This means that more than half of us recognise, absorb and retain information better if it’s conveyed through a picture or video rather than text. Perhaps this helps to explain why images and videos are quickly turning into our primary form of communication when we spend time online.
Some of the world’s most lucrative internet platforms have been pioneers in adopting visual recognition services in response, like Pinterest, Google and Amazon, who have all rolled out the ability for users to search for online content using images alone from 2015 onwards. Users on Pinterest for example conducted 600 million image searches per month in 2018.
Image recognition is changing modern shopping behaviour, with some consumers now using visual search to locate products online. But what does this all mean for retailers who look after the merchandising and promotion of online products, and why should they pay any attention?
How can image recognition solutions be used to enhance merchandising and onsite browsing behaviors for shoppers so that retailers can create outstanding shopping experiences that exceed brand trading goals? The answer lies in visual recommendation technologies that combine the best of image recognition with the ability to optimise the entire onsite experience.
What constitutes visual recommendations?
Visual recommendations is a phrase that describes the unique processes that allow retailers to reveal relevant product recommendations and search results to shoppers when they search for a product using images. Visual recommendation solutions enable brands to automatically recommend products where patterns, style, colour or shape are hard to express in words, but are significantly easier to convey through images.
By using visual recommendations, retailers can empower their shoppers to search their entire product catalogues using images alone. Visual search and visual tagging functions interpret, analyse and tag visual attributes in products to make them searchable and visible in onsite search results. Through the simple swipe of an image, brands can also allow shoppers to browse with imagery in visual recommendations zones on category and product pages, enabling them to shop visual trends and editorial shoots. Products can also be revealed in personalised ways to match and enhance the trading goals of retailers.
When combined, these processes can make a massive difference to a merchandiser’s ability to enhance consumer engagement and revenues.
1. To expose your product catalogue
Retailers need to help guide shoppers to the most relevant potential purchases without undue delay. Implementing onsite visual search and visual tagging can enhance this process by exposing a brand’s vast product catalogue to visual shoppers in an efficient way.
By recognising attribute criteria within images and producing descriptive textual data based on this, AI-driven algorithms tag thousands of onsite products with a range of attributes that makes each product image searchable when a user decides to conduct an image search. Not only does this allow brands to showcase a huge range of their products to consumers based on their distinct visual searches, but it also gives consumers the ability to locate relevant products more quickly through the detection of attributes that are revealed in the visual tagging process.
With a virtually limitless array of characteristics, from colour, brand or design, the visual search service matches these characteristics to the attributes of the uploaded image. In this way, a social-obsessed Gen-Z or Millennial shopper who is glued to Beyoncé’s Instagram account for example can snap a screenshot of Queen Bey holidaying on a beach in St Barts wearing a pink wrap-around floral dress and find a similar item within seconds using visual search – visual tagging will identify all of the granular attributes of this image, from ‘floral’ to ‘beach’ to ‘pink’ and serve up relevant product images and filter menus that match these features. This is particularly important when you consider that in a survey by ViSenze (a visual commerce company) 62% of this consumer group said they want visual search capabilities in the shopping process.
2. To make mobile shopping seamless
Year on year, the percentage of people shopping on their mobile phones increases and given that 53% of visual searchers use their mobiles, it is easy to see why implementing visual recognition services is relevant when it comes to engaging this ever-growing market. Visual search works well for mobile users as we store and upload images on our phones regularly, so optimising the shopping experience for these actions is paramount.
Many brands continue to saturate their mobile shopping experiences with text-heavy navigations and descriptions, making it cumbersome for mobile shoppers to locate relevant products and check out quickly. But visual recommendations offer a handy way to improve these circumstances by enabling shoppers to shop with their eyes in fewer clicks. For example, visual recommendation zones can be activated on category and product landing pages to allow customers to shop multiple products that appear in one single image using ‘Shop the Look’ features, or ‘visually similar’ products can be displayed, to allow users to consider and quickly select items without having to click on complicated navigation menus.
By improving UX and helping consumers shop with their eyes, mobile shopping can be made much more seamless.
3. To enhance product discovery across different use cases
Another key advantage of implementing visual recommendations lies in the ability for Merchandisers to utilise the technologies in a variety of ways to suit different use cases. For example, visual recommendations can improve the in-store shopping process for ecommerce retailers with a brick-and-mortar presence by aiding sales associates in identifying the right products for their in-store customers.
Some brands have also excelled by combining visual recognition services with other services like live agents or chatbots to help customers discover the right products in the most simple and hassle-free ways. Shoppers can upload an image of a product when requesting help via a chatbot, saving them a lot of time and making it easy for them to express their shopping desires for unique cases that can sometimes be difficult to communicate in words (particularly relevant for sectors like lifestyle or fashion, where minute visual details matter most).
4. To empower flexibility in the merchandising process
Visual recommendations can also be viewed as a key medium for merchandisers. Ecommerce teams can tweak visual attribute weightings to produce results that are aligned with their merchandising strategies. For example, “Only return items when the ‘colour’ is 20% similar to what is being currently viewed” or “Only return items when the ‘style’ is 50% similar to what is being currently viewed.” Image recognition technology alone cannot do this, but by using visual recommendations, brands have full control over personalizing the display of images according to shopper and category needs.
By personalising display criteria based on the distinct features of a particular category, merchandisers can create recommendations that are much more in-tune to a user’s needs no matter what the product of interest might be. In categories where the colour black is ubiquitous for example, you need to have the ability to distinguish similar products based on much more intricate details, and flexible visual recommendation systems can allow brands to do this with ease across different categories.
Visual recommendations also give merchandisers the power to return products of a similar brand as well as other attributes like pattern, style, shape and colour. The weighting of these attributes can be manipulated so that merchandisers are free to specify in what situations they want to inspire their shoppers by showcasing an eclectic mix of new styles, or revealing old or ignored styles that simply need to be exposed from the catalogue and shifted to hit commercial targets. The flexibility of AI-driven visual recommendations makes this creative process possible.
5. To generate better results
Image recognition is a relatively new technology and only a small minority of retailers have so far built image search processes into their websites. However, the biggest difference that early adopters are noticing is the impact on performance. In 2017, Gartner predicted that by 2021, brands that have redesigned their websites to support visual search will achieve a 30% increase in digital commerce revenue, and some well-known retail names are already witnessing impressive results. High-street fashion brand Forever 21 delivered a 21% increase to AOV in the first month of adopting visual search.
Other brands are taking things further by using visual recommendations that combine image recognition with the power to adapt the merchandising and onsite search experience.
Generally, ecommerce companies don’t compete on discounts, but rather a seamless product discovery and an outstanding customer experience. Ecommerce teams, therefore, should start to embrace image recognition technologies, and visual recommendations in particular, before they are left behind.
Comments