Your address will show here +12 34 56 78


Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec porta, mi ut facilisis ullamcorper, magna risus vehicula augue, eget faucibus magna massa at justo. Sed quis augue ut eros tincidunt hendrerit eu eget nisl. Duis malesuada vehicula massa... Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec porta, mi ut facilisis ullamcorper, magna risus vehicula augue, eget faucibus magna massa at justo. Sed quis augue ut eros tincidunt hendrerit eu eget nisl.

Mobile has fundamentally changed the way we shop and discover through photos. Instagram users share an average 95 million photos and videos per day and like another 4.2 billion images. But linking this inspiration to commerce is a challenge. First, given an image it is hard to search for related products on social media platforms and content publishers ( such as Conde Nast or Bloglovin ), without a lot of googling and external search. Second, on the retailer side, online commerce websites with the exception of a few players, have lagged behind in leveraging user lifstyle and UGC content to inspire their users. Moreover, the search and discovery experience on online retailer sites lacks visual intelligence. In many categories such as fashion, users shop for a certain pattern, style or visual appearance. Current technology is inadequate to capture the visual appearance of products.

In this blog, we discuss these challenges and possible fixes. We start with the promising news that Instagram is now testing shoppable buttons on their regular feed. Next, we discuss the challenges for commerce and retail. Then, we discuss how emerging technology and artificial intelligence can help solve both these problems at scale.


  • Text Hover
Shopping Coming to Instagram

How do you find out more about that blue dress worn by your favorite celebrity on Instagram, and may be even buy it if the price is right. Now Instagram is testing a new feature that is the first step towards fixing this problem. This allows regular posts (in addition to promoted content) to carry more information regarding the products they highlight. So user can, for example, view more information regarding products featured in an image by tapping a button, which will highlight products with tags. These tag buttons click through to more detailed information regarding the product on the retailer page.
Initial pilot is with a few brands including Abercrombie & Fitch, J. Crew, Kate Spade New York and Levi’s. Currently the feature is available to a group of people on iOS devices within the US. When rolled out to all posts, this can become a powerful tool for retailers and bloggers to help the user find the products quickly, improving ROI on social media significantly.


  • User Generated Content
Adopt User Content and LifeStyle Images in Shop Catalog.

Online retailers feature images of the product and/or a model with plain background for the most part. On the other hand, today’s shoppers, especially millennials, are driven by experience and inspiration. They want to see what the apparel can reflect about them and their lifestyle. A real image of a user wearing it in a lifestyle context is more likely to convey this idea. Discovery and social moments are critical to the shopping journey.

User generated content has become more valuable than ever before. In the case of Instagram, the new feature will help retailers improve revenue from that platform. But user expectation is not restricted to a single platform. It is equally important to provide such content across the board, including product pages, storefront pages, marketing emails, blogs and style feeds, on both mobile and web. For this, retailers and brands can use services to harvest social media images, obtain ownership rights and feature them on their product page.


  • Text Hover
  • Text Hover
Faux Suede Midi Bodycon
  • Text Hover
Splendid Racerback Dress
  • Text Hover
Ribbed tank dress
Automatically Map Street Images to Products in Online Stores using Deep Learning (DeepView.AI)
Use Artificial Intelligence To Map Images to Shoppable Products

How do you take billions of images and map them to products in online stores ? Even taking one image and matching this with the right product is a long and arduous process for a human editor. Here, we’ll need the help of all the advancements in artificial intelligence and deep learning.

While anyone can look at an image and instantly tell a lot about the product and make assessments, machines haven’t been able to do this traditionally. However, deep learning has evolved over the last few years. It is now at a point where we can solve real-world problems. Advanced deep learning can ‘look’ at a user generated image or street style celebrity image and instantly understand the product in the image. Moreover, it can accurately generate attributes for the product and even match with similar products in online stores, as illustrated in the image above where a street image of a celebrity is mapped automatically to shoppable products, with zero human intervention. This means that billions of images generated by users on Instagram can now become shoppable automatically at scale. Imagine the possibilities of such a system. Any image, video or object in the shoppable environment can be ‘shopped’ live with no barriers. This augmented reality search will then become the future of commerce.
Deepview.ai provides solutions which will allow content publishers and retailers to enable augmented reality search systems, which automatically match street style content and user content to buyable products.


  • Text Hover
Improve Visual Intelligence in Commerce using Image Recognition

Keyword search on online retailers has suffered due to lack of search relevance. The process of annotating products with metadata is manual. This process is incomplete and inaccurate, especially at scale. So, keyword search results fall short in number of results and/or relevance.
Currently merchandisers employ people to manually tag products. This is not practical when there are thousands of new products added every week. Quality and accuracy is another problem. How to adapt to current trends in fashion ? How to make sure that tags and are aligned with keywords people use to search for products and are consistent across products, especially when these tags keep evolving over time? How to cover trending style-related keywords which require a significant amount of creativity. How to make sure they are continuously updated on an ongoing basis. For humans, this is error prone, and not effective, efficient or consistent. Also, in general, it covers less than 50% of relevant trending keywords even with a conservative estimate.

Image recognition can solve this challenge by automatically identifying product features in images as a human would. No need of any manual tagging and no more worrying about missed attributes and outdated tags. This in turn makes the search experience a whole lot more visual for the user. We refer you to our blog on fixing search on online retailer sites with deep learning where we discuss this in more detail.


Leave a Reply

Be the First to Comment!

Notify of