ADOTAS – Visual content on the web is on the rise. High-speed Internet is pretty much ubiquitous these days, and it’s allowed us to access high-quality images and video online like never before. Consumers turn to the internet for engaging, visual content. For example, look at the popularity of Ellen Degeneres’ ‘Oscar selfie’ photograph, which received 80,000 retweets in three minutes and over one million in less than an hour, quickly becoming the most retweeted image ever.
Whether it’s through Twitter, Pinterest or Vine, social media sites are flooded with visual content produced by consumers, both from photographs to videos. As such, brands have the opportunity to not only deliver engaging visual content to consumers, but to also deliver the most relevant advertising based on users’ own content. But the challenge remains: When the content is user-generated, how do marketers ensure firstly the content is safe for their brand, and secondly, how to best target their ads alongside it?
Online advertising has benefited from contextual targeting for years now, but video advertising has tended to lag behind, as most tools simply analyse the page’s text rather than the video content itself. Advertisers need to tap into the potential of visual classification as a means of extending the supply of brand-safe targeted content. Put simply, textual ad targeting isn’t cutting it when it comes to targeting ads against video. For example, while keywords might indicate that a consumer has viewed a video about “sports,” visual classification technology can drill down and provide more sophisticated insights by identifying additional elements of the video content. For example, perhaps what was once tagged merely as “sport” is in fact classified more accurately as “surfing” taking place in Portugal. Discovering these specific details, and being able to automate this discovery, can mean the difference between poorly placed and expertly placed ads against video content.
We’ve found evidence that there is in fact an appetite among consumers for better targeted ads: our research found that over half of consumers (57%) believe that online advertising is too simplistic, revealing that this discipline is only as successful as its targeting. The main reason that consumers are dissatisfied with the state of online display advertising is because advertisers are only interpreting some of the data available to them when working out what ad a user might be interested in. This poor targeting can be put down to the fact that the continued visualization of the web is making it harder for publishers to accurately classify their webpage if they are relying on text descriptions alone, particularly for social publishers that host users photos. If brands ignore the visual content that is available to them, advertisers could completely miss the primary point of a web page, particularly true on social sites such as Vine and YouTube that are heavily video-led.
By classifying visual content, both publishers and advertisers have the potential to gain valuable data and customer insights. Publishers and advertisers need to respond to this with more sophisticated targeting to avoid annoying consumers with poorly targeted ads or ads that appear alongside inappropriate or salacious content. If publishers can process their visual data, advertisers will be able to better locate the most appropriate web page in which to place their ad. This opens up a mass of personalization opportunities for advertisers and publishers who wish to unlock the revenue potential held within their visual catalog, giving this content the same value as textual content.
This doesn’t need to be a laborious process for publishers. It’s possible to build ways to automate their visual content classification to ensure that content is safe and classified for targeted advertisers. It’s the fastest and most accurate way to ensure that ads are targeted effectively. With the ability to interpret user-generated video content in this way, everyone wins; brands are more likely to achieve direct engagements and increase click-throughs, and publishers ultimately will increase their revenues.
Automatic visual classification and verification is posed to be a massive game-changer, in terms of its innovative capabilities to correctly identify and classify visual content online. Marketers need to take this on board, or risk getting left behind. Automatic visual classification can also act as a tool for brand advertisers looking to serve up companion ads based on what a user is viewing on television in real-time. With “second screening” on the rise, there is great potential for brands and advertisers to leverage automatic visual classification to recognize visual TV programming using a mobile device and respond by providing targeted mobile advertisements in real time. This capability may still be in its infancy, but it’s not unbelievable that this is a next stop on the evolution of automatic visual classification.
The future is visual and offers many exciting opportunities for those brands who can act quickly to make the most of it.
