Every business leader sets goals and structures incentives against these goals. Incentives are powerful motivators. They can also have unintended consequences.
Marketers care about incremental outcomes—sales and brand metrics. At the campaign level, A/B tests can assess incrementality, but cost and complexity preclude them as a general solution for every campaign. So, advertisers use proxies like click-through rate and last impression served (last touch) to be the measurements of success and create incentive for ad buyer and seller alike.
But our go-to proxies have remained static, even as behavior has shifted and programmatic has changed how we advertise. Their application today sets up a cascade of incentives with unintended, often negative consequences that harm every player in our industry’s supply chain, and if left unresolved will undermine our entire ecosystem.
Click-through: unreliable proxy, engine for fraud. Outside of search, click-through mostly fails as a proxy. It does not reliably measure consumer interest. Since the Internet’s early days, advertising and search have been intertwined. We consume search expecting to click, so the click in this context says ‘this advertising was relevant for me.’ However, applying this yardstick to ads on a publisher’s site conflates two distinct consumer experiences. Consumers visit sites to focus on content—not to redirect by clicking on ads. Moreover, the fact that consumers choose not to click doesn’t mean ads had no effect. In addition, click-through acts as an engine for fraud across the industry.
If marketers valued campaigns on incremental outcomes, there would be an innate counterweight to fraud, as bots don’t buy anything. But as an industry we have incentivized click-through and created the market for fake click traffic: ad buyers goaled on clicks optimize to media partners delivering clicks; media partners, themselves optimizing for clicks, seek out traffic sources offering more clicks. True, at the ground floor fraudsters generate the fake traffic and clicks, but they act in response to the law of demand.
Look closer at last touch. The last touch incentive at least recognizes an ad can effectively influence a customer regardless of click. However, it oversimplifies customer motivation and also disappoints as a metric. Programmatic in general, and real time bidding in particular, took the conceptually easier task of retargeting and made it practically easier.
Retargeting is a much less daunting task when compared with prospecting—i.e., finding net new customers. With close to 300 million online consumers in the U.S., deciding who to target poses a challenge. However, if 150,000 consumers visit your site, that subset is 2,000 times more likely to convert. So, retargeting has always had a better chance at being assigned credit, as there is an implicit—and flawed—assumption that errors in that assignment would be randomly distributed among all impressions.
The advent of Real Time Bidding, however, has dramatically upped the ante. It has made large scale retargeting straightforward–and raised the widespread adoption of the last touch proxy from theoretical problem, to existential challenge. Before RTB, getting reach against a retargeting segment was hard. RTB made it so easy that anyone, and everyone, can do it. With last touch now the ubiquitous standard in every ad server, we mis-attribute credit for net new customers on a grand scale, and waste billions of dollars of ad spend that marketers could better use influencing customers to visit a site in the first place.
Retargeting does contribute value, but our current system dramatically overstates that value. It confuses a correlated relationship with a causal one. After a first site visit, consumers receive retargeting impressions. But that doesn’t mean retargeting causes their eventual conversion. The ad and the customer are simply in the same place in the buying journey, at the same time. Correlation does not equal causation.
Short-term pursuit of credit vs. lasting consumer engagement. Still not convinced we need change? Last touch credit rewards only ad delivery—not quality of engagement—and thereby encourages ad clutter, actively degrading the consumer experience. A content page with ten ad units has a greater chance of receiving last touch credit that a page with two ad units, despite the fact we all intuitively recognize that each ad on the two-unit page is likely to garner more consumer attention than on the ten-unit page. No wonder ad-blocking has become the issue that it is.
Relevancy redux: Simple solutions needed. As an industry, we need to re-align our incentives to better reflect real business goals. I’m offering a straightforward, easy-to-implement suggestion we can put in place today. As we work towards further refinement, this approach will help better assess engagement and provide insight into consumer preferences and motivations.
Split the funnel, give credit where credit is due. Split credit for a conversion, and assign a fraction to the ad shown to a converter before their first visit to the site (i.e., ‘last touch before first site visit’). This way we still recognize retargeting’s contribution via last touch, but also incentivize effective prospecting. This method is by no means perfect—it assigns credit to only a single impression pre-converter’s first visit. But it’s a highly advantageous fix we can apply right away.
Splitting the funnel rewards media partners that create real value, can’t be gamed like retargeting to capture last touch, and streamlines ad spend. It better aligns with good consumer advertising experiences, has the simplicity needed for ad server support and, most importantly, results in higher new net conversions. Today’s ad servers don’t support split funnel as a standard, but they easily could. For now, you have to generate a custom report, but moving forward as an industry, let’s work to make methods such as this a simple, standard metric we can apply to every campaign. Doing so will be transformative.
New year, new resolve. The incentives we rely on today poorly align with business objectives, foster fraud, contribute to poor customer experiences and obfuscate our view into the true component value of the media/advertising supply chain. This is a tragedy. Change has to come from upstream. Advertisers must de-value click-through and last touch, and adopt a new, more relevant, set of common currencies. Doing so will lead to better return on advertising investments and incent clutter-free, engaging consumer experiences.
Albert Einstein defined insanity as “doing the same thing over and over, and expecting different results.” As an industry, let’s revise the less effective measurements we have followed out of rote. Together, we can develop and adopt more relevant metrics that reflect our powerful ability to innovate, create and engage—and truly serve the objectives of all stakeholders.
Konrad Feldman is CEO and co-founder of Quantcast.
This week’s illustrations were created in partnership with students from the Baltimore Academy of Illustration. Click on the images below to view their website.