| |

Attention: a case of 'less is more'? 

Andy Brown: Is attention a case of ‘less is more’? 

Optimising for attention can create better outcomes for media owners and advertisers. But changing attitudes over audience numbers means challenging the status quo.

For attention metrics to succeed, we have to overcome the first rule of audience measurement.

Very early in my media research career, I was taken to one side and reminded by my boss that I had forgotten this rule. It says that even though a particular methodology may be more accurate and provide more granularity, there is one overarching principle that every researcher will appreciate.

That is: “Big numbers are better than small”.

Why TV has been slower to adopt ‘attention’ 

It might sound contrary, but in practice this means that ‘less is more’ when it comes to attention measurement.

I know this because, over the last three years, I have become more involved in the use and application of attention metrics via my role with The Attention Council.  For attention metrics to hold sway across media then we have to overcome this basic rule of audience measurement.

Let’s see why, starting with video.

Whilst we have seen rapid adoption of attention metrics in the digital domain, there has seen a somewhat slower adoption in TV.  When discussing this with TV executives, the objections tend to range from: “we can’t control for creative quality” (I am not sure you can totally with a GRP/Impression count), through to: “attention cannot be factored into an annual agency deal” (perhaps a more valid complaint in a market like the UK).

Although one of the key reasons the buy side is using attention metrics is because they see a strong and consistent correlation with outcomes, we need to be mindful of specific concerns over attention and audience measurement. 

Namely, it appears to risk breaking the first rule.

The big change in incorporating ‘attention’ into audience measurement is that we move from a gross opportunity to see (OTS) to a net OTS. Route, the UK outdoor currency, already does this; Route nets its audience to take account of the visibility of billboards and other sites.

In a TV context, that means we are going beyond presence in the room, to eyes on screen.

In principle this means we reduce the broad quantum of the audience. Although our friends at Thinkbox will show you an ethnographic study conducted by Gorilla in the Room that shows that commercial messages can be consumed and retained by consumers not watching the screen.* 

The question from an audience measurement perspective is whether, in order to create a common planning metric, we should then focus on visual attentive seconds, which is the metric that the majority of users are aligning around for video applications.

I have yet to see an audience measurement system that does not involve some form of compromise. 

Challenging the status quo

If we look at the adoption of attention metrics across the sell-side of the video market (TV and premium video), we see significantly higher levels of adoption on the “digital” side of the business when compared to “TV”.

What are the reasons for this? 

1. The trading system in the UK market

I recently took a call from a leading TV network, who had been approached by three agency holdcos, all wanting to do their annual contract negotiation on attention metrics. The question he put to me was quite simple: how can I manage a deal on that basis as there is no continuous independent measurement benchmark** that would allow a neutral evaluation of a deal based on attention?

It is difficult at this point in time to see a UK agency TV deal being transacted entirely on attention (and price). 

2. Inertia

We should not underestimate the power of the existing modus operandi. It is arguable that TV trading (due to the limits of contracts rights renewal) has not really moved on in 20 years in the UK.  

3. Inventory supply

In the digital domain, it has become a truism to say there is an excess of advertising inventory and of highly disparate quality. This has enabled significant changes to the measurement ecosystem, such as the introduction of verification services like Moat and IAS.

For most premium publishers, they see themselves as being the “wheat separated from chaff” and have upwardly adjusted their CPTs to reflect this. The wider usage of attention metrics in the digital market has seen a further premiumisation of the market.   

However, feedback from the US market would suggest that the recent, softer upfront left the major networks having to work a little harder to sell their less-demanded inventory. This, in turn, stimulated a greater level of investment in attention measurement than previously seen. 

With the growth in streaming options from both traditional broadcasters and newer players, will we see attention being used as a common denominator to evaluate what might be low audience channels with more highly engaged audiences? 


So how have attention metrics changed the industry (albeit more significantly in digital media)? 

It boils down to this: we have seen a further separation of premium publisher’s inventory from poorer quality offers.

These publishers, in many cases, have adapted their ad formats, decluttered their websites, and, in several cases, reduced their ad load. They see this as a virtuous circle, serving fewer ads in more attention-grabbing formats (at a higher CPT), creating a more effective solution for advertisers across the effectiveness funnel. Some digital publishers are now selling attention-guaranteed inventory solutions.

We’re seeing the first rule of attention measurement being challenged. Because, almost by definition, those solutions are being achieved with lower audience counts.

In other words, less can truly be more! 

Andy Brown is CEO of The Attention Council  

*My (only slightly) tongue-in-cheek response to this study: as an advertiser, why I am paying for an audio-visual cost per thousand and getting an audio delivery? To be fair to Thinkbox, and the UK TV industry in general, there is undoubtedly some impact of audio branding.

** TVision runs a 300-home continuous panel reporting attention metrics in the UK. At this   juncture, it is in a pilot phase and would be unlikely to be robust enough for trading. 

Denise Turner, Chief Executive, Route Research, on 22 Nov 2023
“Loving this debate, just catching up on all the comments, thanks Andy, Matt and Ali. Since Route was mentioned in the piece, I feel I should say something, but I'll make it brief! A point of clarification and an observation: Point of clarification - Matt, you say that attention measurement was originally designed for the online world. Actually, I must respectfully disagree, attention measurement in the form of visibility, whether people can actually see a poster or a screen, has been baked into the Out of Home currency in this country, since Postar was launched in 1996, long before online advertising became a thing. And the team at the time used expert academics to guide and inform the process and conduct the research. That continues to this day with Route which was the evolution of Postar. A personal observation after years of working across media channels and running effectiveness teams - we need to be careful about conflating the counting of audience with the outcomes. At Route we deliver a realistic count of exposure to advertising, and by whom. Effectiveness is about overlaying other data sources, and the interaction with other channels to determine whether people did anything as a result.”
Dr Ali Goode, Cognitive Scientist, Gorilla in the Room, on 21 Nov 2023
“Hi Andy Just to add to Matt's comments (though of course I do not speak for Thinkbox but as I led the research, I feel the need to clarify what the research found). Firstly, I do feel obliged to point out it was not an ethnographic study, it was a fully controlled empirical study that was conducted in collaboration with a professor who is a world-renowned expert in attention, having studied it for 20 years. Also, all of what was reported was underpinned by statistical testing to produce robust results. (We did use some film for a study conducted for ITV from 2019-2020 to illustrate the findings). Secondly the key point was that we demonstrated, something that has long been accepted academic psychology community, that attention is a mental resource that is shared around the senses and is not linked to one specific sense. If you deplete the mental resource using one sense, there is less available for the other senses. Talking on a mobile phone is banned when driving as talking depletes the attention left for visual tasks needed to drive safely. We were able to show this with advertising also, if we depleted attention through and auditory task, visual attention got worse and visa versa. Thirdly we showed that auditory attention was more resilient to distraction than visual attention. So with AV ads, auditory content will both communicate in its own right but also mediate where visual attention is directed. Again it is generally accepted that you don't look around for something to listen to, your gaze is attracted by what you hear. Finally we also showed the more the attentional resource was depleted the less that was available for ad processing. A low attention task such as listening to music reduced the attention available for ad processing less than a high attention task such as talking. This was exactly in line with Load Theory the most widely accepted theory on attention. Because of this, hypothesis was that ads with an auditory content do need to be considered differently to those that are just visual alone. My personal view is that low attention processing needs to be revisited, however, it does need to be re-defined as Robert Heath’s ideas generally don’t work.”
Andy Brown, CEO, Attention Council, on 20 Nov 2023
“Hi Matt. I agree that TV is a winner when it comes to attention. However, as I asked at the recent ASI event, what happens when the inventory is delivered and the needle does not move on the outcomes? Attention may be a way of telling the client their "baby is ugly. In some ways the point of this article was to share that for any change in tv sales and buying, there may be a necessary need to accept a lower audience count.”
Matt Hill, Research and Planning Director, Thinkbox, on 20 Nov 2023
“Hi Andy, Interesting article and thoughts as always, couple of headline thoughts from me below, followed by a bit of detail for those interested. Attention isn’t a concern for TV advertising in the same way it is for online display advertising. In any attention study, TV outperforms most if not all others. Attention is not just visual (as you note in your article). Audio attention is vital, and this is partly why TV does so well. It’s true that the TV industry has been cautious with attention metrics – it’s also understandable. Attention measurement was originally designed for the online world whose existing metrics (impressions) are a very poor proxy for effectiveness. Attention is another, better proxy. This is not an issue for TV – ask any econometrician: TV ratings highly correlate with effectiveness. On the tongue-in-cheek – the nerve! – question about why an advertiser might pay a TV CPT (i.e. for audio-visual) when people might only hear the ad: the attention mix you get with TV – audio-only; part visual/part audio; visual + audio – is already baked into its CPT. If TV was always watched with full visual attention then that would likely reflect in its effectiveness and the price would be higher than it is. On the issue of creative quality – with an exposure-based currency the advertiser is paying for the opportunity to gain attention (and so incentivised to do this task well); with a visual attention-based currency, the advertiser is paying for eyes-on-screen. So – through no fault of their delivery and the context they offer – the media owner is out of pocket if the ad itself just isn’t very good at gaining visual attention. Weirdly, this could inadvertently be more cost-incentivising for an advertiser: if they’re only charged for ads that are looked at, the cost-effective thing to do would be to make an ad that isn’t looked at (so it’s delivery is free or very cheap) but still has an audio effect. Unlikely, but we should bear unintended consequences in mind. Attention is a fascinating area for TV, and one where it does really well – especially if you factor in audio. But the TV companies are all busy working on building outcome-based measurement tools for advertisers. Rather than using a proxy measure of effectiveness (like attention) to optimise planning, they’re looking at how to optimise against actually achieving a specific goal.”

Media Jobs