“Suddenly the anti-bots business is booming,” according to Adweek’s Mike Shields in a story this week about several tech companies pivoting into the fraud prevention space. A more accurate take on the situation is that suddenly, more people see a way to profit off of a problem that advertisers have refused to address for years.
Ask any advertiser why they avoid real-time bidding platforms or exchanges and they’ll likely cite one of two problems: lack of transparency into inventory quality, and fraud, which refers to ad impressions generated by bots rather than actual consumers. These are both legitimate concerns, but they also serve as easy crutches for advertisers to lean on.
First and foremost, there’s no reason to fear “bad” inventory in RTB marketplaces. The technology to review an impression pre-bid and give the advertiser’s platform some sense of the page’s content before the bid is made has been available for nearly half a decade. Advertisers who use these technologies can set certain parameters and use the available data to avoid unwanted content adjacent to their ads, or simply avoid undesirable sites’ inventory altogether. If advertisers aren’t using these tools at this point and remain concerned about where their ads might appear or the type of content they are buying, well, that’s their loss.
The second piece – fraud – is a slightly hotter issue now, and some companies are clearly racing for headlines rather than solving the problem. There are more than 50 billion impressions available daily via RTB, according to our internal research. Any technology solution looking to address the problem needs scale across those impressions. That means they need to receive and process all 50 billion bid requests in the various platforms. Otherwise, advertisers are only getting the service on a small percentage of impressions – a drop in a very large bucket.
If tech solutions only offer pre-bid data on 2 billion impressions, they only see into roughly 4 percent of what passes through the marketplaces daily. The Rubicon Project, an SSP, estimated last year that they were seeing 6 billion ad impressions from the comScore 500 alone. That’s one platform, with a small slice of sites, and without factoring in 19 percent estimated growth this year (according to eMarketer). Advertisers may decide they want to use that solution and only that solution, which is fine, but they’re limiting themselves to a fraction of the available inventory and consumers. That means they’re limiting their audience pool and the chance for a successful campaign.
Even worse, some of the companies suddenly entering the fraud prevention space make it sound as if they are using human eyes to detect non-human traffic in their system. Even with an army of interns that is unprecedented in both size and sheer determination, the impact would be minimal. Fraud is a cat and mouse game, and there’s no way a human approach can make a real dent. Blocking suspect publishers is fine, but it’s not a real fraud solution – it’s a publisher integrity solution. Every publisher’s user base changes over time, so if the solution is not dynamic, inventory is unnecessarily restricted.
Truly addressing the problem of fraud in this business requires scale, and only a few companies in this space can process up to 70 billion impressions a day. Bot networks and other perpetrators will constantly try to outsmart the industry. If we want to root out fraud, we need to look at every single impression and give the advertiser a clear understanding of what they’re buying.
There is plenty of buzz around this topic as of late, and my view is that advertisers should take responsibility and use the tools available to them, which are good and solve the problem. However, they should also evaluate their partners to ensure that all claims are in fact accurate. Advertisers need protection – for them and their ads – from non-human traffic across all marketplaces. Full-scale RTB coverage is the only true measure of ones ability to offer fraud protection.