AR Continues Struggle in Retail

An area we spent lots of early time in.  Good overview article of current work.

Despite advancements, AR struggles to take off in retail   By Anna Hensel in Modern Retail
This is the third part in a series by Modern Retail about the technologies that were going to change retail — and where they are now. See our previous stories, about the future of robotics, and the truth about RFID.

Three years ago, retailers began spending money on AR.

In 2017, Williams-Sonoma acquired Outward, an AR and 3D imaging startup for $112 million. The following year, Ulta Beauty also acquired an AR startup for an undisclosed amount, as did L’Oreal. Retailers like Walmart and Nike acquired other startups that specialized in virtual reality and computer vision, but also with the goal of using them to create experiences that incorporated some type of virtual reality component.

That ushered in a stampede of retailers eager to experiment with AR. Michael Kors and Warby Parker were some of the first companies to experiment with running AR in Facebook ads. Macy’s launched AR experiences for both furniture and beauty shoppers.

AR right now is big among two categories: beauty or furniture. Sephora, Ulta Beauty, Wayfair, Williams-Sonoma and Ikea have all launched virtual try-on experiences that use AR. Retailers in apparel have also experimented with adding these types of tools — the most recent being Asos — many of those efforts remain firmly in the experimentation phase. Because of the nature of the technology, it’s really those three industries where it can make a real difference — especially in fashion.

But even for the retailers who have invested heavily in AR by building their own virtual try on experiences, they haven’t been able to point definitively to how much their AR tools have resulted in a sales increase.

“There’s still some hyping [around AR], no doubt,” said Andrew Lispman, e-commerce analyst at eMarketer. “But I think the reality has set in when you see the initial use cases…are mostly edge cases. They may work well for their category, but it’s not a use case that extends between all or even most categories.”   ...."