Right, wrong or otherwise, online technology innovations were – andare – sometimes driven by pornography, (See PORN: The Hidden Engine That Drives Innovation In Tech), gambling (See VR is a smart bet for the future of gambling), and crime in general.
Leaving aside moral judgements on those industries, they’re generally considered on the seedier side of business in general even if they’ve become more mainstream. I recall early in the digital business, (around mid 1990s), there were some companies that just didn’t want business from such industries more due to a higher incidence of fraud and such more so than any other reasons. Over time, of course, ad fraud has become more prevalent everywhere. It seems everyday there’s something about ad fraud in at least on of my industry news feeds. I’m starting to wonder what – if any – spinoff benefits there might be from AdWars. Pick a study, any study, and you’ll find ad fraud to be in the billions of dollars. Here’s just a small sample…
- SlideShare – State of digital ad fraud 2017 by Augustine Fou
- Businesses could lose $16.4 billion to online advertising fraud in 2017: Report
- The Bot Baseline: Fraud in Digital Advertising 2017 Report
- Adweek’s Ad Fraud Section
- ‘Biggest Ad Fraud Ever’: Hackers Make $5M A Day By Faking 300M Video Views
Where’s the money going? Who knows. From a criminal perspective, there may be some effort to set up some servers in some backwater $@%%hole country somewhere, but it’s not like you have to risk getting shot breaking into the quickie mart for a quick dollar. If you’re a terrorist in need of funds or a cash rich drug lord looking to expand your business without having to worry so much about shipments being intercepted, this seems like a pretty good business. On the other side, it’s of course not so great if you’re a brand who just wants folks to buy your New & Improved breakfast cereal. After all, you maybe used to have to worry about shelf space at retail and Share of Voice in the marketplace. But funding drugs and terrorists via the digital ad industry’s piss poor performance at curbing fraud? And also maybe having your ads showing up next to hate speech content? Yeah, this is kind of new. (And yes, I have read that ad fraud is projected to be down this year. And yet, it’s still BILLIONS.) Of course some isn’t to cloak and dagger bad guys, it just goes to some seemingly legit businesses that have gotten greedy. (See Confessions of a Fake Web Traffic Buyer.)
Here’s a press release I’ll bet they never taught you how to craft in MBA and Public Relations School…
Um… yes… To Our Dear Customers… We were also distressed to learn that the advertisements for our 100 year old family owned and oriented Happy Family Cereal Company have been running next to hate speech videos and on networks that apparently share revenue with terrorists, drug dealers and the companies who have been out of work from the crash of the Fidget Spinner market, who have since turned to crime as they were too antsy to wait and seek other new opportunities. We’ll try to do better, but we expect about 20% of our ad spend to continue to go such folks because the industry really has no solid control over such things. We always thought 50% of our ad buy was wasted, (that’s old news), but that was really just a branding thing, not outright theft. Oh well. Please continue to enjoy our cereal. We’ll try to do better. We’ll pull some ads for now until things settle down, but… you know… there’s not much we can do.
Any Silver Linings?
Ever the optimist, I wonder if we’ll get anything useful out of all this. If we lived in a world with no crime, sudden illness or injury or risk of fire, we’d save trillions in costs of police and emergency services. Similarly, the fraud in the ad industry costs not only from a Direct Cost perspective, but all the costs that go into Fraud Combat. Sure, there’s a lot of jobs in the cottage industry of ad fraud, but that’s not really the silver lining I’m wondering about.
Are there other uses for the technology being used to combat ad fraud? Let’s take a quick look at some ad fraud methods. (And sadly, this is just a partial list.)
- The most prevalent and obvious: Robots, (‘bots) view and click on ads. Since advertisers are paying to get to prospective customers eyes and not robots, this is pure fraud. ‘bot fraud, (or even manual fraud), gets more severe when dealing with traffic sourced from third parties.
- Faked mobile traffic.
- Hijacked devices.
- Hidden Ads. (ads served in single pixel stacks that will generate impressions, but are effectively not viewable.)
- Networks of fake sites with no real people place ads on pages via programmatic exchanges.
- Cost Per Click (CPC) fraud. Sites with no ’real’ content, but optimized for high value keywords.
- Cost Per Action (CPA) fraud. Fraudsters will automatically take actions, (like registering or other actions), to get a bounty for the action, even though it was a robot doing it.
- Agency Fraud: If that external bad guys weren’t enough, there’s been proven contract rigging, bid rigging and more going on, even at supposedly reputable top firms.
- ..and the list goes on.
Hey, You Said “Silver Lining!”
Oh, that’s right. Sorry. Yes, linings…
Out of all of the ad fraud, we’ve collectively been developing countermeasures. These countermeasures could spawn technology useful for other purposes. What are just some of the ways people are coming up with to combat ad fraud?
- Whitelisting. This seems fairly simple, but would require a trusted party to manage the lists. It’s probably easiest to sort out the white list for advertiser entities than publishers, but either way, at least a partially manual effort at a non-trivial cost. (Still, maybe worth it.)
- Viewability Tech. This tries to confirm that an ad unit is actually seen by someone. Now, to be seen by someone usually really means that the ad had the possibility of being viewed. That is, it was viewable within the viewport of a browser; not obscured “below the fold” or within a single pixel and so on. As well, that the browser was being used by a person and not a robot.
- Behavioral Analysis: To determine if it’s real people looking.
- “Big Data” methods to be able to process all of the ad fraud tech at high volume and high speed.
Over time, it’s likely some of the companies developing these technologies will realize there’s potentially new markets for similar technology. Maybe, just maybe things like viewabilty will be useful for other things. Perhaps usability for synthetic vision, AR/VR displays. That is, learning what elements are used and useful for users based on their usage. Maybe behavioral analysis to segment ‘bots from people can be enhanced to segment user cohorts. For example, do mouse movements behaviors differ based on customer age, gender or other factors?
What got me thinking about this was that I recently put together some presentations on ad fraud and other issues around the same time I was working on some Product Management efforts involving looking at things in different ways, how to discover new opportunities and so forth. As well, if you look at the Display Ad LUMA scape, it’s pretty clear there’s going to be some ad industry consolidation at some point. Not everyone will win in that vertical marketplace. But maybe there’s other places to go. (Since I don’t really work in ad tech directly, I’m not going to be the one to go and seek these strategies further. So I’m just throwing all this out there for fun.)
And What of the Advertising Industry Itself?
The technology being spawned within the ad industry seems to be of perhaps three general types:
- Operational: That is, just managing the complexities of the industry.
- Delivery: This is the whole supply chain with everything along the way from Demand Side to Supply Side and everything in between.
- Fraud Management: From view ability to theft and so forth.
It’s really only the 2nd and 3rd categories concern here. Operational SaaS products from asset management to client management to insertion orders and more are kind of a “ho hum” area. They’re important to business and there are valuable companies providing services here, but they’re not where the core action and concern is right now.
Delivery and everything about it is part of the problem. The industry is not very transparent. Programmatic has become such a complex ecosystem it’s allowed for all the weak points that have provided infection vectors for nefarious players. Even though most recently there seems to have been some decrease in fraud levels, I think this will get worse before it gets better. Consumer use of video has been spooling up big time, and it only makes sense that fraud ‘bot owners will follow them. The anti-fraud tech will catch up, but they’ll only be able to anticipate so much. The “good guys” will have to wait to see what the fraud patterns and methods are first.
Maybe all this is just wishful thinking on my part. But if we’re lucky, (or smart), we’ll perhaps find some other useful benefits from Ad Fraud industry technology For all of this cost to be going into what is essentially a non-productive effort seems a tragic waste.