SAS: We’re being swindled by robots
Posted: 1 June 2015 | Author: Rachel Alt-Simmons | Source: SAS
For marketing analysts, one of the beautiful things about digital advertising is it’s “trackability.” Every interaction on the web generates a piece of data, so in theory, you know where and when your ad was seen, who clicked on it, and any subsequent conversion.
But what if the online advertisements you were paying for weren’t getting to your target audience? Or what if robots were running rogue on the Internet and hijacking your impressions? Robots?? Bots – software that performs automated tasks - and Botnets – basically robot code that coordinates with other bots across a network – are bad news for consumers and advertisers.
The bad bots generate false impressions and clicks. Really bad bots can create phony user profiles and enter information into web form fields (anyone who has a blog and doesn’t turn on their spam filter will know what we’re talking about). They infect our computers. They redirect traffic to phony sites allowing the bad guys to falsely collect advertising payments through middlemen.
The “Chameleon” botnet discovered by spider.io in 2013 had infected 120,000 host machines and was emulating human visitors on certain websites, resulting in the display of billions of ad impressions. On the 200+ websites targeted by the bot, 65% of web traffic was accounted for by the bot, and resulted in an estimated $6.2 million in additional (false) advertising costs per month. Worldwide, the Interactive Advertising Bureau (IAB) estimates that 36 percent of all web traffic is generated by bots.
In addition to the problems that this causes in trying to measure the impact of your advertising message, the cost of the advertisement is based on impressions and clicks. It's estimated that corporations lose between $5-10 billion annually due to bot fraud. With an estimated more than one-third of all online advertising coopted by bots, marketers have no way to accurately manage or measure the impact and efficacy of their messaging and campaigns.
An industry consortium led by a number of marketing trade organizations has embarked on a “Making Measurement Make Sense” (3MS) project. Currently, marketers pay for online advertising space based on served impressions – which represents the number of times the ad is served to a “consumer” (and bot). The industry is trying to move towards a metric based on viewable impressions, a more accurate way of tracking whether an ad was seen by an actual person.
The challenge for marketers is that digital marketing is too critical a channel to ignore or defer until the problem is fixed. But while marketers continue to invest money in this space, they are more aggressively trying to monitor how that money is spent. Bob Liodice, CEO of the Association of National advertisers said, “When you bundle bots, click fraud, viewability and lack of transparency, the total digital-media value equation is being questioned and totally challenged.”
Many organizations are making investments in digital fraud detection technologies, combining session-based analysis with big data weblogs to track down, identify and block the bots. In the meantime, marketers can take action as well: Reduce digital fraud exposure risk by identifying traditional web metrics that can easily be faked by bots (impressions, click-through-rates, completion and last-touch attribution models) and using more concrete ROI metrics for campaigns.