Criteo’s Jeremy Crooks and Sylvain Piquet consider the issues around measuring effectiveness in mobile advertising.
The IAB’s announcement of an industry-backed ad viewability standard is good, but not groundbreaking, news for advertisers and their digital agencies. It is however, without doubt, a step in the right direction.
We certainly need to see more accountability in display advertising, especially with such significant spend shifting back into that area. The latest report from IAB Australia/PWC reveals that display is showing the strongest growth in the online advertising market with a 27.5 per cent increase for the quarter ending 31st March this year, putting it ahead of search and directories.
A major focus for advertisers is the channel’s real impact on brand awareness, consideration and purchase intent. This is even more important when you consider the rise of programmatic trading and the diversification of ad formats, placements and inventory sources.
Independent studies, including one from C3 Metrics in 2013, have shown that over 50 per cent of display impressions are not viewed by people. Whilst the technical debate has only just started around what should be the ultimate ad viewability currency for each display type (IAB formats, Video, etc.) and device type (Desktop, Tablets, Smartphones), we are surely a long way from the end in an increasingly complex landscape. This is exactly why this very legitimate and healthy debate should not distract marketers from the bigger picture – how can they measure true ad effectiveness and manage display campaigns accordingly?
The placement and format of an ad can impact significantly on its effectiveness. On one premium news publisher’s website, the MPU within the article achieved double the user engagement (measured as click through rate x post click conversion rate) of the Leaderboard at the top of the page.
Not incidentally, independent research from Millward Brown (MarketNorms, March 2014) shows that this is also true about brand metrics: MPUs are 54 per cent more effective at generating message recognition and 44 per cent more effective at generating brand awareness than leaderboards. This can vary from publisher to publisher depending on audience behaviour – for example, users of a newspaper website may have a stronger propensity to scroll down the page than the average.
Another explanation would be that users are more receptive to display ads once they have read through most of the content they were looking for in the first place. Whatever the reason, these considerations are equally as important, if not more so, than focusing on viewability alone.
Display’s true effectiveness – incrementality – can be measured pretty accurately through AB testing protocols. In simple terms, AB tests measure the uplift in behaviour by comparing two statistically significant groups of users: the first exposed to the campaign-specific display ads and the second shown non-related display ads (eg. charity or house ads). So long as sampling is not biased and volumes observed allow for statistical significance, AB testing can be relevant for any type of display activity, including:
- Brand display: measuring uplift in brand awareness/consideration
- Prospecting display: measuring uplift in visits from new/non engaged users
- Retargeting display: measuring uplift in sales and conversion rates
However, even though AB testing is a robust statistical tool to quantify “how much” incremental activity is generated, it does not (yet) tell “which impressions” are creating incremental value.
We will undoubtedly get there in the not so distant future, but until then, buying on a cost-per-click (CPC) model remains a good way for advertisers and their agencies to ensure that their display campaign objectives are met in a transparent and accountable way. At the end of the day, if an ad isn’t seen it won’t be clicked on.
There are obviously downsides of using CPC – some customers who see but do not click on ads may still be influenced by them. While this does need to be taken into account, the metric remains for the time being the best proxy to measure performance. There is also the fear that CPC can encourage fraudulent behaviour. Again, steps need to be taken to address this and measuring closely post-click efficiency (conversions, new unique users coming to the site, etc.) will enable the problem to be countered if and when necessary.
So how do you ensure that a vendor’s and advertiser’s objectives are aligned through a CPC buying model? For this “ecliptic alignment” to occur, 3 golden rules need to be observed:
- The vendor’s algorithms should optimise towards end conversions, not solely on clicks
- The profile of display clickers should be in line with the advertiser’s target audience (especially your audience of potential converters)
- Reliable ad verification mechanisms should be in place to prevent issues such as click fraud and click farms.
Following these golden rules outlined above creates a virtuous cycle where impressions are bought on a metric focussed on effectiveness. Compared with the blunt tool of ad viewability, this methodology is more exacting and more valuable. Publishers who have highly effective inventory will also be rewarded with higher CPMs as buyers optimise towards the inventory sources, which deliver results for their clients.
All things considered, whether a display campaign has direct response or brand objectives, the ultimate goal for advertisers is to achieve these effectively. If these three rules are observed, a click will always be an excellent indicator of effectiveness and therefore value.
Jeremy Crooks is Managing Director of Criteo Australia & New Zealand.