Centro's Programmatic expert, Ratko Vidakovic, presents at the 2015 ad:tech New York on Ad Fraud, its many forms and how to best protect yourself from it.
Who I am
Housekeeping items
Q+A at the end
Link to deck online
Some basic knowledge assumed
Twitter hashtag: #stopadfraud
1. Centro is unifying digital advertising. It has software, it has services, it works with thousands of brands and agencies, and tens of thousands of campaigns.
We are going to be covering lots of ground today around fraud
What it is, how big of a problem is it, examples of fraud, why it exists, who’s responsible, how do you guard against it, and so on
Respectful of time, so I’ll move fast (if I don’t jump into detail, ask in the Q+A)
Jugular problem and why
In the news recently
Blew up with BusinessWeek feature
ANA/WhiteOps study that got attention
Stats
Footnotes about the stats
ANA report was aggregate
Based on what we see, much more nuance
Lots of issues mixed under the “Fraud” banner
Let’s start with a quick definition before we move forward into examples
Biggest mistakes I see is that many definitions of fraud limit it to non-human traffic, but it goes far beyond bots.
Human users can also constitute fraud, especially when the ads are misrepresented.
Read definition
Here are some examples of non-human or bot traffic.
This is where most anti-fraud companies focus
And where most of the media attention and studies go
Here are some examples of fraud that are committed against human users
More sinister because they are harder to police (in some cases)
Another important distinction is between fraud and low-quality
Just because publisher is low-quality (viewability, performance, brand safety, etc) doesn’t mean fraud
Examples…
Here are some examples of how low-quality can invade the marketplace, but isn’t exactly fraud (grey area)
Again, all of these things up here are human traffic, not bots, so harder to detect and police
One of the most insidious is “arbitrage”, which we will talk about quickly for a second…
Real quick, if you see a site in your reports that looks fishy, or has suspicious performance, follow these steps
Go through steps
Add bad sites to your blacklist, and possibly report them to your DSP
Now we’re going to talk about why fraud exists and persists in the ecosystem
Main reason (#1) is open marketplace (pros and cons)
The only alternative to open marketplaces are closed platforms (like Google + Facebook), which have their own tradeoffs
The second (#2) biggest reason fraud exists is because it isn’t illegal, yet highly lucrative
Furthermore, higher payoffs mean higher levels of fraud (e.g., video)
Not advocating or endorsing fraud, but bottom line: it’s low-risk and high-reward (no brainer, easy money)
Third biggest reason (#3) is the misaligned incentives of the players, practically across the board
Read Ben Thompson quote
What Thompson means by improper incentives
Biggest victims of course are the brands and advertisers that are financing fraud
The second biggest victim are Internet users (now taking things into their own hands: ad blockers)
Bottom line: primary victims are advertisers and internet users. Secondary victim is the ad-tech industry as a whole.
Sounds cliché, but it’s a shared responsibility by everyone
But everyone is scared to talk about it (Pandora’s box)
What ends up happening is the Chain of Pressure and blame
As we go through supply chain, let’s start at the source: Publishers
Read quote from Businessweek
What this quote suggests is that some publishers are committing fraud intentionally, and some accidentally
Brings us to ad-tech ecosystem, specifically the SSP and DSP (this part is tricky and important, so pay ATTENTION)
Before there were ad networks
Now we have divided supply and demand… starting with SSP, then DSP…
Not a perfect world, so what can advertisers do to protect themselves from fraud that makes it’s way through?
These best practices really all tie into each other
Read them all
How the first two feed into third practice