Commentary
Let's Not Forget: Digital Advertising Moves Products
- by Josh Chasin , Op-Ed Contributor, November 11, 2015
Sure, we want to optimize for viewability. Sure, we want to ensure a trustworthy supply chain, and to purge fraud from the ecosystem. And sure, we want consumers to willingly accept our ads.
All of these things will make digital advertising better. But right now, digital advertising is delivering results for thousands of advertisers, including the world’s largest and best-known companies, who shift proportionately more of their precious adspend to our space each year. So today, let’s focus on some new findings that remind us of the continued efficacy of digital advertising, and of ways to use it even more effectively.
Latency planning can improve
campaign effectiveness
A colleague of mine recently published a paper that demonstrates that ad latency (how long an exposure “works”) can be empirically measured at the
campaign creative level. The paper suggests that planners could improve digital advertising ROI by optimizing based on latency, as opposed to simple frequency caps: the longer the latency, the less
frequent the need for incremental exposure.
Conversely, though, latency planning suggests that incremental exposures are required to maximize efficiency as the length of the campaign increases. Rather than the somewhat blunt instrument of a frequency cap, latency planning suggests that exposure should be ongoing throughout the campaign, with time between exposures varied by latency of individual creative executions.
Another interesting finding in this paper was that the latency of video impressions is significantly longer than that of display ads. Some video impressions demonstrated latency (that is, had measurable impact) in excess of a week.
Non-viewable ads work too!
Yes, I know that statement is paradoxical. The catch
is, we’re talking about ads that made it onto the screen, but in a fashion insufficient to qualify as viewable based on industry standards for either % of pixels (50%; 30% for large formats) or
duration (1 second; 2 seconds for video).
A second paper found that ads that made it to the user’s screen, but failed to qualify as viewable based on the industry standard thresholds, can (but don’t always) contribute a measurable, positive impact on overall campaign performance. Still, the paper concludes that “In all cases, viewable ads were substantially more effective than non-viewable ads, supporting both the definition of viewability (for display) and viewability measurement.” This also means that ads work better if more of the ad makes it on-screen for a longer time -- intuitively obvious, but nice to have validated.
Context matters
In the TV space there is a lot of research to demonstrate that program
environment (and viewer engagement) has a material bearing on ad performance. In, digital though, with so much business shifting to programmatic channels where the advertiser often doesn’t know
in advance where the impressions will run, I think we’ve paid far too little attention to the impact of environment on ad performance.
So I was happy to see that in some of the work my company is doing on ad effectiveness, we are seeing measurable variability in ad performance by publisher environment. It will be important to dig further into this topic, but I’m certain based on what we’ve seen so far that the same piece of creative shown to the same consumer will perform better if the exposure is via a prominent placement on a site offering premium branded content.
Mobile ads are high performers, particularly at the bottom of the funnel
As consumers continue to shift their share of time spent with digital media from
computers to mobile devices -- in September, 65% -- mobile ad effectiveness becomes an increasingly urgent topic for advertisers and publishers. We’ve seen mobile ads delivering over three times
the lift of computer ads with respect to purchase intent -- perhaps because consumers engaged with mobile media are by definition closer to the point of purchase for any goods or services purchased
offline.
While small screen size is an issue that advertisers and agencies need to navigate, mobile’s real promise may well lie in the extent to which mobile platforms enable advertisers to reach consumers when they’re ready to buy.
To be sure, we all still have homework to do about viewability, fraud, ad blocking, and whatever issue rears its head tomorrow. But we should remember the strong place we’re starting from, and be optimistic that continued research advancements will only make digital advertising stronger.
"A second paper found...In all cases, viewable ads were substantially more effective than non-viewable ads..." And you wonder why we have a credibility problem in online advertising?
Hi Josh. Where can I source a copy of the second paper you refer to.
It seems counter-intuitive that something that is not viewable has an effect. Maybe James Vicary was right after all (even though he admitted he faked the results)!
So I was wondering whether it was possible that in these cases there was mis-attribution of causation based on some other correlated variable of which the 'unviewable impression' was merely a cohort.
I'm not necessarily debating the points made here, but it's bad form to mention sources that prove your points without linking to them.
I continue to be amazed that there are those---not you, Josh----who may still need to be convinced that ads that can be seen as complete stories or sales pitches are more effective than ads that are "viewable" for only one second---the "industry standard". Why this obvious fact needs documentation is beyond me.
As for TV, while there is a huge amount of evidence that involving content---serious dramas as opposed to silly sitcoms or reality fare----has a positive effect on commercials this is mainly a function of commercial exposure, due to less zapping. Consequently a somewhat higher proportion---perhaps 10% more than usual---of the program's audience, gets a chance to see an average commercial in a more involving context. However, this does not mean that those extra viewers who do see the ads buy what they are saying as this, the most important part of the equation, is determined more or less independently and the advertiser's claims must stand or fall on their own merit.
Yes, can someone PLEASE provide links to the quoted sources - especially the paper that "proves" that non-viewable ads work too.
I suspect that "the paper" in question is part of the IAB's preliminary work that was used to establish its unrealistic "industry standard" for ad "viewability". I'd be very surprised if it's a really objective or definitive study---but I could be wrong on this. Let's see.
Folks, I think pou're missing the caveat of the "non-viewable ads work" point. Of course you have to see the ad for it to have an effect. I was very clear--
"The catch is, we’re talking about ads that made it onto the screen, but in a fashion insufficient to qualify as viewable based on industry standards for either % of pixels (50%; 30% for large formats) or duration (1 second; 2 seconds for video)."
The ads are on screen and register with the consumer. But they do so either for less than a full second, or for less than 50% of pixels. Therefore they are not defined as viewable ads. The point is that even impressions that get to screen but don't qualify as viewable based on the % pixel or duration requirements still can contribute to campaign effectiveness.
As far as links to the papers, both were presented by Steven Millman at the PDRF conference in London in October. The PDRF makes the papers available for general download after a fashion; right now they are only available to attendees. Soon they will be online. In the interim, shoot me a note (jchasin@comscore.com) and I'll be happy to propvide both papers. By the way, they earned Steven the prestigious Chairman's award at the conference ("prestigious" being Steven's word, not mine).
Josh, I can't speak for the others but as for myself, and focused only on video commercials, not static ads, I have a hard time accepting the idea that less than 100% visibility is worth much to an advertiser who's commercial is telling a story or making a sales pitch that has a beginning, middle and end. I'm sure that one can go out and do a study that demonstrates that those exposed to a random second in the commercial have a little value, in terms of ad awareness relative to thoss who could not see the ad at all, but I'm afraid that this is not going to impress many branding advertisers---at least not those I talk to. If digital ad sellers want to ask for CPMs that are higher than TV's average---not primetime broadcast, but all forms of national TV--- for video ads, they will have to abide by the same definitions of exposure.
I don't think that anyone believes that digital advertising doesn't move products, but so does TV as well as radio, magazines, newspapers, out-of-home billboards, etc. All of the non-digital ad sellers offer what amounts to a 100% exposure opportunity per ad, in the sense that if the audience is so inclined it can see the ad in its entirety. Digital does not provide the same degree of exposure opportunity for many of its ad "servings", yet it wants to get paid for any and all exposure opportunities, even if they come in bits and pieces, and there are those who actually maintain that advertisers should pay higher CPMs if they want useers to have a full opportunity to see their ads. This is what bugs many of us, hence requests to see the research in question.
What a load of old bollocks... All it talks about is "exposure" and "viewability." Not a word of proof that it actually "moves" products. Something we used to call "sales." If in doubt... Obfuscate!
I can't speak for how this author arrives at the idea that "Digital Moves Product". But what I've seen in my pretty extensive attribution work is that digital is very often given full credit for moving product when it's role is quite small.
Essentially, it takes a village to move product. But the easiest thing to analyze is last touch - the advertising immediately prior to purchase (one of digital's roles). So, in 99 out of 100 cases, companies accept last touch measurements and conclude "our digital works fantastically".
There are many truths they miss - here's a couple... Many digital attributions are sales that would have happened anyway - WITHOUT the digital advertising. This is particularly true with SEO. Probably 40% of SEO driven leads were paid for unnecessarily. And also, 90% of the cost to develop someone into "ready to take action" happens outside digital - with traditional advertising. Yet companies calculate digital in isolation - and overestimate it's impact by a factor of 10.
Honestly, this was the most innovuous piece I've written here. I'm shocked at all the hubbub.
The point-- stated up top-- was that "...we’re all collectively guilty of 'burying the lede': that digital advertising works, persuades consumers, moves products" So, "let’s focus on some new findings that remind us of the continued efficacy of digital advertising, and of ways to use it even more effectively."
Viewability is essentially akin to network clearance data in TV-- an assurance that the ads make it onto the screens that advertisers have paid for. Nothing more, nothing less. Just because impressions not qualifying as viewable still have a measurable impact, that doesn't mean anyone is arguing for advertisers to optimize against half an ad on screen for thee quarters of a second. It's just GOOD NEWS that even fractional exposures still can contribute to campaign performance. We do, after all, say that performance of ads on screen more fully and for longer durations, exceeds that of ads classified as non-viewable to a sufficient and consistent extent to justify the thresholds in place (which, of course, have NOTHING to do with when an impression makes the best impact, and everything to do with when a publisher should be paid.)
George Parker, your comment was flippant and offensive. Send me your email address and I'll send you the papers. Of COURSE the impact I'm talking about is measured in sales.
There's no "last touch' attribution in any of this stuff. I'm not a fan of last touch at all.
I'm sorry now I that wrote this piece without having the papers available-- but I thought it was far too innocuous a post to create a stir. Again, I would have linked to these papers if they'd been online. They will be publicly available soon at the PDRF's site, and my company will probably post them at our site as well. Steven did excellent and rigorous work; don't judge it by my petty little article, read the papers.
“In all cases, viewable ads were substantially more effective than non-viewable ads, supporting both the definition of viewability (for display) and viewability measurement.”
I'd like to know how to quantify what "substantially more means" in this context and by what standards are the viewability for display and measurement are from, the IAB?