Retailers must widen the data sets they look at to develop collections that perfectly suit their customers’ needs, says data scientist Clive Humby
Last month brought the news that the fashion industry is in decline for the first time in six years. Brand loyalty among fast fashion consumers is fleeting. Retailers are over-buying stock that simply does not appeal to their consumers and long supply chains mean that wrong decisions on stock are harder to rectify.
Retailers are at risk of losing touch with what their customers want, so the need to listen to and understand consumers has never been greater. But how do you generate sales from an increasingly hard-to-reach consumer base, one that is information-rich but attention-poor?
Victoria Beckham was identified as the most prominent influencer across all segments of H&M’s target consumers
Producing a generic collection based on the latest trends is one option, but research shows that consumers prefer genuinely personal offerings. Alternatively, fashion brands can use the data they have on their customers’ shopping habits, using the past as a predictive guide. There are, however, problems with this approach. Customer data alone is not a strong enough guide, given that clothing retailers see many of their customers so infrequently.
To gain a complete picture of consumers, brands need to blend their own customer data with other data sources, including social media. This will not only enable them to tailor their marketing to better suit their customers, but will also help attract new customers.
Social data is one of the most powerful sets of data now available to brands. Through analysis of more than 50 billion interactions between users on social media, we have discovered a lot about what consumers really care about.
In a recent study, we took five of the highest profile fashion brands – H&M, Asos, Topshop, River Island and New Look – and analysed their customers’ passions. What our data revealed was that, in the effort to generate sales, collaborations with celebrities are often the simplest and most effective way of engaging consumers.
But fashion retailers must be wary of collaboration for the sake of it – the best collaborations are tailored specifically to the interests of a brand’s customer base. In H&M’s case, our analysis revealed that the brand was underperforming at the higher end of the market, which is dominated by 18-to-24-year-olds looking for investment-piece purchases. Victoria Beckham was identified as the most prominent influencer across all segments of H&M’s target consumers, meaning a collaboration with her could pay off handsomely.
The lesson here is clear: by using social data you can be more effective as a fashion brand
New Look and River Island were also brands that attracted consumers looking for investment pieces. But in an example of how careful brands need to be when considering collaborations, we found that, rather than working with top-end fashion designers, the likes of celebrity Alexa Chung and Made In Chelsea’s Millie Mackintosh would drive more sales for these retailers.
Cast your net wider
Fashion brands should not assume that the faces on today’s magazine covers are the best to engage with. Topshop and Asos are brands that have a specific appeal within the young, “preppy teens” customer segment. Within this group, we found that the biggest influencers were the likes of former X Factor contestants Frankie Cocozza and (winner) James Arthur.
To the average onlooker, these stars may seem long out of date, having emerged from The X Factor in 2011 and 2012 respectively. However, both of these stars broadcast to a large online audience that matches the Topshop and Asos target customers. Cocozza has more than 1 million Twitter followers and Arthur more than 2 million, underlining the importance of brands basing their marketing strategies on hard data rather than presumed popularity.
The lesson here is clear: by using social data you can be more effective as a fashion brand, targeting the specific consumer types that really matter to you.
Consumers should never be perceived as homogeneous groups, but – instead – segmented, examined and engaged based on their shared passions, likes, and interests. Nor should influencers be assessed on intangible measures of fame and reputation. By looking at different data sets, and taking the time to better understand their customers, the question of “how do we generate sales?” will become less daunting for fashion brands.
Clive Humby is chief data scientist at consumer insight company Starcount