UK Oracle User Group

Let's Talk About Image Recognition

13 October 2015

Oracle’s Maggie Schneider Huston speaks with Ditto Labs’ CEO David Rose about the future of visual listening.

Recently, we announced the integration of image recognition company Ditto Labs with Oracle Social Cloud’s Social Relationship Manager (SRM.) Oracle’s Maggie Schneider Huston spoke with Ditto Labs’ CEO David Rose about the future of visual listening.



Maggie Schneider Huston, Oracle: First, let’s go over the basics. What is image recognition?


David Rose, CEO Ditto Labs: Imagine you’re sitting in a public place and observing the behavior of those around you. Image recognition is the digital version of this. It’s an ethnography tool. Brands are using this data for customer insights.


MSH: Can you give me a specific example?


DR: Sure. One of our clients is the University of Illinois at Chicago. They used our tool to understand trends in teen smokers. They used to conduct surveys, but now they can look at the photos that teens are posting online. They’ve discovered that teens are taking cigarillos, cutting them open, and mixing the tobacco with weed.


MSH: Is that a violation of the individual’s privacy?


DR: We draw a very bright line between public data and private data. We are only reading public photos - usually those that are posted to Twitter and Instagram. We abide by the rights and regulations written in the social privacy statements.


MSH: How does this help brands?


DR: Brands are putting photos to work in three primary ways:


1. Social listening: This is a great way to learn about how your customers are using your product. For example, if you’re Red Bull, you can discover when people are drinking Red Bull, what they’re mixing it with, if they’re drinking it primarily outdoors or at a bar, etc. It’s a fast way to glean insights from information that already exists.


2 Visual Analytics: Companies put big money into advertising and social media. With digital listening, you marketers can see how often brands appear and how often customers engage with them. You can slice and dice that data by geography, gender, and other features. It lets you know how well you’re customers are adopting your product and keep an eye on the competition.


3. Customer Engagement: We give brands access to the social media handles of people who are using their product (or their competitor’s product.) So, for example, if you’re North Face and someone takes a picture of themselves wearing your clothes, you may want to use that user generated content in your advertising. Or, you may want to say something like, “Hello! It looks like you’re an outdoorsy person, would you like a coupon to REI?” We find that customers love it when a brand reaches out to them. We can also determine the influence of the person who is posting the photo by analyzing their tweets, followers, Klout score, etc. With this information, customer service representatives can triage posts more effectively.


MSH: What about brands that don’t have a natural use case, like Oracle Social Cloud? There isn’t really a good way to photograph our product.


DR: If it’s an object that doesn’t show up in photos, brands may want to explore affinities. For example, Dell could gain customer insights by tracking the people who follow Dell on Twitter and Instagram and exploring commonalities. That will allow them to understand the lifestyle of their customers.


MSH: What’s the takeaway from this?


DR: The conversation around brands is transitioning from text to photos. People aren’t describing their world in 140 characters; they’re taking a picture and sharing with their friends.

For more insights into visual listening, come see Ditto Lab’s Mary Tarczynski at Oracle’s OpenWorld.

Source: Oracle Social Spotlight


We'd love to get your feedback on this; you'll need a UKOUG login to provide it, so if you don't have one, please click on 'create a web profile' first.