What you need to know:

• Cinema ads score 80 per cent active attention, 20 per cent passive attention and therefore zero wastage of ad budget, according to early data from Amplified Intelligence. 
• That’s a higher level of active attention than the firm found in its recent study for Nine (72 per cent active attention for BVOD ads watched on mobile, 37 per cent for linear TV ads) and the highest overall attention score it has recorded globally, according to CEO Professor Karen Nelson-Field.
• Val Morgan MD Guy Burbage said the data quantifies what the company has long known – cinema captures more undivided attention than other media, and therefore cannot be compared like-for-like on blunt metrics when planning media buys.
• Media agencies say brands getting to grips with attention, but that reach and frequency remain mainstays of planning. Though some – such as OMD and Hatched – are starting to reorder the hierarchy of media plans based on attention data.
• Hatched's Andrew Pascoe says one financial services brand notched a 12 per cent active attention gain without spending more or losing reach by reallocating spend between channels.
• NAB Executive, Operations, Planning & Partnerships, Tom Dobson says appetite increasing at the bank for attention metrics and use of attention within media mix modelling. Also keen to understand whether contextual ads drive any increases in attention, or whether it could put in generic 'home loan' ads to AFL sponsorships and earn the same attention.
• Karen Nelson-Field pushes back on critics and says no point in sprinting for attention CPM, because it will just be "gamed". Far more data required to displace other currencies.

Seeing, but not believing?

“About 75 per cent of all online ads that you pay for against MRC [Media Ratings Council] standards are actually not getting any human attention at all,” according to Professor Karen Nelson-Field, CEO and co-founder at Amplified Intelligence.

She says there is a role for two-second nudges and lower funnel performance ads with lower attention values. But they must at least register with the viewer or marketers are wasting their money. “Without attention, ads have no impact. It is as simple as that.”

There is a vast difference across what is considered ‘viewable’, per ad industry standards, and what is “viewable with attention” across different formats and channels, says Nelson-Field.

The flow-on effect of such “inequitable impressions” is that “any concept or methodology or system that relies on equitable impressions will fail.” Which means key marketing tools such as share of voice (SOV) modelling based on impressions “is going to fail you.”

Screen test: zero wastage

The findings confirm what Val Morgan already knew, says MD Guy Burbidge.

“It’s common sense … but we haven’t had a business like Karen’s measure it in an independent way with great technology and deliver those consistent results that we need to talk about it in market,” says Burbidge.

Scoring 80 per cent active attention – where people are directly watching the ad – and the entire remainder ‘passively’ paying attention, may seem high. But Burbidge thinks the numbers “pass the smell test”.

“Having zero wastage makes a lot of sense. But the biggest impact for us is having a more equitable start line of analysis of what is delivering and driving real value on comms plans,” says Burbidge.

“Right now we’re measured by reach and frequency and cost – CPMs. That doesn’t tell the full picture, and we have consistently lost in that area.”

But he sees attitudes starting to shift as attention data is being integrated into agency planning tools.

“As that happens, we’re seeing the thresholds where cinema becomes effective after driving reach on other channels drop from a million or million and a half dollars to half a million dollars,” says Burbidge.

“So we’re already getting greater access to client plans by more of them planning with a filter of attention rather than planning just to cost efficient reach.”

Both holdcos and independent media buying agencies agree the pendulum may be starting to swing back.

Attention vs. reach & frequency; SOV fails

Some brands are struggling to reconcile a shift away from pure reach and frequency (R&F), says Andrew Pascoe, Head of Planning at Melbourne-based Hatched, with others trying to grasp whether attention is a replacement metric or a filter. But Pascoe says clients – including a financial services brand – have notched wins by acting on attention-based planning tools.

“We’ve been able to say ‘here’s a media mix with R&F of x, but if you retool the mix and reallocate the spend … you get about 12 per cent more active attention for the same spend and the same reach mix’,” says Pascoe. “They said, ‘of course we’d do that’. So that was a sign-up.”

Showing clients an “attention-adjusted share of voice” is another way to land attention metrics, says Pascoe.  “We’ll pull out some of standard simple [post-campaign] data and rebalance it so we get an attention-adjusted share of voice.  Suddenly, clients, especially the heavy screen spenders, go ‘Holy Shit’ I’m not where I thought I was. I’m not a 30 per cent share of voice, I’m a 15 per cent share of attention … sales are flat-lining.’ And then they are interested.”

Penny Shell, Chief Product Officer at OMD, agrees brands “have an appetite” to better understand attention planning, but stresses “reach and frequency … are still an important consideration” in planning. However, she says attention is “changing the hierarch any balance of channels” if not the overall mix.

Tom Dobson, Executive, Operations, Planning & Partnerships at NAB, echoed those views – marketers are keen to better understand attention-based planning, but have to use the right tool for the right job.

“There’s an appetite [for attention] and there’s also a sense of denial, because while we all know there is wastage, we also know that direct response is required, so we have skewed our spend to digital over the years,” says Dobson.

NAB’s latest Marvel-inspired brand campaign leans heavily on the reach and frequency delivered by TV, cinema and out of home. “It lends itself to cinema and we really wanted reach and awareness. But it is hard [to push harder into attention],” says Dobson.

“A pure attention metric would be amazing,” he adds, but “different campaigns have a different nature of attention required”.

That said, Dobson sees a disconnect in the level of researching, testing and tweaks that take place before an ad is launched “and then [once it is in market], you don’t really measure the level of attention being paid to it”.

Attention feeding MMM?

Both NAB’s Tom Dobson and Hatched’s Andrew Pascoe hope attention metrics can shed more light on how media investment has moved the needle.

“I would be overselling it if we said we were there [now] but I want to get to use that attention adjusted activity data as an input to some of the econometric modelling [we do],” says Pascoe. “We’re a bit wary of shifting them there – just doing econometrics is a big enough job in itself for clients – but if I can find the right way to do it I would like to be able to build a separate model [in parallel] using the attention metrics to see if there is any correlation.”

Dobson says he would “love a consistent metric” to use “within the media mix modelling that we do… an attention variable … to be able to adjust for [attention] would be really useful”.

Contextual, creative versus channel

NAB’s Tom Dobson challenged the likes of Amplified Intelligence to come up with data to prove whether or not contextual advertising yields more attention.

“I want to know if the attention for an AFL ad is going to be higher because you are watching it. Or could we just have put some [standard] home loads ads in and got the same attention?” asks Dobson. “If you’re spending half a million dollars making an ad, do I really have to tailor it to the environment?”

Val Morgan’s Guy Burbidge is looking to crack the other side of that equation.

“Something we’re looking into for the next phase, as we get more committed around the data collection, is the idea of genre. So do bigger titles, more cultural media moments garner more attention because they're more lean-in, they're more anticipated, they're more talked about?” says Burbidge, likening the cinematic release of Spiderman or Top Gun to an ARL or NFL Grand Final on TV.

“We already know that those really big moments in cinema deliver better results. But we want to overlay an attention quality layer that shows clients that these cultural media moments are an absolute must buy. How you treat a TV special is where cinema, in our perspective, should be in that ranking,” he adds.

“If 80 [per cent active attention] is the benchmark, I think you might see a bit of fluctuation on either side [within different genres] because of the nature of those cultural moments.”

While cinema shows highest active attention that lasts across the full length of the ad, Nelson-Field says channel, not creative, is king.

“The creative plays a role, but it’s not the principle,” she says. “If you put the same ad on a cinema screen and on the general web, you’re not going to get the same amount of attention [on the web].

“There are always outliers, and interestingly they are often movie trailers. But not to the point where you will get 10 times more attention.”

Rush for attention CPMs ‘misguided’

Rushing attention CPMs will backfire, warns Nelson-Field, with buyers defaulting to minimum costs brands ending up with what they paid for.

“CPM is a dirty variable in the sense that you pay what you want to pay from bidding perspective. You might be paying too much, you might be paying too little, but it's got nothing to do with the output of the impression,” says Nelson-Field.

“I think it's too soon to be looking at an attention CPM, because you’re just going to optimise to the lowest attention average – basically looking at CPM divided by [the number of] attention seconds. So you can actually game the system, because if the CPM is low enough, you can still buy an ad that's two seconds of attention, and it looks like it's a good price to pay, but it doesn't really represent at all the quality piece.”

Will attention ultimately become a traded currency?

Attention does not create memorability

Some of Nelson-Field’s detractors would equally argue that catching the eye does not equate to creating memorability and mental availability.

She agrees “one hundred per cent, and we have said this from day one.” But she suggests there are many imperfect metrics.

“I get criticised on this. Be we went into this trying to solve a currency problem… You can actually buy against viewability, and half the time, people aren’t actually looking.”

“So we went into this to think how can we build attention in a scalable way? I have said all along to the neuroscientists out there that this is not a cognitive measurement, but you can't build a scalable currency with a sample of ten,” says Nelson-Field.

“We don't understand what people are thinking and I would hate to think we could, but what we can see is systematic patterns. So when you've got enough data and you look at it at individual level, you can start to see how this thing called attention, active and passive, is related to outcomes,” she adds.

“How people feel and think? Not at all. But we do see systematic patterns, and that's what we're building product towards – to be able to capitalise on that for the advertisers.”

The making of Top Gun: How Amplified Intelligence built cinema attention rating

For the big screen, Amplified Intelligence effectively scaled up the tech and methods it has used to work out attention paid to smaller screens.

It found facial recognition cameras that could cope with low lighting, worked out where they should be placed, then built a basic model and started to collect more data so that the model could “train itself, because it is true AI”, says Nelson-Field.

For the next phase, it built a “friends and family” panel consisting of “roughly 80 people” across three movies in Sydney in March, which forms the basis of cinema’s early attention rating.

The model can “pick up a person’s face and it will make a prediction based on the trajectory of their face against the cameras and the distance from where they are sitting as to whether they are looking at the ads or not,” says Nelson-Field.

While early days and effectively a ‘beta’, Nelson-Field states the study is “the first in the world” to take that approach.

“We've built and quantified a phase one or a baseline, which is completely verified against all the other data that we collect. We still have the capability of ‘active’, ‘passive’ and ‘non-attention’. So we're pretty excited that we'll be able to use cinema data in the product that we build over time as we get more sample.”