By: Scott McDonald
The recent brouhaha surrounding Google’s inability to provide 100% assurance that brand advertisers won’t find their ads adjacent to objectionable content is just a dramatic manifestation of a long-simmering problem. As digital advertising moved out onto the Long Tail, advertisers traded some certainty about brand safety for lower ad prices. The problems accelerated as programmatic trading increased the reliance on complex data stacks and impenetrable targeting algorithms, with all of the opacity that results from such a technically complex ecosystem. For more than two years now, the industry has grappled with the problems of fraud and “bad actors” who exploit that opacity – with only fitful success. Now with the “discovery” that some ads follow their target consumers to dreadful places, the question of brand safety has become Topic A.
To some degree, the outrage is disingenuous since advertisers always knew that advertising on the Long Tail entailed more risk – especially when dealing with gigantic volumes of uploaded user-generated content. Mama always said that you get what you pay for, and she was right. This is not to say that Google couldn’t do a better job of opening its walled garden in the interest of brand safety reassurance. It surely could.
The incident ought to focus the industry more squarely on the neglected subject of valuing advertising context. Context is defined simply as the environment or situation in which the ad is experienced. In recent years, targeting has gotten nearly all of the excitement and attention, a reflection of the big advances in targeting technologies. We want to serve the ad to just the right person at just the right time. But the focus on targeting has tended to set the value of the context at zero, despite decades of research indicating that context has a big influence on advertising recall, brand perception, and persuasiveness.
Traditionally, advertisers used the context for targeting. Consumer packaged goods advertisers bought TV spots in the daily soap operas in order to reach housewives. Fashion advertisers bought pages in Vogue in order to reach fashion-oriented affluent women. Contextual targeting ruled and the media were quick to alert advertisers to potentially embarrassing situations that might arise, giving them the option to reschedule their ad to avoid problematic adjacencies.
However as targeting evolved in the digital space, ad calls were based on cookies, device IDs, 3rd party data overlays, and ever-more complex sets of supposedly person-level attributes. What’s more, the real contexts in which ads were seen multiplied dramatically. Contexts came to include not just editorial environments but also different contexts provided by platforms and devices, widely varying rational and emotional contexts (at work, standing in line, multitasking), and different social contexts. The proliferation of context abetted an expanded number of ads – with current estimates being that we are now, on average, exposed to 5,000 ads a week. So any individual ad now is more apt to be affected by the context set by the ad(s) that preceded it.
At last week’s annual meeting of the Advertising Research Foundation (an organization which provides education and research on advertising and where I was recently appointed CEO), a number of presentations dealt with this timely issue. Several sessions dove into research that used up-to-the-minute neuroscience tools to measure consumer attention, memory transfer, and emotional response. Most importantly, the ARF’s own research staff presented initial results of a multi-pronged, multi-year inquiry into context effects. This research engaged more than 25 ARF member companies in a partnership to measure the advertiser value of contemporary contexts.
Initial results align closely with the older literature. Some context effects work through attention transfer – a fancy way of saying that advertisers benefit when consumers are paying closer attention (ie. are engaged). Other context effects work through halo effects, in effect transferring some of the authority, legitimacy, coolness, or style of the context onto the advertised product. Consumer mood can be very important and also very volatile. One intriguing part of the study looked at the negative impact of having your ad follow a (negative) political ad from the recent Presidential campaign. Regardless of whether the ad was for Trump or Clinton, regardless of the political views of the consumer, the mere fact of being adjacent to the negative ads set a context that reduced consumer willingness to take the ad to heart.
The results presented last week are only the opening volley. A deeper dive is promised in the next installment to be delivered at the ARF’s Audience Measurement conference in June. But the timing is excellent since recent events have served as a useful reminder that targeting isn’t everything. Advertisers need to pay attention to context – now more than ever.