Brand Sentiment and Deep Learning - Indix



Brand Sentiment and Deep Learning

There are many wonderful things happening in the world of the visual web. Here’s a crazy number: there are more than 1.8 billion pictures uploaded to the web every day. Wow, who’s sharing all these pictures? It’s us! It’s all the pictures that we upload to social media, Flickr, share on WhatsApp etc. The potential and need to leverage these photos is as obvious as the sky is blue. Technology is shaping the way that we discover and buy products on the web. Now the more that brands know about us, the better they can serve us.

Direct conversation between consumers and brands has never been as prevalent as it is right now. Hence, brands really need to do whatever they can facilitate this conversation further. One way that they’ve been trying to do that is by engaging in brand sentiment analysis. They use social media listening tools to analyze text and filter out the positive and negative words used with respect to a brand. The photos that people share on these channels is also a treasure trove of sentiment data, although it is more challenging to parse.

Enter, image recognition and deep learning technologies. Both these technologies are making radical changes in the world of Pervasive Commerce, where every interaction with a consumer is an opportunity to inform, educate and sell products. Image recognition technologies help businesses develop Product-Aware apps. What this means is that if you like your friend’s purse, you don’t have to just feel jealous and sulk. You can take a photo, upload it to an image recognition app like ASAP54 or Neiman Marcus’s Snap.Find.Shop, and the technology analyzes the shape, color, material etc. of the product and finds matches from across the Internet.

shutterstock_283272575When applied on the other end (to images that consumers share on say a brand’s social media page), image recognition combined with deep learning can discern logos and emotions on faces, among other details and help brands understand how and where their products are being used. Some social media networks have APIs that enable this kind of analysis, and there are also companies dedicated to this work. One company that has been doing this for a few years now in Ditto Labs. How amazing is it that technology can now parse pictures to determine the brand, the product, the sentiment of the person using it (smile or no smile?), the setting (what kind of tent are they using for stormy weather camping?), and so on. They can also determine what combination of products are being used.

What does all this amount to? A really deep and nuanced understanding of the target market and consumer behavior. What products are people using? Which ones are more popular? Where are they being used? This understanding can then be combined with advanced targeting techniques so people are offered the right product at the right time and at the right place. It can help make informed decisions on assortment and future product development. It can permanently replace the controlled and doctored environments of focus groups. Just the sheer volume and range of data is invaluable and cannot be replicated in any physical setting.

Every time such a trend gets adopted, it’s super exciting. What seems like magic when described is actually a reality that is changing our lives, making every decision one that is impacted and influenced by big data and science.

  Download the Pervasive Commerce White Paper
Product Awareness White Paper

Leave a Reply

Your email address will not be published. Required fields are marked *