At Curalate, we’re fascinated by the future of computer vision and machine learning. Those technologies have made serious strides over the past few years and the future looks incredibly bright.
Curalate is doing its part to define how artificial intelligence meets commerce with Intelligent Product Tagging — technology that can analyze an image and use machine learning to identify the products depicted within that image. For example: If you have a photo of a woman wearing a floral dress, our technology can identify that dress, then visually match it with the corresponding product in a brand’s catalog — making the image shoppable. We expect to start introducing this tool to clients in 2017, but it’s actually a much longer-term research effort for Curalate’s talented product development team. Read all about it here.
I caught up with two Curalate product managers to learn more about how artificial intelligence can modernize the commerce experience.
What are the benefits of Intelligent Product Tagging for Curalate clients?
Yiyi: It greatly speeds up the process by which brands can create intelligent content (photos and videos that are aware of the products inside of them), making it easier for their consumers to discover and buy products shown in photos and videos. We identify an object in an image and make a really good guess at what we think that specific product is. This makes it easier for Curalate clients to identify products in visual media that they’d like to tag. Not only does this save our brands time, but also it can be especially helpful for a new employee at a brand to get started in Curalate without first needing to become familiar with the entire product catalog.
Many brands also champion the photos and videos their fans share online. Maybe they find a cool photo on Instagram but the necklace that the person is wearing is not the exact same as yours — but you have a really similar item. Our technology lets you recognize that there is a necklace in the image and then search for similar necklaces in your brand’s catalog. From a shopper’s perspective, an exact match is great but similar products can be just as interesting — if not more — as they discover what resonates with their specific tastes.
Malini: Intelligent Product Tagging will make it far easier and faster for Curalate clients to tag products within the content (editorial or user generated). We were excited to hear from clients that they loved seeing the similar product suggestions, where previously a human would need to have a deep knowledge of the catalog to make this type of connection. Ultimately our clients find value in pointing their consumers to a product they might love, even if it’s not the exact product pictured.
Why is it so important for brands to tag their imagery?
Yiyi: Consumers discover products across many different touchpoints, especially social media. But too often, consumers are left asking questions like “What product is this?” It takes a lot of friction to get people from seeing a desirable product in a random photo on the web, to a place of purchase. Product tagging makes it easier for consumers to buy products where they discover them. It’s far more efficient than making consumers perform a series of searches to try to find that exact product. Tagging will lead to increased brand awareness, more sales for the brands, and a greater return on ad spend.
Malini: Curalate’s core mission is to shorten the distance between discovery and purchase. Just like being in a physical retail store, any time a consumer discovers a brand’s content online, that is an opportunity to make a sale. This can only be achieved if the brand’s content is made intelligent through product tagging, and then distributed to as many consumer touchpoints as possible.
What are some potential future applications of computer vision technology?
Malini: The potential future applications of computer vision are incredibly diverse, be it in image recognition, face detection, visual search, reconstruction or video analysis. I believe that over the next few years, strides in computer vision and machine learning will greatly impact e-commerce experiences.
Image recognition is a really exciting and hard space, and we are not the only tech company working on this. However, I do believe Curalate’s partnerships with 850 highly recognized brands sets us apart — through access to their product catalogs, and clients tagging products within content. We will only get better with time as our content library and partnerships grow. Intelligent product tagging was our first use case in this space, and it certainly aids our clients. We’re definitely excited about the potential future applications of the underlying technology, including:
- Powering customer-facing visual search.
- Showing related products when a certain item is out of stock.
- Helping to find cheaper versions of high-end products.
- Transforming recommendation engines to suggest complementary products based on previously curated looks.
- Helping brands increase catalog coverage with editorial imagery.
Yiyi Zhou studied Digital Media Design at University of Pennsylvania, and has built products at Microsoft and funnels at Julep before coming to Curalate. She would like to tell the whole world that Dargon is the best cat ever.
Malini Jagannadhan studied electrical engineering at Georgia Tech, and worked on the Photos, Videos & Camera Apps at Microsoft prior to joining Curalate. She enjoys a fully charged phone, a good poke bowl, and building a badass computer vision algorithm.
Curalate powers consumer discovery for the world’s smartest brands — turning pics and videos into portable, actionable content assets. It enables consumers to purchase your products wherever they discover your brand — social, ads, blogs or anywhere else. Get a demo now and learn how we can increase your average order value, return on ad spend and unique site visitors.