As we enter into a period of sunny Spring weather and Covid restrictions have all but disappeared, it’s worth reflecting on what will become the ‘new normal’ in media this year, with many things seeming the same as before the pandemic.

But some things will be different. This year, there is much talk about the novel concept of ‘attention guarantees’: promises from broadcasters to deliver not only defined numbers of impressions, but minimum amounts of consumer attention (as defined by attention suppliers like TVision).

This is not the first time that an ‘attention guarantee’ has been mooted: AB InBev piloted the approach with some success last year with the A+E network.

And there is much to recommend it, specially for a famously creative advertiser such as Budweiser. TVision’s data allows AB InBev to identify which TV ad slots are most likely to get a lot of attention, and which are most likely to be ignored.

Sharing this data with the networks enables a transparent dialogue about where the value in any individual media buy lies; reporting against this metric helps assess if the value has been delivered. If everyone knows what they are getting (and what they are giving) then you can partially de-risk the deal and give a better price for the delivery. You might even make a forward-looking contract to deliver a specified amount of attention on the basis of this data: an attention guarantee.

But AB InBev’s subtle and sophisticated use of the attention data may not be followed by the rest of the market that is now following their lead. We may rue the day when we first thought of ensuring – or even insuring – attention.

The concept of a guarantee rests on the assumption that attention is a commodity. And, to an extent, it is. Attention data providers such as TVision and my own firm, Lumen, have created a common currency of attention which is consistent, quantifiable and fungible across media.

This attention data is relatively regular and predictable. Big ads almost always get more attention than small ads; ads that are on screen for longer tend to get noticed more than ads that whizz straight past you; an ad on Site A will, nine times out of ten, get more attention than the same ad on Site B.

But these predictions are dependable only up to a point. Models based on media and contextual characteristics can explain a lot – but not everything.

The first fly in the ointment is that not all brands require the same attention levels to get the job done. There’s a big difference in the attentional requirements of a ‘Sales Ends Tuesday’ ad compared to a big brand ad.

A reminder message can get to the point much more quickly than a new product launch ad. A famous campaign, with well-established distinctive brand assets, does more communicating in less time than an ad for a start-up that no one had ever heard of. Treating attention as a standardised commodity risks collapsing these important distinctions.

But there is a second and perhaps more important reason to be wary of simplistic attention guarantees: the creative. Two different ads can both be put in the same spot, but if one of them is made by Colenso BBDO, one of them is going to get a lot more attention than the other. But the TV station or the publisher on the other side of your attention guarantee has no control over that. Yet they stand to lose a significant amount of money if the ads you send them are duds.

In English law, there is a concept that insurance claims can’t be made against suppliers when the goods they sell have an ‘inherent vice’. If you are shipping robust stuff like TVs and refrigerators around the world, you can insure them against loss or damage. But other items, like eggs, which are by their very nature fragile or breakable, are more complex to guarantee.

In the words of the UK Supreme Court, there is a demarcation line between “damage caused by an external fortuity and damage resulting solely from the intrinsic nature of the insured goods.”

Attention to advertising, it would seem to me, offers a good example of ‘inherent vice’ (and not just because of the similarities between many old school ad men and the film of the same name). You can take all the care in the world to buy the best media, but if you put a dog of an ad on air, you only have yourself to blame if no one watches it – or acts on it.

Of course, there might be an upside to this. Will we see broadcasters refusing to air terrible ads? Might they offer to take the bet only if they know the ads are good? In that context, it’s unsurprising that A+E were prepared to guarantee attention to the company who invented ‘Whassup?!’

But there could be downsides too: thresholds for ‘attention guarantees’ set so low as to be effectively riskless to the publisher – and effectively meaningless for the advertiser. The weaponisation of attention data to wring further discounts from media owners. The incentive for brands or agencies to develop or run attention-grabbing ads for attention’s sake, irrespective of their impact in market.

I think the insurance paradigm is probably the wrong way to think about attention data. Insurance is most appropriate when you are dealing with ‘risk’: when all the potential outcomes and their likelihood of occurrence is known to the decision-maker. Attention to advertising is, however, often ‘uncertain’: we even don’t know – and certainly can’t control – all the factors that influence the attention that ads receive.

In this sort of reality, I think it’s best to adopt what you might call ‘anti-fragile attention strategies’. Use the data you have to listen to and act in the world. Make lots of small bets and adopt a learning mindset to make sense of the results. Use attention data as a means to achieve your true ends, rather than making it an end in itself.

The more we learn about attention, the more we understand how much more there is to learn. Attention research is too young – and the world too complex – to reach for the actuarial tables yet.