• by Josh Chasin , Op-Ed Contributor, April 9, 2015

In the last year and a half, a good deal of light has been shed on the issue of fraud in the digital supply chain. Trade publications (including this one) and, notably, the Wall Street Journalhave been reporting on the problem of non-human traffic (NHT) and fraud, raising awareness among publishers, agencies, and advertisers. Anyone who attended the IAB Leadership Meeting this past February knows that cleaning up the supply chain is one of, if not the, most important initiatives for that group in 2015.

At this point, it would be tough to blame advertisers for being skittish about their investments in the digital ecosystem. We’ve just finished telling them that half the ads in digital aren’t even viewable, and now we’re telling them about how rampant fraud has become. Let’s face it; as far as I know, there isn’t a lot of organized crime in the TV supply chain. We’re five years into 3MS, and undermining advertiser confidence is exactly the wrong place we should be right now.

But maybe things aren’t all that bad. In November 2014, my company conducted an analysis of NHT across the universe of all digital ad campaigns we regularly track. (It’s important to note that these campaigns skew toward larger advertisers and agencies -- and thus, the sites they tend to buy). For the most part these campaigns are branding campaigns, not CPA (where a lot of the fraud still lives).

We found that the extent of NHT (which includes, but is not limited to, fraud) varies greatly by campaign. A small number of the worst-performing campaigns generated the lion’s share of all NHT impressions observed. But the positive news is that four out of five campaigns in the analysis had less than 5% NHT:

• 79% of the campaigns had less than 5% NHT, accounting for 25% of the total NHT impressions.
• 14% of the campaigns had between 5% and 20% NHT, accounting for 45% of the total NHT impressions.
• 7% of the campaigns had greater than 20% NHT, accounting for 30% of the total NHT impressions.

From the publisher’s perspective, we observed that the top 50% of sites, ranked by lowest incidence of NHT, had less than 1% NHT, and the top 85% had less than 5% NHT-- indicating that the vast majority of sites perform well in providing inventory free from fraud and other types of NHT. It is only in the lowest-performing 15% of publishers across these campaigns where NHT was a sizeable problem. The good news is that tracking and measurement can allow an advertiser to optimize campaigns to avoid problematic traffic sources.

So where does the NHT come from?

First off, it should not be a surprise to anyone paying attention that cheap, blind exchange inventory is far more likely than premium direct to be fraudulent. Surely is it not a coincidence that as large swaths of the digital marketplace have migrated to the transacting of blind inventory on a real-time basis, the twin issues of viewability and fraud have become more pronounced.

Second, we have often seen publishers fall prey to an influx of NHT when they’ve begun working with a third-party traffic provider. When you source traffic organically — through search marketing, and via other, more traditional strategies for audience development and retention —the results tend to be, for lack of a better term, human.

But maybe a good rule of thumb is, there are no short cuts to audience development. In addition to introducing NHT to a publisher’s traffic, third-party traffic sourcing can also dilute the target audience of the site. One publisher targeting a particular, narrow demographic wanted to know why we showed their audience composition against that demo declining. It turned out, the traffic they were buying had a very different demographic skew — when it was human — than their earned traffic.

Third, there is still plenty of fraud in the CPA space. Click a button, get a penny — it’s an irresistible model for fraudsters.

The good news is that there are tools in the market that allow advertisers and agencies to understand which inventory sources are free from fraud, so that buys may be optimized accordingly. And while there is certainly way too much fraud for anyone to feel good about, more good news is that we saw four in five branding campaigns deliver in excess of 95% of impressions that were free from NHT of any kind. My pals at the MRC (George! David! Shout-out!) are working on standards for detection and filtration of invalid traffic, which, once ratified, will be applicable to all accredited digital audience measurement services.

So my message to advertisers is this: Don’t let fraud deter you. Digital works, digital free from fraud works even better, and the means exist to circumvent the fraudsters in planning and buying digital advertising.