NEWS REVIEWS

BANNER FRAUD

Emily Steel wrote an interesting article in the Wall Street Journal on 12th October – on the subject of internet banner fraud, and assistant professor Ben Edelman’s work in the field of fraud monitoring. I thought it an article worthy of comment for several reasons.

Banner fraud comes in several obvious forms, not merely those mentioned in the article. It can include fraud in the form of misplacement, geo-fraud in the form of serving client advertising into inappropriate geographic territories, and other frauds including those discussed below.

So, if you have not read the article – click here to read it and then come back to me. OK? OK.
Most amusing was the presence of the Century 21 advert that appeared alongside the article. As I sat in London, I mused on the sheer irony that Emily’s words carried with them an advert promoting the October “open house” season for this very American realtor. How wonderful it must be to publish articles about banner fraud whilst committing geo-fraud at worst or bad client care at best.

Perhaps it is wrong of me to suggest that even if the client didn’t realize they were “probably wasting their money” when advertising their services in Uganda and Uzbekistan this October, a wholly professional financial publishing operation surely would respect notions such as “best advice” – and insist on delivering Century 21 campaign solely to US eyeballs. Sure, I know all the arguments that can justify this kind of geo-scattering – but still. Just because the client assumes all WSJ.com readers are American does not mean that the WSJ does not “know” this is untrue.

Does the “law of misrepresentation” apply to negotiations in which one business intentionally or negligently fails to represent the facts clearly to another? If the WSJ asserts that 65% of all dotcom display inventories are US audience, and 35% are international, does it fall on the client or the publisher to insist on US-only inventory in the buy or sell?

But let me cut to the chase:

Three years ago, Goallover developed a product we called “Digital Bobby” – referencing the British bobby - a policeman “on the beat” being one on foot patrol covering his area. This tool reported to us all the URLs upon which our client advertising was delivering. It allowed us to catch “misplacement” and also to spot check the sites delivering our client advertising for content that was unacceptable to our standard terms and conditions of placement.

This tool was perfectly good at catching more than 90% of all banner misplacement. It is available for license and costs approximately $2000 USD per month. It is invisible to the seller of advertising space, who will not know when they are being spot-checked.

So, much like the British Bobby, this is a spot checking tool that cannot be in all places at all times. But the point is that it does not need to be. The vast majority of suppliers, the professional ones, don’t misplace. So, when Bobby gives them a quick “aye, aye, what have we here?” the majority of suppliers stand to attention and do any house-keeping that is required. And once they know there is policing and, as with the penal system, recourse for recrimination, the vast majority of the population fall into line.

What we discovered was that by aligning the deployment of Digital Bobby with the correct terms and conditions to deal with misplacement and other forms of fraud, there really was no long-term problem to solve. Sure, there would always be a percentage of rogues in the market, but over 90% of the problem was solved in one fell swoop.

That left us deciding whether there was merit in developing a banner-fraud monitoring tool on a larger scale, and I concluded that there was not. I appreciate that some people have raised millions (an ex-associate included) to go off and build the “all singing and all dancing” Digital Bobby, but I think their investors and their efforts are a fruitless waste of life ambition.

Why? Because if they build a bobby capable of ubiquitous attention, then the fraud will disappear. It’s a law of diminishing returns. It’s like selling “usability” technology that tells you how your site can improve. The vast majority of the improvements will happen within the first three months – and then there’s just the endless rounds of A/B testing – which personally I find quite mindless.

And because – bear with me here for the technical part – DoubleClick, Atlas, Mediaplex, and other leading client side professional ad servers worth their salt could very easily deploy Bayesian logic – overlaying impression delivery and IP address analysis, post-click and post-impression mapping, on any particular domain – to “spike” and “flag” abnormalities or “surges” in impression delivery. They could integrate this kind of monitoring very simply, and put to the sword any company that thinks it has a claim to monitor and police the ad servers.

Now, if half a dozen leading industry ad servers agreed also to report in, and provide the IAB a checklist of domains known to cause unusual impression “surges” without the typical profile for clicks and resultant follow through, then not only can each ad server immediately suspend impression delivery to any apparently dodgy domain – pending a human operator check and manual approval – but they could also collaboratively develop a list of the known “perps”… you know, the usual suspects.

I could go further, and suggest that reputable ad servers do not provide their service to any website whose ownership is not clearly visible – because secrecy is usually associated with hiding something – and make a bunch of other suggestions here. But my point is that if you make a buy and make some good terms and conditions – do some Digital Bobby patrolling – instead of leaving it all to that nice Edelman fellow, this issue would be insignificant.

And thereafter, this matter would be pretty much cleaned up – banner crime would no longer pay.
So, it does not take a professor’s assistant to solve these issues. It takes a banner jockey from North London! Albeit a good one, a sharp one, perhaps one that has not covered all the bases here but:

1) If you want to use Digital Bobby to spot-check your ad placements – email us.
2) If you want to invest in or build an all-singing all dancing Digital Bobby – email us.
3) If you are the Wall Street Journal and you want to write about banner fraud – email us
4) If your name is Emily Steel and you want to know about even more disgraceful antics - email us
5) If you have already invested in or are building an all-singing all dancing Digital Bobby – oops!

© 2009 GOALLOVER.