The impact of 100% viewability on advertising economics

Nico Neumann
By Nico Neumann | 23 February 2016
Nico Neumann, senior research analyst, University of South Australia

Recently, we have seen two big holding groups pushing the agenda for viewable ads in Australia. Last November, IPG Mediabrands launched ‘project quality’ to aim for 100% viewability. GroupM even went a step further and announced in January 2016 that clients will not pay if an ad is not 100% viewable.

As most readers probably know, the Interactive Advertising Bureau (IAB) and Media Rating Council (MRC) offer viewability guidelines too, but whether or not these standards are adequate has been discussed fiercely among publishers and agency representatives: 50% of a display ad has to be in view for one second to be deemed viewable.

Is this definition really sufficient? Let’s be honest: who would pay for a TV ad that that is only 50% on the screen?

Hence, the two initiatives from IPG and GroupM are clearly steps in the right direction. It was about time that big agency networks put their foot down on behalf of their clients. Recall that Unilever, one of the biggest advertising spenders worldwide (and a client of GroupM), has demanded 100% viewability since 2014.

However, there is still a lot of confusion as to what “viewable” actually means. For example, when defending digital media in terms of viewability, people often bring up the argument that consumers may also go to the kitchen, bathroom, or simply check their phone when TV advertising is broadcasted.

But this argument misses the point as we need to distinguish between consumer attention and ‘technical viewability’, or better said, the opportunity to see an ad. Put differently, we also do not know whether someone actually pays attention to a mobile or desktop ad even if it was 100% loaded on screen. Any effects of media on customers can only be established through continuous research and analytics, such as well-planned A/B tests or attribution modelling.

In addition to attention versus opportunities-to-see, we need to differentiate between (1) unseen ads, (2) nonhuman traffic and (3) fraudulent ad schemes.

Firstly, many ads may represent legitimate placements but they just happen not to be seen by anyone because they run on browser tabs in the background or were at the bottom of a page. This problem is exacerbated by the fact that some publishers use auto-refreshes, automatically switching ads after a user has been on the page for a certain amount of time. Auto-refreshes inflate both website duration and impression statistics, thus generating questionable revenue.

Secondly, in contrast to unseen ads that just did not make it to the eyeballs of a person because of browsing behaviour, some ads will be served to bots, which can make up to 60% of internet traffic. Note that there are many ‘good bots’, which are needed for internet services, such as search-engine crawlers. In fact, some traffic will always be non-human and be responsible for wasted ads.

Thirdly, we have to deal with fraudulent practices: bad bots intentionally mimicking human behaviour (created to load impressions, generate clicks or fill out forms), ghost sites, or other dodgy methods, such as ad injections, intend to ‘steal dollars’ from digital advertisers.

Fortunately, many vendors offer technology to battle fraud and measure viewability of online ads. The only remaining question is:

Who along the ad tech supply chain will absorb the costs to guarantee 100% viewability?

Clearly, publishers will lose revenue due to unseen, wasted ads and the prices of digital ads are likely to go up sharply for three reasons:

1) Demand and supply

A recent study by Quantcast reveals that 97% of inventory is not available at 80%+ viewability thresholds. This also means that advertisers may need to adjust their expectations as to which audience they can reach realistically with digital media. Cookies could be created and bought in an infinite number, but reaching real people is a different story.

2) Adserving-opportunity costs

Even if an ad was not seen by consumers because of their browsing behaviour or bot traffic, the ads were still designed and served on the page, leading to ad-operation costs. Thus, viewable ads will need to generate enough income to cover the additional expenses for ads that could have been viewed but were not.

3) Verification costs

The buyers and resellers of online advertising – publishers, DSPs, ad networks/ exchanges, and trading desks – will need to use third-party technology to show that their ads were viewable and served to humans. The required analytics and verification tools will incur further costs, often bundled into CPMs or media spend percentages. Brands need to be aware that using this standard fee-bundling into volume metrics leads to an economic incentive to charge for as much traffic as possible. Therefore, advertisers and media buyers should request full transparency, including cost structures and viewability reports. In particular, any report should be obtained directly from the independent third party, otherwise there may be another case of "grading your own homework": there is a conflict of interest if the agency or DSP, which executes the media buying, also summarises, collects, or processes the information on how much they should be charged for their service.

Overall, we are likely to see major market disruptions with improvements in third-party viewability measurement and analytics. While marketing budgets are still shifting towards digital, CMOs may realise that they have overestimated the power of online advertising due to inflated numbers and by relying on easy-to-fake KPIs, such as click-through rates or served impressions. Given that clients may now understand that only a fraction of their budgets were actually used to show ads to humans and that digital ad prices will increase, it will be more important than ever to examine which channel, medium, or publisher provides a positive ROI.

Let’s all hope that this will not be done through last-touch attribution.

By Nico Neumann, senior research analyst, University of South Australia

comments powered by Disqus