What are your biggest opportunities & challenges for next 12-24 months?
The big opportunity is to unlock all the visual information about brands and products held in videos and images and make it useful. Recent advances in machine learning and computer vision are going to bear fruit in widespread adoption of new tools for media analysis. The critical thing about these new tools is that they are visual, taking over from text analysis. Text search and analysis is simply easier to do, so lots of people do it. However, text search has clear limits. The big challenge for those of us in the visual analytic space is scale and accuracy, but we are making great strides as an industry in this area.
What keeps your clients up at night?
People are struggling to tell the story of right now. It’s obvious that something has happened and conversations have moved to social media. We all know there is a relationship between in person, traditional media and social media engagement, because we all do it. But, very few of us can articulate exactly what is happening and why. Marketers are still just getting their heads around it and frankly they lack good tools to do this kind of analysis. This is even more evident when considering the visual impact of marketing. Everybody knows that video and images are much better at grabbing people’s attention, but aside from video ads and some efforts at tracking virality, it’s not being measured very well. If you can’t measure it, you can’t talk about it, plan for it or influence it.
What’s the hardest thing to educate clients about?
The capability of machine learning has really exploded. Two years ago, we were at a plateau in terms of computer vision, what kinds of objects could be recognized accurately and how to train the algorithms. Machine learning and convolutional neural networks have changed the game, creating opportunities for people to do incredible things. These technology advances, in turn, power new kinds of strategies for marketers, showing them the graph of their target audience and helping them get in front of the right people visually.
What are some unmet needs in the marketing technology landscape?
I still think there is a need for an automated, large scale, in-image ad capability. Lots of people have tried it and thrown in the towel. GumGum has started experimenting with metrics. I loved Stipple and they folded. Luminate got bought by Yahoo and was never heard from again. This capability would be one of those tools that would be so powerful and obviously useful. No more key wording or tagging, no more guessing about what will engage – just track organic engagement, ID the object, pop in an ad and away you go. We are seeing new efforts by Pinterest and Instagram to add the “buy now” button to images, which is really exciting. But, as far as I know it is still being done manually.
What social network do you anticipate accelerating growth in the next year?
I actually like Twitter more than ever. They are addressing the troll issue pretty firmly, which should make people feel safer when participating. For real time activities likes sport and politics, it’s an awesome complement to TV. And for pre- and post-game activity, it even has the power to replace TV altogether. In addition, the efforts being made to create channels around events is a step in the right direction. I even see them breaking out of the media and breaking into the real mainstream this year, especially in the US.
–Yasha Spong is Director of Business Development at Zample and an expert in media fingerprinting.
