"Data is imperfect", says panel of retailers

Retailers need to understand that data analytics cannot always provide a perfect view of customers, said a panel of retail data analysts. 

"Data analysts need to understand that data is imperfect and needs a lot of interpretation to understand what we're seeing," admitted Nicola Travis, head of marketing and CRM at The Fragrance Shop.

Travis was speaking alongside Rhian James, customer insight manager of The Dune Group, at the Big Data & Analytics Retail conference in London this week, who agreed. 

James said: "We would love to know where all our customers are based so we can generate location specific store openings and promotions. But we haven’t captured the country, and even if we had they might have moved."

Instead, Dune uses Movable Ink email client which allows marketing content to change depending on factors such as the time the email is opened or the weather. This product allows Dune to understand where customers are geographically when they open emails, so the following week they can be marketed to based on this location information.

"It might mean someone was in New York last week on work, and they will receive an email based on that information when they have come home, but it's fairly accurate for the majority of people," she explained. "So we're not worrying about not having hard data.

James said Dune is being flexible with its customer insights. 

Also sitting on the panel was Ian Jarvis, VP of business development, marketing and strategy services at The SmartCube. He said many retailers are determined to build up profiles of their customers.

"But there needs to be a balance building [profiles], it's not going to be perfect, there will be gaps and data you just can’t get your hands on."

James also explained that when Dune receives purchase history data tied to customer information, it is not always helpful to fashion retailers. 

"We don’t use their past purchases to market what they might want next," she said. "The real challenge as a fashion retailer is that seasons come round, you’re selling something they didn’t know they wanted. For instance, pool sliders weren’t in fashion a few years ago. Someone’s last five pairs of shoes, do they really give any indication of what they buy next?”

James – who is the only employee working on insight at Dune – also said the retailer’s new cloud-based single customer view solution is so effective and practical she will not need to hire a data scientist to crunch the data. 

“We’re not looking for a data scientist because my solution is drag-and-drop and user friendly,” she said. “I won’t need to pay the price for a data scientists, but someone fresh out of university who is curious with problem solving skills, not an SQL-coding person.”


What’s Hot on Essential Retail?