KLUMME : The day started out the same way as any other. I walk into our Copenhagen office. I greet my colleagues. I get an espresso. I sit down at my desk to begin the morning ritual of reading and replying to emails.
To my surprise, I had received an email from a senior colleague of mine with the subject line simply stating: “Report”. Followed by a short message stating “this is why we combine offline and online”.
The report he was referring to is entitled “Quantifying Online Advertising Fraud: Ad-Click Bots vs Humans” written by Adrian Neal and Sander Kouwenhoven and published by Oxford BioChronometrics SA January 2015.
The paper which seems to be very well written (using “quantifying” in the title certainly makes it sound serious) describes the research findings by the authors, conducted over a 7 day period in early January 2015 testing ads on various popular platforms (Google/Facebook/Yahoo/LinkedIn) to determine the ratio of ad-clicks that are human initiated against those that are automated by “bots”.
Overcharging?
The report concludes that between 88 and 98 pct. of all ad-clicks were by a bot of some kind and goes on to state there are perhaps few industries where overcharging on such a scale as demonstrated here would be tolerated
Needless to say as a social media strategist that spends quite a bit of time advertising online and creating strategies and campaigns on behalf of clients, I found the report to be quite disturbing. I understand why my colleague felt it necessary to send me the report for further review. What I didn’t understand was the message that followed: “This is why we combine online and offline”.
His message was loud and clear, this is why we have little trust towards online medias and advertising online in general. I wouldn’t be sincere if I told you that I’ve never heard his concerns echoed in the voices of previous clients. After all, social media (as we know it today) is still in its infancy compared to offline medias, and who is to say that the number of impressions or clicks that the clients receive are real – up to 98 pct. are fraudulent as this report “clearly documents”.
A sales pitch
The only problem is that the report in question is not an academic paper. The format and presentation might replicate what an academic paper would feel like, but make no mistake this is not academia.
I clearly see why some people would make this error, as “Oxford” is plastered on the front of the report. It is a great compliment that Oxford University is such a well known institution (and brand) that by simply using the word “Oxford” in the title/company it immediately gives a degree of credibility to the writer and a sense of assurance to the reader. But this report is simply a sales pitch. Nothing more.
Oxford BioChronometircs is a company that unsurprisingly sells (you guessed it) bot fraud software. I find it curious that they wouldn’t state this very clearly in their “research” paper. Equally curious, I would like to point out their lack of transparency with their so called research.
What type of ads did they create? In which territories? What were the assumptions made to differentiate bots from humans? What wording were used? Were the ads made to attract people at all? What was the percentage per platform? Why is there so much missing information? Where is the big data to back up these claims? Why publish a “report” making such claim without even attempting to replicate the findings? Have 3rd parties to control the results, and so on.
Bot fraud exists
In my own professional experience of running hundreds of campaigns over the past years I simply cannot recognize or concur with the conclusions of the report. If I create an ad campaign to get signups on my clients website, and Facebook tells me they got 10,000 clicks, and the client tells me they got 3,000 sign ups, then I know without question that 30 pct. of those clicks were from real people.
Should I then conclude that 70 pct. of the clicks were bots? And 100 pct. of the people that clicked the ad signed up? Absolutely not.
To be very clear, there is certainly bot fraud online and the problem does exist on all platforms, it would be ludicrous to tell you otherwise, but the problem is not as rampant as some companies would have you believe.
Kate Dreyer at ComScore (a publicly traded leading internet technology company) recently published (“Ad Fraud and Non-Human Traffic: How Rampant is the Problem?”) that nearly 80 pct. of campaigns had less than 5 pct. non-human traffic. Let’s not forget; Facebook, Google (especially), Yahoo and LinkedIn all have massive teams dedicated to fighting this exact problem that has plagued their industry.
I have no data to back this claim up, except my own common sense: If statistically 88-98 pct. of ad-clicks were bots, people would know. Online campaigns would have little to no impact, and clients would not keep returning, but they do. Why? Because online campaigns have clearly proven to have an impact.
Fear mongering
Oxford BioChronometircs’s “report” is a poor attempt at fear mongering to sell their products. Their “sensational” findings have apparently worked to garner attention as some medias have picked up on the story.
In fact, I would like to take the time to publicly challenge Oxford BioChronometircs to replicate these results in a real trial, together with a bigger team, 3rd parties to control the experiments and data. I’m sure we can find a respectable institution that would love to conduct this type of research.
The business of online and social media marketing and advertising doesn’t just boil down to trust. Where traditional media can be touched and held – a poster is visible to the client, a TV or radio ad can be seen and heard, an online advertisement, more often is not.
Reporting to the client, and statistics on paper are not the only tangible links clients have that “something was executed”. Advanced enterprise tools such as Hootsuite, Falcon Social, and Komfo (to name a few) exist to help clients, online advertisers, and social media advertisers to track social media and the impact that ads and campaigns have on the client’s bottom line.
A click or an impression has never been a good indicator of performance for any online ad campaign. Focusing on measurable results, having clear ROI indicators, and tracking client’s budgets vs conversions will assure the transparency and trustworthiness of online and social media advertising in the future.
I attempted to contact Oxford BioChronometircs to confront them about their “research”. Unfortunately no number is listed on their website. I have also tried sending a message through their web-based contact form, and have yet to receive a reply.