This week there has been a lot of media attention on Derek Muller, a popular YouTuber who has recently been unhappy with the performance of his Facebook fan page.
Muller created a couple of "exposé" type videos in which he talks about his problems with Facebook. This week, he discussed an "experiment" that he conducted to "test whether Facebook likes were worth the money." Links to more details and the videos here.
However, throughout the discussion, the media has overlooked one major detail: both Muller and the BBC reporter, Rory Cellan-Jones, whose work he references, made an obvious, easily preventable, clear anti-best-practice rookie mistake in the set up of their test campaigns that is the primary cause of their "scandalous" problems.
The Mistake
Both Muller and Cellan-Jones targeted their ads to include random developing countries like Pakistan and Egypt and were then shocked and outraged when they paid for clicks from those countries. Those users aren't even relevant and they are clicking! They announced. There must be something CRAZY going on!
So basically, the root of the complaint is that click fraud exists and that it is particularly prevalent in developing countries - welcome to 1999.
To anyone with even a passing familiarity with digital marketing, this is like making an exposé that investment bankers use spreadsheets. It is not exposing even a tiny morsel of information that is not widely known. It is, in fact, a key part of a long-established ecosystem that marketing professionals and ad platforms optimize every day.
As long as click-based ad platforms have existed, click-fraud has existed and every ad platform has had to invest resources in controlling it. Could Facebook do more to control Click Fraud? Yes. All ad networks can. Click fraud is driven by a revolving door of new gamers, requiring on-going clever and dedicated engineering resources to keep it in check. Is it slightly more puzzling why click fraud would be prevalent on Facebook versus a display ad network where page owners would make money from those fraudulent clicks? Perhaps.
But ranting about anecdotal examples of click-fraud caused by incorrectly targeted ads is not a smoking gun. It is, in fact, quite a distracting thing to do.
The Real Takeaway
The real takeaway that both Muller and Cellan-Jones missed, to the detriment of their audiences' knowledge, is that properly targeting your ads is a requirement for success on any digital marketing platform, including Facebook. When you don't properly target your ads, you shouldn't be surprised when your new irrelevant traffic isn't useful. This is digital marketing 101.
Targeting countries where you don't actually have users is both ripe for click-fraud, as they both discovered, and it doesn't help their businesses. In Muller's case, had he used his Facebook credits that he used for his "experiment*" to run ads against his target audience (based on location, interest and demographics), he wouldn't have been able to receive clicks from Pakistan. If Facebook had secretly targeted his ads to Pakistan anyway (a much bigger scandal), he would have clearly seen it in his Insights the next day when his US-targeted ads caused an increase in clicks from Pakistan. But this was not his accusation nor his "evidence," this bigger scandal didn't happen.
What happened was that Muller wasted his test budget targeting ads to the wrong countries, and then gained useless fraudulent followers who then hurt his Edgerank by not engaging with his page content. A cautionary tale perhaps, but again, not a smoking gun.
Perhaps it is Facebook's fault for not making it clear how to geo-target his ads? Perhaps they were implicitly driving click-fraud by making it complicated to limit his location targeting?
I present to you, Exhibit A - The Facebook ad interface, targeting section:
Not only is it very clear how to limit the location targeting of your ads (it is at the TOP of the audience targeting wizard), but the DEFAULT is set to only the US (it knows I am based in the US). This means that it would have taken proactive effort on the part of Muller and Cellan-Jones to add irrelevant countries, which to me indicates that perhaps they were digging for a scandalous story with the hopes that no one would notice that they achieved their irrelevant traffic through their own incorrect targeting strategy.
To me, this whole situation is a cautionary tale in a few important areas:
Learning From The Cautionary Tale
1) Properly geo-target your ads! Don't pay for irrelevant traffic or an irrelevant fan base, it is not good for you in the short or long run! (See the posts on CTR v. ROI and Social Media Must-Do Guidelines for more details on this point, as it is a core principle of digital marketing success).
2) Be wary of uneducated or inexperienced people making claims about how an ad system does or doesn't work. More likely than not, they did something wrong.
3) Don't blame others, including ad networks, when you make an obvious strategic error. It's bad form, and it distracts from your ability to optimize. Muller was so focused on exposing some sort of fraud at the heart of his bad performance that he missed the main learning opportunity - that it is more important than he realized to be careful in how he sets up his ads, and in how he uses a test budget. He is right that he would be better off had he not used the free credits. He is wrong that it is Facebook's fault. He could have easily made that money work for him.
Conclusions
In the end, this is a good reminder to all that there are rippling repercussions for making basic mistakes with digital marketing campaigns, even at a small test-budget scope. The good news is that there are tons of resources out there - from the platforms themselves and from blogs like this one, that can help advertisers avoid such obvious mistakes.
In the meantime, whether the media admits it or not, Muller and Cellan-Jones will have to dig deeper for their Pulitzer-worthy exposé, for user error and classic click-fraud are certainly not news.
Stay tuned and subscribe for more posts on Social Media and many other topics near and dear to the 2014 Digital Marketer's heart!
For help customized to your business needs, contact us at www.DigiMarketeer.com!
************
*Muller's "experiment" is by no scientific definition an experiment given that it has no hypothesis, no control, and no comparison of different variables of any kind. The fact that he is a science blogger makes this situation even more puzzling, for he must know the importance of properly structuring an experiment. Read the posts on The Science of Marketing and Scientific Ad Testing for how to apply scientific principles to marketing.
No comments:
Post a Comment