Written by ViSenze
Data in the retail industry comes in many forms. The type of data which often receives the most focus comes in the form of numbers and letters as that type of data may be easily analysed, massaged, and put to work within the industry. When people think of “Big Data” they think of alphanumeric data.
Alphanumeric data generated by consumers and retailer activity has been used for centuries to increase revenue and margins. This is achieved through understanding and predicting trends, attracting consumers, and presenting those same customers with the right information at the right time.
An increasing number of retailers are turning to the data locked up in images and using that data to achieve some of the same goals that alphanumeric data achieves. We live in a visual world, with the average human eye processing 1,000 images per second. Our digital systems may be set up to process alphanumeric data, but our brands are obsessed with visual data. How can that visual data be harnessed in a way that helps retailers achieve their business goals? How can retailers take our human obsession with images and make their product and style images work harder for them?
At ViSenze, we have obsessed over these questions for the past 9 years. Using artificial intelligence, we have trained algorithms to recognise images accurately and at scale, thereby liberating the data therein. On top of these algorithms, we have layered a suite of solutions that harness this artificial intelligence. These solutions are easy to install and consistently deliver business value for leading retailers around the world every day. We will discuss some of those solutions below.
Smart Search
One way to realise the power of visual data and unlock the hidden data within images is to put the power in the consumer’s hands. Letting the consumer search by means of an image instead of text is a trend we see becoming increasingly popular with younger shoppers. Amazon, Pinterest, and Google have all capitalised on this trend, which has further popularised visual search. Not only does visual search convert up to 4x the rate of text search, but the search queries uploaded by consumers may be analysed (in a GDPR-compliant manner) to better understand customer intent.
Smart Recommendations
Online shoppers are denied the use of the sense of smell, touch, or sound. These shoppers can only rely on their sight, and it is important for retailers to both recognise and respond to it. One of the ways to respond is to provide recommendations that are visually similar to the hero product on the product detail page.
We have seen conversion rate increases in those retailers who have A/B tested logic-based recommendations against a visually similar recommendations engine, thereby proving that retailers who can access the power of visual data will have an advantage. As an example, a UK-based furniture retailer is on course for £20M+ in extra revenue simply by unlocking the power within images through this solution.
Shop the Look
Retailers spend time and money on beautifully styled model shots as they know that shoppers are more likely to buy a product that they can see someone else wearing. However, many of those same retailers waste the opportunity to up-sell the shopper on the other items that the model is wearing. For those retailers, visual data within the model images remains trapped in the model images. Our technology frees that data by recognising those other products on the model and bringing back the closet matching in-stock products. As an example, the global fashion marketplace that we work with is set to achieve $40M+ in additional revenue as a direct result of adopting this solution and freeing visual data.
Enrichment
Although we live in a visual world, alphanumeric data is still central to the way that the internet operates. In that alphanumeric world, both text search and search engines rely on good, accurate metadata to function effectively. We take large amounts of visual imagery and extract metadata providing accurate metadata related to style, occasion, colour, and other attributes, tailored to the unique nomenclature of each retailer. This approach has helped retailers to save time and money, as well as standardising their data. Our enrichment capabilities brought us into contact with Retail247, who have incorporated this into their PIM solution, Origin. ViSenze image analysis is called to determine, rich product attribution which is then used to rapidly enhance product data.
Forward-thinking companies like Retail247 understand the power that is locked within images. Companies like Retail247 have an edge on their competitors, who see the world in a flatter, 2D manner. We continue to enjoy collaborating with those forward-thinking companies who see the potential in the images all around them.
To learn more about ViSenze, email ross.gillanders@visenze.com