Where Are All Those Clicks Going?

Last week Business Week decided to scare the world with a shocking cover: Click Fraud, The Dark Side of Advertising. This is about as timely as Roger Waters launching a tour where he plays the Dark Side of The Moon in full. If you recall, 12 years ago his fellow former Pink Floyd band mates had done the same. In both cases, you got to hear the Dark Side of The Moon but you didn’t get the full picture.  Which is more or less what you got in the Business Week article.  Yes, it’s true, there are cases in which as much as 15% of the clicks may be fraudulent, and yes, it’s true, the search engines may not be equipped to detect them.  However I felt very reassured when I read this article from an August issue of Fortune.  Apparently Nielsen has found a solution to the issue of measuring the impact of online advertising.  They have 30,000 households who will act as a sample, and they will be able to project from that sample the amount of  traffic that is really generated by online advertising.  Mmmm. Where have we seen this before?If you think I am taking the issue lightly I am not. 

These two articles point out a few major issues that as digital marketers we need to take extremely seriously.  First of all, whether we like it or not, the debate on the impact of digital marketing is now in the mass domain. The fact that mass media outlets are now covering the topic however does not change the fact that it is still a complex science. As an industry, we have a duty to set our standard higher than the standard expected by our clients or by the masses.  At the same time though it is extremely important to act as the voice of reason, and reframe the debate.  It is true that there are instances in which the measurement is not perfect, and it is true that there are instances of click fraud.  At the same time, for every 15% of fraudulent clicks, there is an 85% of legitimate, targeted and effective clicks.  The answer is not to go back and try to mirror the old media model, or even worse, to give up on Internet advertising. The answer is to keep working on improving the measurability of the web.

Comments

[Charlie Ballard]

While every article we see on click fraud is questionable on its own merits, together they do raise an important conversation around the value of each click, and who should be ultimately responsible for defining what a "click" actually is and how they're counted.

In the past six months we've run a number of test campaigns across a variety of new behavioral networks -- BlueLithium and DrivePM in particular -- and for the most part we were very surprised to see that the number of clicks the networks claimed were being driven to our clients' sites were an order of magnitude higher than the number of visits the client sites' measurement tools said they were receiving (via tools like Omniture and WebTrends).

Now, certainly there is always going to be fallout between a click and a measured visit, via browsers being closed, referrer information blocked, and other technical issues, but on average we never expect to see more than a 15-20% fallout rate. And here we were seeing over 80% of the clicks we were being charged for not arriving at the site.

Who's right and who's wrong? Are the networks measuring correctly but for some reason Omniture is malfunctioning... but only on the new networks? Or are the networks themselves measuring clicks in a way that doesn't translate into a visit in the way we've seen in the past?

Whatever the technical issue, the solution was clear: they refused to admit to any problem or begin to help resolve it, insisting their numbers were right -- yet we weren't getting clicks or sales. So in the next campaigns our media dollars were spent elsewhere.

As Dino says, the "[only] answer is to keep working on improving the measurability of the web." And until the publishers help us make this happen, we're going to just have to stick with what we know.