About the Coresight Matrix
With this report, we introduce the Coresight Matrix, a new product designed to help our subscribers better understand a number of segments and key players within the retail-technology industry. Our analysts are seeing the digitalization of the retail world and are helping clients achieve their digitalization and innovation strategies.
From our work with clients across large companies and small startups, we discovered that many need information on cutting-edge retail-technology, such as computer vision, visual reality, augmented reality and mass customization. They want to know who the leading players in the market are, which firms are driving improvements in these technologies, which startups might be the next rising stars and which startups big companies should be looking to work with.
The Coresight Matrix helps to answer those questions, by leveraging our expertise at the intersection of retail and technology, supported by our proprietary qualitative data analysis. We identify, evaluate and then position key players within the selected market based on two criteria:
- Innovation Effort: How much effort and progress the company has made to improve its products, technology and innovation strategy in general.
- Market Power: Where the company sits in the industry now and how much impact it could have in the market.
Later in this report, we include a market overview and provide more details about our methodology. We begin by detailing the topic of our first matrix: Computer vision for product search. The companies in our matrix represent large technology and software firms, retailers, and startup vendors that work with retailers.
[caption id="attachment_87758" align="aligncenter" width="700"]
Source: Coresight Research[/caption]
Companies Featured in The Matrix
Tech giants such as Google, Microsoft, Amazon and Alibaba are already leveraging the power of computer vision technology in their services. Below, we detail the companies featured in the top right corner of the Matrix, representing strong market power and leaders within the space.
Google’s Lens technology enables smartphone cameras to identify and search for objects based on how they look. It can now recognize more than 1 billion products from Google Shopping, Google’s retail and price comparison portal. The number of products covered quadrupled in the past year. Google Lens is leading the market with its broad application and its ability to analyze whatever content is in the camera at any moment.
Microsoft’s Bing Visual Search enriches customers’ shopping experience with visually similar images and products and makes recommendations accordingly. It also identifies barcodes and extracts text information from images. Microsoft Bing is leading the market through its high technical ability to find images that are similar to the composition of the target photo.
Amazon Rekognition allows customers to identify objects, people, text and activities. Amazon has also leveraged the power of computer vision in building its checkout-free convenience store, Amazon Go. Amazon Rekognition is leading the market through high sentiment analysis capabilities, deep learning algorithms, advanced face comparison and search features.
In China, computer vision is a key technology supporting Alibaba’s “New Retail” ecosystem, from searching for relevant products to running unstaffed stores. Alibaba is leading the China market by creating a well-rounded artificial intelligence ecosystem in which computer vision is a major focus.
Retailers are lagging in terms of developing their own computer vison products. Big retailers, such as
Walmart, just started to build out computer vision capabilities in recent years. Most other retailers, unlike tech companies, usually don’t have the funding for nor plans to build their own, so they usually work with tech startups.
Large companies are increasingly recognizing the enormous potential startup tech offers. We expect large retailers to continue to seek acquisitions of and partnerships with startup companies in 2019. Companies across multiple sectors have launched collaborative efforts to gather disruptive new ideas, harness new technologies and achieve competitive edge.
[caption id="attachment_87766" align="aligncenter" width="574"]
Source: Coresight Research[/caption]
In 2019, we expect more retailers and brands to adopt computer vision through collaboration with startups focusing on this technology. Below, we detail the companies featured in the bottom half of the Coresight Matrix, representing strong innovation effort and disruptors.
Markable.ai provides computer vision-based visual-recognition application programming interfaces (APIs) and software development kits (SDKs) to brands and retailers. Viewers can click or hover over the clothing they see while watching TV shows, movies and other digital media, and Markable generates both exact and visually similar product results. Markable has expertise in applying its technology to the fashion industry.
Cortexica provides artificial intelligence based visual search, video analytics and image recognition to retailers, pharmaceutical companies and manufacturing. The company’s findSimilar Fashion software matches images of fashion items and finds similar products within the retailer’s database. Its findSimilar Shoes software can recognize shape, as well as color, pattern, and distinctive marks. Cortexica’s team has a strong background in bio-engineering. The company has built on its innovation, replicating the functionality of parts of the human visual cortex.
Syte.ai develops visual search solutions for retailers, publishers, influencers and consumers with a focus on fashion products. Syte.ai works with retailers such as Marks & Spencer, Boohoo, Farfetch, Khol’s, Myntra and Intu. It also partners with Samsung to provide visual search technology. Syte is an outstanding computer vision startup that has worked with leading retailers and is expanding its business.
ViSenze provides AI-based visual search and image recognition solutions that help e-commerce retailers improve revenue and conversions. ViSenze works with retailers such as Rakuten, Uniqlo, Zalora and ASOS. It is supported by MasterCard Startpath and Unilever Foundry. ViSenze is backed by the Japanese ecommerce giant Rakuten and has a close connection to the National University of Singapore. We believe ViSenze has great potential due to its financial and technical capability.
Slyce, a visual recognition company, has partnered with at least 50 companies, including retailers such as Home Depot, Bed Bath and Beyond, Neiman Marcus and Macy's. Slyce works with many leading players in the retail industry, assisting them to digitalize their businesses.
While existing giants and startups in computer vision improve algorithms and expand applications for visual product search, we expect to see more new entrants.
Market Overview: Computer Vision for Product Search
Computer vision has seen rapid adoption in retail over the last few years while artificial intelligence (AI) is emerging an important new technology in retail. As the result of growing demand, more companies are implementing computer vision solutions, which will potentially drive sales and improve conversion rates. Improved computation power has led to major strides in the development of deep learning applications as it relates to visual search technology.
These solutions analyze more layers of data related to an image to determine objects and come to automated conclusions. Computer vision technology enables the faster identification of similar items that are in stock on a retailer’s website. It works by comparing the pixels in an image to identify and return results that are similar or an exact match to the initial image.
Online, computer vision technology provides shoppers with new avenues for to explore products and interact with retailers and brands. According to an eMarketer study, nearly three-quarters of Internet users in the US say they regularly or always search for visual content before completing a purchase, while only 3% state that they never do. This shows the importance of being able to quickly search for and see images of products.
Another advance in search capability is the use of computer vision software to analyze images on a pixel-by-pixel basis. Recently, large tech companies have made strides using computer vision algorithms for both text-based and camera-based images. Google’s Lens technology and Microsoft’s Bing Intelligent Search both enable smartphone cameras to identify and search for objects based on how they look.
According to a survey by Field Agent in March 2018, of 2,000+ US smartphone users 18 years or older, 83% responded saying that “product images/photos” are “very” and “extremely” influential in digital purchase decisions. Retailers put an abundance of product images on websites and in advertisements, but the images are difficult to sort through, especially finding what is available. Often, when searching for a product, a traditional retailer’s website will return an item that is out of stock.
How Many Retailers Are Using Visual Search Technology Today?
Today, many retailers and brands tag products individually and manually when cataloging them online, assigning each item keywords corresponding to the model name or number, size, color and other specifications. This limits the number of keywords significantly, making searching more difficult. To compensate, retailers generally display relevant items based on similar searches or purchases by other customers. Therefore, retailers tend to direct customers to a small fraction of available products.
As a result, recommended products generally relate only to items for which other users have searched, promotional items or ones with matching keywords. Search results and recommendations are therefore not directly related each other based on aesthetics but are linked to each other by manually input search terms and marketing data points. Furthermore, natural product discovery raises one of the most difficult searching issues: How does one search for a product that one does not know how to describe?
According to Alibaba, 10% of global shoppers use camera search by uploading photos to shop for products. Difficult-to-search items, especially those that shoppers do not know the names of, represent a challenge for traditional search engines. Computer vision software can help solve this issue by interpreting images captured directly in an app or uploaded from a camera. Like the audio recognition app, Shazam, software providers can make searching non-verbal material simple for smartphone users.
Retailers can leverage visual search technology to provide more efficient and engaging e-commerce experiences:
- Computer vision technology can automatically tag images with many possible words that the algorithms thinks apply to the image.
- Auto tagging eliminates the need to manually tag images, and significantly increases the number of terms per image.
- Visual search technology allows faster identification of similar items that are in stock on a retailer’s website, offering more complete-the-look features that often rely on information that developers input and hard code manually.
How We Create the Coresight Matrix
Coresight Research rates companies based on proprietary quantitative and qualitative analysis methods to demonstrate market trends, such as direction, maturity and participants. By leveraging our analysts’ expertise in the intersection of retail and technology supported by key data sources, we developed a methodology to determine the top platforms offering computer vision-based solutions to enhance the product search experience.
We evaluate companies based on two criteria: innovation effort and market power. This is what we found:
Innovation Effort (X axis): Coresight Research assesses companies’ innovation effort by evaluating product development, algorithm optimization, application expansion and technical research, looking at areas such as company patent filings and human capital investment. We normalize our innovation effort rating scores based on selected companies’ different backgrounds, so smaller startups can compete with larger industry leaders on our matrix.
Market Power (Y axis): Coresight Research assesses companies’ market power using two major metrics: value appropriation and value creation. Under value appropriation, we evaluate companies’ ability to operate efficiently and effectively, and to secure customers. Under value creation, we assess companies’ ability to make customers react to their products efficiently and effectively and to create an emotional experience for customer engagement. In addition, we consider companies’ backgrounds and factors such as funding, market position and number of employees.
After evaluation, we place each company into one of four quadrants: Leader, Disruptor, Legacy or Innovator. We base this as follows:
Leader: Companies that fall into this quadrant have high market power and high innovation effort. Generally, they possess a strong industry background with enough funding and technical ability to conduct research and develop cutting-edge products and algorithms. We see their effort and progress as bringing products and team to the next level through further investment and development. We expect them to continue leading the market. These companies demonstrate a clear understanding of market needs, they are innovators and thought leaders, and tend to have a presence in several geographical regions with a broad platform to support clients.
Disruptor: Companies that fall into this quadrant have a high level of innovation effort. They have a strong core business team and a good reputation in this specific market. Most are highly specialized in the field and have a clear understanding of market needs. On the normalized matrix, we sometimes find companies in the disruptor quadrant have even higher innovation effort scores than industry leaders, due to disruptors’ highly focused effort and market understanding. These companies have the potential to become leaders as they build more credible market positions and gain resources to sustain continued growth.
Legacy: Companies that fall into this quadrant hold large market share. The companies generally have a good market position and reputation in the industry but may have only recently started developing technology in this specific field. They have the funding and capability to develop a strong business in the field, but may still be looking for the right strategy, and they have the potential to move to the leader quadrant.
Innovator: Companies in this quadrant have high growth potential. Most of them are in the process of improving algorithms and technologies and are expanding the applications of existing products. They are seeking funding and have the potential to move to the disruptor quadrants.
