Predictive analytics vs Prada

Can technology predict next season's top fashions? All of the high tech techniques described in this week's story on the art and science of predicting fashion can help.

For example, predictive analytics can look at all of the characteristics that make up a style -- from fit to color -- and, based on historical trending on those attributes, predict the success of a new fashion. Social media can also help by allowing for real-time interactions with groups of people who have been identified as fashion predictors.

But there's one thing you can't forecast, argues Leslie Ghize, senior vice president with fashion consultancy The Doneger Group.

"Technology never captures free will. People buy things for reasons that never really can be quantified," she argues. Sometimes, she says, there's nothing to predict. "A lot of people don't know what they want until you show it to them." (The high-tech world analog here would be Steve Jobs and Apple creating the iPod/iTunes store, the iPhone and the iPad tablet, each of which redefined its market.)

While analytics can help with things like quantities and distribution strategies once a product has a track record, it can't make the final call on new fashions that are completely different from what has sold before. Analytics can influence, but the decisions still come down to experience and intuition. "There has to be a balance," Ghize says.

Right now the "balance" meter is pointed all the way over to human intuition and experience.

In some quarters it belongs there. A few influential designers, such as Miuccia Prada, carry enough weight to establish an innovative new fashion as a winner. "She sends stuff down the runway that is very advanced. It takes peoples' eyes a long time to get used to the pattern or color. But eventually it's so editorialized, it's picked up in so many magazines, that it becomes what people want. It has nothing to do with technology or predicting anything. She is inspired to develop what she wants and she strongly influences fashion," Ghize says.

To some extent, fashion is a self-fulfilling prophesy that's built on relationships, advice and experience within the industry, Ghize says. For example, bright colors were the big story of the spring season. "Everyone buys into the philosophy, everyone in the industry buys that, and that's all the customer has to buy."

But that kind of insular group thinking seems like a bit of an anachronism at a time when fashion-conscious consumers take their cues from the Internet and their peers through social networks. They have a world view that's much broader than the marching orders that traditionally have come through fashion runways and retailers. There is an opportunity cost to following the crowd inside the fashion beltway.

Ghize sees a disconnect between the way in which people create an offering and how customers perceive it. "They're not interested in the trend," she says. "Customers want to be excited and find something great and buy it."

So will fashion retailers have that it item when consumers come looking?

A more rigorous application of analytics may help reinforce a decision to take a leap and go with something that runs counter to the industry's conventional wisdom. And the application of analytics to accurately predict the demand and optimized distribution mix for more established fashion items may free up designers to take those risks.

Computerworld's IT Salary Survey 2017 results
Shop Tech Products at Amazon