How machine-learning is helping Footasylum gain better tread

Sportswear brand Footasylum has struggled to keep apace in a turbulent market. Last autumn, the company reported a pre-tax loss of £4 million for its first six months, and issued a profit warning for 2018/19 following a disappointing Christmas. In April shareholders agreed to sell the brand and its 69 stores to rival JD Sports for £90 million.

But the retailer believes there are reasons to be optimistic. Online sales have continued to perform strongly, growing 28% to £36 million for the 18 weeks ended 29 December 2018, with the brand having changed direction to focus on web and data.

“We knew we needed to be doing something with data,” says Tom Summerfield, head of commerce at Footasylum. “And we were ready to be educated about what direction we wanted to take.” 

A key part of that strategy included a partnership with AI specialist Peak at the beginning of last year to improve its social marketing via machine-learning algorithms.

That ultimately led to the creation of a “predicted customer view” of customers likely to be interested in certain products, with segment profiles then used to target ‘lookalike audiences’ via social media in order to acquire new customers.

For every £1 spent on platforms like Facebook and Instagram it gets £84 back in revenue, an “unusually high” return on advertising spend says Summerfield.

However, the bulk of the project so far has involved using information from its customer database to determine how the company should contact existing customers. 

“Using a host of transactional behaviour data” the company can “predict whether a customer has a high propensity to buy something, or if they are not in the market, we leave them alone," he says. "It’s about popping up at the right time with the right product in a really natural way.”

Step change

The result has been a 28% increase in email revenue. “We have people on our email database who don’t shop on the website, so it's also providing a better journey and experience for those customers,” he says. “It is a core component of the web growth, but it's also helping our bricks and mortar business as well.” 

Seasonality data is also fed into the system. “As the weather is hopefully turning now, we might ask the algorithm to consider that and use data from a weather website to help build out the recommendations further.”

He insists the project is more than just a smart use of analytics. “The AI allows us to make decisions based on an original set of trading strategies we put in at the start, but then does the rest for us. It’s creating efficiency and effectiveness…through automation of the action.”

Summerfield hopes to use the system for greater personalisation, customer acquisition and to monitor and predict consumer demand for inventory management.

But with the JD Sports acquisition on the cards for this year, how much can Footaslyum really plan for the future? Not surprisingly, he can't say much while talks are ongoing. But he believes the two businesses "complement each other" and "nothing has changed" in terms of its digital and data strategies.