The Problem With Online Ad Buying: Sonal Patel of AppNexus
In this two part series, Sonal Patel shares her thoughts about building a better, brand-safe Internet.
Recent incidents of online ads appearing alongside content that endorses violence and hate speech have put programmatic — and the use of algorithmic tools that automate the placement of online advertising — firmly in the spotlight.
But before we can address the current dilemma, it’s important to understand how digital advertising really operates.
Before the internet, the purchase of advertising was entirely manual — whether in print, broadcast, or out-of-home. Brands, usually represented by agencies, bought advertising inventory directly from trusted outlets. In most countries, print and broadcast media has faced strict restrictions on the images and language they carry. Traditional advertising therefore carried minimal risk to advertisers, since buyers essentially knew what to expect when placing an ad.
However, as consumers increasingly consumed news, film, music and information digitally, manual processes alone were no longer sufficient to handle the astonishing scale of advertising opportunities. When a consumer clicks on a page or opens an app, they produce ad “impressions,” or opportunities to place an ad where the user will see it. So to deliver billions of ads against billions of impressions in real time, each day, buyers and sellers increasingly use “programmatic” technology that stages an auction for each impression, decides which creative to show the end user, and delivers the ad — all in real time.
The challenge for brands, of course, is that it becomes more difficult to know where their campaigns are appearing. With millions of websites and apps, there are some that publish material that no brand would want to be associated with.
The issue affects both Singaporean and global brands alike. In February, it was reported that a video ad by the Singapore National Environment Agency (NEA) appeared on a website that has articles supporting the Islamic State in Iraq and Syria. An investigatory report from The Times (UK), also published in February, highlighted a whole host of incidences where household names, including luxury brands such as Mercedes-Benz, Sandals and UK-based charity Marie Curie, were unwittingly being placed alongside sites promoting extremist or pornographic content.
At the centre of the controversy is Google and Facebook, companies that control highly in-demand inventory that collectively accounts for 48 percent of all digital advertising spend.
YouTube, which is owned and operated by Google, has been in the headlines recently for allowing publishers promoting hateful and violent content to sell advertising space to some of the world’s most renowned brands. The scale of the problem is huge – each day, around the world, people watch up to one billion hours’ worth of content on YouTube. That’s a huge number of impressions for brands worried about being placed next to unsavoury content.
Not only is this an issue of brand safety, the practice is likely to generate tens of thousands of pounds a month for extremists. According to The Times (UK), an ad appearing alongside a YouTube video, for example, typically earns whoever posts the video US$7.60 for every 1,000 views. Some of the most popular extremist videos have more than one million hits.
Facebook meanwhile has also been at the epicentre of the debate, accused of not doing enough to combat the publication of misleading and outright false stories that spread quickly via the social network. As more and more people digest news online and via social media networks like Facebook, it’s more important than ever to identify and stop sites publishing fake news.
In part two, Patel will explain how to combat unfortunate online ad placements.
Sonal Patel is the Managing Director of AppNexus Singapore. AppNexus is an internet technology company that enables and optimises the real-time sale and purchase of digital advertising. Our powerful, real-time decisioning platform supports core products that enable publishers to maximise yield; and marketers and agencies to harness data and machine learning to deliver intelligent and customised campaigns.