Three Ways Image Recognition Technology Will Change Retail Marketing In 2017

www.lumsdainephotography.com

In this opinion piece, Jon Stubley, VP ANZ of GumGum, analyses the development of image recognition and the powerful impact it’s having on the retail industry.

Carlotta Vittori
Posted by Carlotta Vittori

Visual merchandising has always been critical for the retail industry. From the dawn of time society has been engaged in some form of barter or trade which has been driven first and foremost by aesthetics.

From putting the ripest peach on top of the basket to displaying the shiniest trinkets on the roadside stall, a picture has always told a thousand words.

It was in the nineteenth century however that visual merchandising really came into its own when traditional dry goods stores morphed into the large scale department stores of today. Store windows became a pivotal marketing tool that remains a constant today (like many of you I’m soon to take my kids to see the David Jones and Myer windows) and then crept instore to make indoor display part of the interior design.

In the twentieth century the store windows really did become works of art with artists like Dali and Warhol creating some of the elaborate displays.

Fast forward to our twenty-first century world of ecommerce and omni-channel retail combined with advances in image recognition technology and we are now seeing new forms of retail merchandising.

Indeed image recognition has now reached a level of sophistication that means it is beginning to have a serious impact on the retail industry. I can say this with authority as I work for a company that invented in-image advertising back in 2007 and it is a much more precise beast in 2016.

Our own solution can reliably identify both the granular detail and, most importantly, context of an image on a web page and then serve an ad over it.  That ad can be as specific as identifying an image of a curly haired brunette and then serving an ad for curly haired brunettes over it, for example.

This is just the tip of the iceberg for retailers though as we are now seeing image recognition being used to drive purchases.

Last month it was reported that eBay is introducing visual search technology to drive its new shopping venture, eBay Collective – allowing homeware purchasers to browse an online ‘look book’ with technology then matching to similar looking items for immediate purchase.

Consider also that the visual web is growing exponentially with two billion images already uploaded and shared daily across the world, and the power of images has been documented by MIT researchers who report it takes the human eye just 13 milliseconds to process an image (even though it takes 100 milliseconds for your eye to blink).

It all points towards an immense and looming opportunity for retail marketers. To make the most of your chances we have three areas retailers should focus on:

  1. Visual search

Although visual search has been around for a while (in 2008, TinEye became the first image search engine to use image identification technology), recently things have been hotting up, particularly in the area of ecommerce. Amazon is already active in the space and Slyce is emerging as a key player with its slogan ‘Give your customer’s camera a buy button’.

When you consider that we are on the cusp of augmented reality and virtual reality entering the mainstream, the possibilities for visual search in the physical world start to fall into place.

Most of us will have already experienced the world of Pokémon Go; imagine that instead of hunting Snorax and Lapras you were watching a fashion show in Paris and ordering your favourite pieces.

There are already apps like Blippar that connect brands with their audiences via augmented reality. Consequently, it isn’t hard to see the power of when augmented reality and visual search come together to allow a consumer to interact with a product before buying it.

  1. Frictionless buying

We are seeing more and more ways for consumer to buy with ‘less friction’. Similar to eBay Collective, US based Houzz (a hub for remodelling and designing homes) recently launched its ‘Visual Match’ tool which uses image recognition software to scan more than 11 million house photos to identify products, then matches them to the six million products available in the Houzz marketplace.

It has the potential to be incredibly lucrative; the company partners with brands like GE and Black and Decker and takes a 15 per cent commission on all sales and offers a completely new sales channel for product marketing.

Instagram is also trialling ‘in-app purchasing’ by introducing shoppable purchase tags that allow brands to tag their products, in the same way users can tag friends, with a link to purchase.

  1. Impact on social marketing

Machines equipped with image-recognition software are now able to sift through the rapidly growing number of user generated images on social media (Facebook alone reports that its users upload more than 350 million photos to its servers every day) and make sense of what is in them.

For retail marketers this means it is now possible to spot rising trends using visual-centric social listening.  This has the potential to be much more accurate than text solutions; effectively becoming a virtual ‘cool-hunter’ roaming around the social media world planet reporting on what brands and styles of clothing people are wearing (at a fraction of the cost of the real-world ‘cool-hunters’).

It also means that retailers have the opportunity to evaluate their products ‘in the field’. This allows them not only to understand how their products are being used – image recognition is getting extremely good at deciphering context – but also to fine tune their product marketing.

Product logo and straplines look very different under the harsh light of the ‘real world’ and brands are using social media images to ensure their branding is up to par.

All up it’s fair to say that even more change is coming to retail marketing and you can be sure that most of it will be powered by visuals. Make sure your brand is not left behind.