Nearly Two-Thirds Of Non-Direct Inventory Deemed Non-Viewable
- by Tyler Loechner @mp_tyler, November 18, 2014
The digital ad industry’s
viewability woes just keep getting bigger.
Data from the third quarter of 2014 finds little more than a third (36.7%) of all display ads purchased on networks and ad exchanges were deemed “viewable” per the Media Rating Council (MRC) standard of 50% of the ad being in-view for at least one second. That’s down from the 45.3% rate posted in the second quarter, which was down from the 51.3% rate posted in the first quarter.
A silver lining is hard to find in the report, which comes from Integral Ad Science.
Viewability rates plummeted if the standard was extended from one second to five seconds, as only 21.6% of ads were in-view for more than five continuous seconds. Similarly, only 12.5% of ads were in-view for more than 15 seconds. All of these figures are down from the previous quarter.
Integral also reports that 13.7% of display ads traded on networks and exchanges in the third quarter were fraudulent, which was a major reason that networks and exchanges received a 540 “true advertising quality” (TRAQ) score, which is based on a mix of fraud, viewability, ad clutter, brand safety and professionalism ratings.
The TRAQ score has also been on a downward spiral this year, with the 540 score (out of 1000) below the second quarter's mark of 576 -- which was below the first-quarter mark of 578, a figure that was itself down from the fourth-quarter 2013 mark of 597. Integral notes that 17.4% of all display ad inventory traded via networks and exchanges last quarter was considered “moderate to very high risk” inventory in the brand safety department.
The latest Integral Ad Science Media Quality Report also includes data on video advertising.
“We’re excited to include video in our latest Media Quality Report and will continue to provide quality benchmarks for use across the industry,” stated Scott Knoll, CEO of Integral.
Over one in five (20.7%) of video ads bought on networks and exchanges in the third quarter were deemed “moderate to very high risk” for brands, per the report. The entire industry -- including exchanges, networks and direct-with-publisher deals -- combined for a 30% video viewability rate on the quarter, per the MRC standard for video viewability, which requires the ad to be in-view for two continuous seconds. Of video ads that were completed, only 20% were viewable throughout their duration.
If the stats, cited above, for video ads are correct, this is a huge problem that must be dealt with by digital ad sellers. Video is, by far, the most appealing ad format for branding advertisers because it allows them to use TV, or TV-like, commercials. But, if only 20% of the digital placements run in their entirety ----with a typical ad being either 15 or 30 seconds long-----imagine what that does to the cost per viewer figures for digital compared to TV, all of whose ads run in their entirety.
Just because a TV ad runs in entirety, doesn't mean it is viewed in entirety.
A TV ad may not be viewed in its entirety but at least it is entirely viewable.
I agree that TV really suffers the same viewability issue, though due to lack of equivalent measurement in TV, it's hard to quantify...your TV spot loads (~"ad impression" online); the viewer walks away to grab a snack, or switches channels momentarily, or minimizes the PIP, or skips the ad (~"not viewable" online). In either scenario - on or offline, you pay for the "impression". Wouldn't it be swell to pay for TV spots if you had some guarantee of viewability?
Keith, I agree that it's unrealistic to assume that Nielsen can actually measure whether a person "views" a commercial---even though it reports commercial minute ratings. I have long maintained that significant percentages of such "audiences" leave the room, become distracted, etc. and probably do not see an average commercial. Despite that, Nielsen does provide a measurement that tells the advertiser how many TV sets were tuned to the channel when his spot appeared, which is a pretty close approximation of a "page view". Also, unlike digital, TV spots really run in their entirety. So the opportunity to see them and get a complete message----not bits of one, or none at all-----is far greater than what digital has to offer---assuming that all of these reports about viewability, fraud, etc. are accurate.
Ed, totally agree. There are counter balancing things though. For example, someone may have viewed TV in a public place and seen an ad. That is currently outside the definition of "TV viewing". I am sure the same applies to much internet delivered content and ads. But regarding "out of the room", we do co-incidental studies here in Australia. Basically we telephone the panellist's home and ask who was watching TV, in which room, and what they were watching and on which channel immediately prior to the phone call. This is then checked back against the metered viewing statements. I am not a liberty to divulge the levels but they are pleasingly high - much closer to 100% than you would ever expect. In essence the panellists are doing what they are supposed to the gross majority of the time, meaning that the ad-avoidance factor is at least being accounted for, and as far as we can tell, pretty close to what it should be.
John, the out-of-room type of avoidance is only part of the equation. Indeed, it is often over emphasized by respondents who tend to over claim the number of times they leave the room while "watching". By far the largest form of avoidance is cessation of attention, which is almost impossible to measure without some form of bias causing observational methodology--a camera, for example. What I've seen in numerous studies and report on in our annual, "TV Dimensions", is that leaving the room accounts for a per commercial audience loss of about 5-10% while inattention, often involving some other activity, runs to about 35-45%.
Fair points Ed. The thing is when "out of room" (or more correctly out-of-range) it is impossible to have any attention. The thing with inattention is that most studies are based on claimed inattention which I believe, just like out-of-room, is over-reported. But I agree that inattention is definitely a larger quantum than out-of-room. But, and it is a but with little robust empirical evidence, I think you may be under-estimating the power of the human brain to both store, process and recall memories both rapidly and off a very small (but personally relevant) sub-set of ad content. I refer to, and I assume you have read, Robert Heath's work on low involvement processing. In one article he hypothesised that the average number of 'triggers' for an ad was five - things that worked well were music and jingles, but also the obvious things like colours, characters/talent. But one of the keys is (ironically) the sound. Someone can be pay little visual attention, hear the trigger sound and (I) either choose to watch the ad again, (b) recall the ad, brand, product etc and choose to go back to what they were doing (c) recall the ad and reject both the ad and the product. So, while inattention may be high, a good ad with lots of good triggers can still work acceptably high even for inattentive viewers. Further, those triggers can work even without the ad. Now this is showing my age, but if I hear The Flower Duet from Lakme, or Malcolm McLaren's take on it 'Aria on Air', I immediately think of BA's 1989 "Face" ad. I am sure there are many, many more examples that we can all think of.
Agreed, John, but I'm referring to an average situation, to keep things simple. There are many ways to trigger a response to an ad, such as "image transfer", by which a sound or musical signature, for example, "reminds" the viewer of previous exposures to an ad campaign, in some cases, cutting across media----like a radio message stimulating recall of a TV commercial. Even within TV , a simple visual cue, even if that's all that is seen, can have the same effect. There's also the reaction of a consumer the first time he sees a commercial for a new campaign relative to the second, then a third exposure and, over and above that, the length of time between exposures. Also, we have high and low interest products and/or ad campaigns to create still more nuances. But I'm really trying to lump all of these situations together into an "average" exposure context, primarily so the data is more easily understood by those who are not familiar with the research and all of the ramifications.
Ultimately, the most important measure of advertising is whether the desired consumer response was obtained (sales lift, greater brand awareness/engagement, etc.) After reading this article and the well-articulated comments that follow, it's easy to see why many companies are increasingly focused on advertising outcomes and ROI.
Of course Mike that is the end game. Metrics such as brand awareness are generally a 'soft' measure (at least for established brands) while 'ad engagement' is really an input to the marketing mix (i.e. I spent $X for Y contacts with Z 'engagement' ... what happened as a result). But the thing is that NO medium is in a position to report back to an advertiser the ROI of a campaign, because the days of a single medium campaign are long gone! And even back in the days when you could be 100% TV or magazines, there were still other marketing levers (such as price, distribution, packaging, promotions etc.) So who IS in a place to do this? Media agencies have a lot (not all) of the requisite data and if a client provides their other marketing metrics they are in a good place to perform the econometric analysis as well as to make any recommendations as to changes in the strategy or implementation. Third party specialist analytics companies can also perform such tasks but are they more data specialists or media and marketing specialists. To me, the IDEAL is for the advertiser themselves to employ some smart young graduates and to do the hard yards themselves rather than lean on the individual media or to rely on their media agency , or to outsource it to a data agency. In the latter cases what happens if have an account review and change agencies - all that knowledge is lost.
John, I don't think we disagree. But media and marketing firms could get much closer to understanding cause/effect than they are now. And to say "NO" media can measure revenue outcomes would be an overstatement. We work with a few folks who do just that. I was just saying that we should focus less on "access to an audience" and more on measurable activation of that audience. That's where the industry is headed. Great discussion!
Mike I agree that said firms can get closer to the fiscal outcomes of marketing and advertising campaigns. The problem is that when a media owner within a medium does so, and then tries to quantify their contribution to the brand's success, in my experience the results are always off the scale. Let's face it advertising is a 'weak' force with respect to sales (not so much with respect to awareness, preference, brand image etc.). For example, distribution and price dwarf it. In a good econometric model paid advertising may just hit double figures regarding its contribution to changes in sales levels. Yet I see things like a TV network claiming they were responsible for 70% of the sales growth. Cough, cough. My point around 'no medium' - and I actually meant to say 'no media owner' - was that they don't have the totality of the marketing and advertising campaign inputs. You then run into the grey area between correlation and causality. Yes, your business may have a 70% correlation with sales movements but that does not mean you were responsible for 70% of the incremental sales. I had an example in NZ where the leading ice cream manufacturer was extremely concerned that there media spend showed good correlation to their sales - which made them happy. But they were extremely concerned that the #2 brand (and growing) was also correlated ... how could we stop their competitor getting a 'free ride'. A quick look showed they hadn't factored in weather (temperature and rain were the important drivers). A deeper look revealed that distribution of their products into smaller shops and dairies was even more critical than the weather which makes sense as it is an impulse purchase. My point is that no media owner that I know of has both the breadth and depth of product data to do such analysis. The media agency has a pretty large chunk of it. The brand has it all.
Great comments so far. A few thoughts. 1) Viewability, especially for video and especially for brand advertisers is simply not optional. 2) The "yeah, but people ignore TV" argument is silly and unworthy of the digital media business. People ignore ALL media, but the baseline requirement is that it is possible to SEE it. 3) We can throw all the numbers that we like at attribution, but it will never be perfect. For a start, success will always have a million fathers, and executives will always have biases about what really works and what does not. Additionally, we should take care not to mistake the map for the territory: just because we have spreadsheets and attribution models does not mean we have the truth. Lastly, even if we had a flawless understanding of the past (e.g. we knew precisely what worked and did not work last quarter) it is not necessarily predictive of the future. It is in our nature to hunger for crystal balls and alchemy, but we will never really have either. At best, all we will really have is a better set of guesses about the truth.