×

A New Currency for Measuring Ad Quality: Viewability

In association with OpenX

Wherever you sit in the digital advertising ecosystem, viewability matters. For brands, it’s a marker of high quality ads; for publishers, it’s the basis of inventory appeal; and for technology providers, it’s the currency everyone wants to trade on. But what makes ads viewable — and how can this be tracked when deals are automated?

These were the issues explored at the last OpenX School of Programmatic session for 2016. Taking a look at everything from mobile and video-specific challenges, to fraud, ad blocking and inconsistent standards, Robert Perine, senior product manager, OpenX, explained the current complexities of viewability and how to move beyond them. For those who missed out, here’s a run down of the key takeaways:

What is a viewable ad and how is it measured?

The concept of viewability is simple: it describes whether an ad has been seen by a user. The clear technical definition is that to be deemed viewable, ads must appear on whichever part of the page is on screen, known as the ‘viewport’.

But things get complicated when it comes to measurement. Opinion on how much of the ad should be on screen, and for how long, is divided. The Internet Advertising Bureau (IAB) state that 50% of a display ad’s pixels should be in the viewport for one second, two for video. Yet other organisations, such as Group M, state that 100% of a standard ad, and at least half of each video ad, should be seen.

How viewability is calculated sounds straightforward: code (often a verification pixel) is deployed, which measures the ad relative to the rest of the page and transmits data offline for processing. Yet implementing such processes brings a new set of variables, hazards, and best practices that must be taken into account.

The four considerations of viewability

Before marketers, brands and publishers can even start measuring viewability, there are four central areas they need to consider:

 1. Fraud detection should come first

As buyer demand for ads with high viewability rates has risen, so have the number of fraudsters cutting into brand budgets by faking user views with bots. This means that before viewability can be measured, it’s essential to identify and disregard all non-human traffic. Guidelines from the Media Rating Council (MRC) stipulate all vendors should be able filter out two types of fraud from reported metrics: general (including bots, spiders, and blacklisted IPs), and sophisticated (malware, ad stacking).

 2. Don’t forget about blockers

The second most important thing to bear in mind is ad blocking. Fraudsters don’t use blockers – they are actively trying to enhance viewability rates – so traffic from humans who have blocked (and not seen) ads will remain after fraudulent views are removed, which makes it vital to monitor and remove it as well. Yet, publishers must beware of overwhelming users by putting several ads on each page to bump up viewability scores, frequently serving ‘sticky’ ads that are always in view, or automatically refreshing ads. Those who haven’t decided to block may do so if the quality of their experience is impaired by advertising overload.

 3. When ads are not viewable

Once measurement begins, there are several instances to watch out for where ads fall outside the viewability threshold. They may be loaded in a hidden tab behind the one a user is looking at, obscured by another window, or fail to render completely. The latter of which highlights a further hurdle with latency – if ads take longer to load, they are in-view for less time, and viewability measurement cannot start until ads are fully loaded.

 4. Mobile and video are in a different league

While the industry has developed strategies for delivering increased viewability on desktop, we have not yet reached the same level with mobile and video.

The user experience on mobile is inherently different. Smaller screens drive unique behaviour – users zoom in and out on pages, shifting the viewport – ads in games are loaded long before they are served, making them hard to track, and mobile ads take up a lot of bandwidth, which can lead to latency. At present, mobile ads are still assessed using the same viewability standards as desktop; but this will have to change if the user experience and viewability rates are to improve. What’s more, there is inconsistent adoption of the Mobile Rich Media Ad Interface Definitions (MRAID) standards, on which viewability verification platforms are based.

The video space also brings a catalogue of variables. As video will soon need to be user-initiated, measurement of viewability will also concern whether users clicked on ads, chose to play them, and had the sound on or off. There are specifications too; VPAID (the Video Player Ad Interface Definitions), which players must support, and the Video Ad Serving Template (VAST), which should be used to deliver ads. And there’s more: issues like server-side stitching can make verification tech harder to use by stitching ads into the content stream before it is delivered the page.

So, in an ecosystem with countless browsers, screens, and devices, it goes without saying that accurately determining viewability is no easy task. Verification companies have their work cut out for them, as do publishers, brands, and ad tech vendors. What then does the future hold for viewability and the industry as a whole?

The next steps for viewability and beyond…
  1. Improving measurability

Moving from served to viewable impressions is a step in the right direction, but more work is needed — if viewability is to improve, we need better methods of measuring it at scale. In the meantime, we’re likely to see a shift towards private marketplaces (PMPs) where viewability and ad quality are easier to monitor and maintain.

 2. Focusing on user attention

The time users spend with ads is as important as viewability. Some would say more important. So, isn’t it time to start quantifying user attention? In the next few years, tracking the time each ad is in view, and how long users spend engaging with it, will help us understand the value advertising delivers for users – and how to do better.

 3. Creating clearer standards

If anything is consistent about our existing issues with viewability, it is the need for universal standards. There is a growing requirement for guidelines that define how the industry should transact and assign value to ads. If we acknowledge that user attention – not just viewability – is the end goal, finding a way to measure it in a way that’s meaningful, transparent and standardised is crucial.

That’s about it for the viewability round up, but not for the School of Programmatic. OpenX will be back with a brand new curriculum in 2017, bursting with insights, tips, and tricks. Keep an eye on the OpenX School of Programmatic page for more information.