reported being dissatisfied with their data capabilities in a survey last year by the Luxury Institute, a research and technology partner for luxury brands.

“I won’t tell you who the CEO was, but it was a large multi-brand retailer, famous in the world, and he said, ‘Milton, our team doesn’t even know what to do with the data we have now,’” said Milton Pedraza, the Luxury Institute’s chief executive.

The situation can be particularly daunting for brands just beginning to use data. The good news is experts say there are some basic steps they can take to get started.

Start From the Outcome You Want, Not the Data You Have

A common mistake businesses make is to look at all their data and try to puzzle out what insights they contain. That’s actually backwards, according to Robin Barrett Wilson, industry executive advisor for fashion at SAP, a software and technology vendor. It’s more effective and efficient to take a “reverse engineering” approach: determine the outcome you’re trying to achieve and then figure out what data you need and if you have it.

“Where people run into trouble is [saying], ‘I have these data. What question can I answer?’” said Nicole DeHoratius, an adjunct professor of operations management at the University of Chicago’s Booth School of Business. “It’s what questions need to be answered, and then let’s go get the data that allow us to answer those questions.”

To figure out which outcomes should take priority, Ekimetrics, a data-science partner for businesses, has developed a framework for companies that starts with gathering a cross-functional team to understand the results the company is after and potential use cases for data.

“There is a process of mapping all of those use cases against what’s going to deliver the most value versus what’s the most effort and investment to bring it to fruition,” said Sona Abaryan, a partner and retail and consumer lead at Ekimetrics. Something like demand forecasting that affects several areas of the company offers a lot of value but is complex to develop, whereas a company might choose to focus on quick wins first.

Focus on Data Quality, Not Quantity

Bad or incomplete data can produce inaccurate insights that point you in the wrong direction. It’s common to see problems like incorrect inventory records or mislabeled products that come up as the wrong item when they’re scanned out of a store. Companies often have a bias for focussing on what’s easily measurable, too, Abaryan noted, which isn’t necessarily what’s most valuable. That can vary by business model. Wholesale businesses may have different questions to answer than DTC brands.

One data priority for brands should be having a clear picture of sales, stock and sell-through at a granular level, such as how much of a particular SKU sold in a store the previous day, according to Carlos Sánchez Altable, a partner at McKinsey.

“This is what really drives the fundamentals of a fashion brand,” he said. “You don’t need real-time, but you need yesterday.”

Brands need to understand their customers, too, which isn’t always easy. Often they have heaps of third-party data, but it’s the wrong data, the Luxury Institute’s Pedraza said. Knowing a shopper’s birthday isn’t as important as their size and style preferences. He suggested brands communicate clearly and honestly with customers to gather data directly and reward them when they provide it.

This first-party data is becoming more important as governments impose online privacy measures and tech companies move to disable tracking cookies, making it more difficult to accurately target shoppers online. In a research note last December looking at major themes for 2023, investment bank Cowen emphasised that companies need “to earn customer trust and build the infrastructure to collect this information,” suggesting they embrace tactics like loyalty programmes.

Break Data Out of Siloes

If the marketing department has data that’s valuable to the merchandising team but doesn’t make it available, it’s not much use. Having a centralized data source in the cloud, rather than storing information in individual spreadsheets saved locally on employee computers, can help.

It’s also imperative that data from disparate sources be organised and labelled similarly. A spreadsheet full of SKU numbers and another with columns of styles, sizes and colours might have complementary information but can’t be easily brought together.

“I am amazed that even today I ended up working with a retailer where we couldn’t merge their warehouse data with their procurement data because they didn’t have a common field,” DeHoratius said.

Companies that manage this challenge effectively often have a directory everyone can refer to, according to Barrett Wilson, and a person or team that understands technology and retail.

“Really it’s about having that bridge between IT and business,” she said.

Of course, the key to making this information sharing effective is making sure teams across the company are communicating and thinking holistically and not just about their specific roles.

Find Talent — Or Train It

Fashion can struggle to compete with industries like tech to attract talent with the skills to process and apply data. Sánchez Altable said one issue companies run into is getting hung up on specific roles like “data architect” when such a defined skillset isn’t necessary.

“You just need smart software engineers,” he said.

There are other places retailers can find talent, too, depending on whether they need to hire someone with a background in technology.

“MBA students right now come out with incredibly good skills for data and analytics,” DeHoratius said.

Hiring isn’t the only option, however. Companies can also train existing staff. Not every company may be able to do what Levi’s did and create an internal bootcamp on artificial intelligence, but there’s a lot of excellent training on data analytics available, DeHoratius said.

Make Clear Data Is There to Assist, Not Replace

One unexpected obstacle brands may face integrating data is cultural. DeHoratius has found employees may think data is important but not fully know why. Others may see it as a threat.

“The data doesn’t know more than I do,” DeHoratius said, summing up the response. “It isn’t that it does … You need individuals that understand the business well to ask good questions of the data, but you need the data to conduct the analysis in a rigorous way.”

Companies may have to make clear that the point of using data isn’t to replace human insight but to supplement it. SAP’s Barrett Wilson even cautioned against being too reliant on data without scrutinising it. Sometimes there’s information that isn’t captured in the numbers, like weather events that affect sales.

“It is always an art and a science,” she said.