"Data-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Nicolle Pangis, global chief revenue officer at Xaxis.
With $6.3 billion destined for illicit hands this year, according to the Association of National Advertisers, ad fraud is clearly a major problem. But rather than impacting all players in the space equally, fraud is increasingly born by those who willfully, negligently or unmindfully make poor choices in their digital advertising efforts.
To understand why this is the case, we need to face two hard truths.
First, while one of the initial promises of Internet advertising was to provide a virtually limitless pool of inventory, and thus drive down prices, the amount of high-quality inventory on properties consumers actually view remains remarkably finite. Anyone can set up a website to accept ads and they do so to the tune of millions of new sites a year.
The more competition for ad dollars between these millions of undistinguished sites, the more likely they are to engage in questionable practices, such as buying traffic. Buying traffic means publishers pay external sources to deliver visitors to their site – the more visitors, the more chances to sell ads. With the incentive on volume, these “visitors” too often aren’t people at all but computer programs, or bots, designed to mimic human behavior.
The further out you get into so called long-tail inventory, the more likely fraud becomes as the competition for traffic and ads becomes fiercer and the consequences of engaging in fraud diminishes. If My-Fraudulent-Site-A gets shut down, it’s no skin off my back because I also own My-Fraudulent-Site-B through Z.
The second truth is that some of the ad world’s favorite metrics, such as clicks and views, are inherently susceptible to being gamed and must be vigorously policed to have any meaning. A bot can click away at an ad all day long. What it won’t do is actually make a purchase, which can be a more accurate way of measuring effectiveness.
In trying to optimize to potentially unstable metrics, advertisers and agencies are often in the paradoxical position of encouraging strategies that amplify the amount of fraud they encounter. A high demand for cheap clicks or CPMs begets more cheap clicks and CPMs. This is counter to what basic economics would suggest – namely that higher demand would lead to a lower supply of cheap clicks or CPMs.
Advertisers are waking up to the fact that those cheap CPMs often aren’t all that cheap. A $5 CPM with 75% bots translates into an actual CPM of $20, a whopping 300% price increase.
Brands and agencies that understand this industry backdrop – a limited amount of high-quality inventory and performance metrics that are prone to manipulation – have a significant edge in sidestepping fraud. Rather than chase CPMs across a vast pool of questionable inventory with questionable traffic, savvy brands can prioritize verified human audiences on inventory with high media health.
Being the smart money comes from having clarity on the two primary elements of the media transaction: who is viewing the ad and where they are viewing it. As any CMO who has ever pulled back the curtain on a dishearteningly high amount of ad fraud knows, these seemingly simple questions can be maddeningly complex.
Getting inventory directly from high-quality publishers and using traffic filtering and verification technology to vet each impression helps advertisers ensure they’re getting legitimate placements and human traffic. As soon as inventory enters the wild it’s subject to multiple forces that can quickly decrease its value. Most notably, as impressions enter the wider marketplace they can quickly pass through several middlemen in the vast ecosystem, each taking a cut and driving up the ultimate price that advertisers pay.
Think the direct route limits scale? In a way it does, though not in the manner someone arguing for the benefits of open exchanges might have you think. When one realizes that significant chunks of so-called scale, particularly the mouthwateringly cheap variety, are in fact fraudulent, avoiding them becomes a benefit. Focusing solely on high-quality, publisher-direct inventory, advertisers can currently gain scale across billions of impressions each day on the best media properties around the world and across all digital channels.
Publisher-direct media buying doesn’t obviate the need for filtering and verification of traffic. Even before getting to who (or what) is viewing their ad, advertisers need to analyze the viewability of each impression to make sure someone is actually going to see it.
Verifying humans versus bots is done in several ways. For example, by enabling pre-bid technology on the buy side, advertisers can be more confident a human will see their ad. On the sell side, suppliers can implement technology that verifies before the impression is even surfaced to an advertiser that it is being created by a real person.
Probably the biggest gut check for marketers looking to drive down fraud is expanding their performance metrics beyond a clicks- and CPM-only diet. These metrics have become the Achilles’ heel of the industry – fraudsters know they’re important and they’re conveniently easy to manipulate.
Rather than relying solely on these familiar standbys as a proxy for a campaign’s effectiveness, marketers should also introduce metrics that are harder to game. Sales are an obvious choice for measuring direct marketing campaigns. Other useful metrics include brand health, which analyzes whether a campaign enabled a lift in brand awareness or loyalty.
While ad fraud is a real and expensive issue it’s also largely avoidable. Optimizing toward publishers with high media health, accessing inventory at the source, verifying traffic and incorporating harder to game metrics are powerful prophylactics against digital fraudsters. However many billions this year’s ultimate loss tally is, smart advertisers can increasingly ensure it’s not coming out of their pocket.