Lumen at the Insight Show

Come and hear us talk about getting an ‘unfair share of attention’ at the Insight Show in Olympia next Wednesday at 11.10am on the Showcase stage – it’s part of the broader Marketing Week Live event.

We’ll also be exhibiting at the show, and demonstrating our new POS testing tools and mobile eye tracking. Come and say hi!

Join Me_INS-Speaker-Banners-MSG4-728x90.gif

Location, location, location

Finnpanel Conference.jpg

We were in Helsinki last month, at the guests of Finnpanel, talking to media planners and buyers about the reality of attention to online video. There was a very big screen.

Size matters. As we have noted before, big posters get more attention than small posters. Big billboard banner ads get more attention than MPUs. Video ads on big desktop screens get more attention that video ads on little mobile screens.

But size isn’t everything. What we find time and again is that attention to advertising is a function of attention to the surrounding content. The more you engage with the content, the more you engage with the advertising. This is why an ad placed next to some high quality journalism that you savour slowly will get more attention than when it is placed next to train times that you glance at quickly.

And the best performing ads are placed as close as possible to the most engaging content. Below is a meta-analysis of over 200,000 page impressions we have captured on our UK-based desktop panel since 2016, showing where attention tends to go on a typical web-page.

chart email.png

You can see that the heat is often at the top of the page, and follows the content down the centre left of the page.

Which begs the question, where would you put your ads? In line with the content, where they are likely to get noticed, or to the right-hand side, where they are much more likely to get ignored?

As Kirsty and Phil will tell you, it’s all about location, location, location.

Secrets of the skip button

Secrets of the skip button

Digital video delivers lots of attention, but is it the right kind of attention? If you force people to watch an ad, won’t they just stare at the skip button, poised to strike as soon as the time is up?

We recently did an analysis of over 1000 YouTube ads that we had collected from our panel to investigate this issue.

First, the bad news, or at least, the expected news. People do look at the skip button. You are not alone. Everyone does it. The chart below shows the percentage chance of a person looking at different areas of the screen, and there a big red blob where the skip button sits.

skip button viewing.png

There’s also a smaller blob of heat in the bottom left of the screen. This corresponds with people looking at the time left in the ad, counting down the seconds until they can see their cat videos.

But that’s not the whole story. Sure, people look at the skip button, but they don’t hate-watch it. The time spent looking at the skip button is brief. The second chart, below, shows the time spent looking in different areas. This reveals that time spent looking at the skip button is short: the majority of the attention goes to the centre of the screen.

skip button dwell.png

What does this mean for advertisers? BRANDED SKIP BUTTONS. Simple as. Someone has to do this. Please, for the love of God, someone has to do this.

Hard Pivot

Hard Pivot

Thank you to everyone who has noticed our new venture: an over fifties dating app. Yes, it’s a hard pivot for an attention technology company, but hey, you have to do what you love.

This isn’t the only brand extension that we have launched recently. Even I was unaware that we appear to have opened a retreat centre in Euston and a restaurant in Paris.

lumen paris.jpg

There are also ambitious plans for us to launch an SI unit to measure light. Watch this space!

The power of pre-roll


We were at the asi conference in Athens last week talking about differences in attention to advertising across media channels. Regular readers will know, of course, that not all impressions are created equal: big ads get more attention than small ads; print display ads often get more attention than digital display ads; video ads have a better chance of being noticed than static ads; and so on. These attention differentials have a knock-on effect for both long term brand building – as Ebiquity have also found – and short term sales.

But the audience in Athens was interested in TV and OLV. How does pre-roll compare to the other media we track?

The short answer is that it rocks.

Below, you can see the ‘attention curve’ for ads in different channels. The data in the subsequent charts either comes from our UK desktop panel, or aggregated results from bespoke tests we have conducted for OOH, Print and Mobile over the last five years. It shows the percentage of people who notice an ad at all, and the percentage of people who look at ads for at least 1 second, 2 seconds, 3 seconds and so on. Two things to note: this data is cumulative, so people could glance at an ad three times for half a second each time, and we would record the aggregate attention to the ad as 1.5 seconds. And we have applied the same viewability standard (50% of the ad, for 1 second or more) across all the media, so that we can compare like with like.


What it shows is that pre-roll ads, either on publisher sites or on platforms such as YouTube, get a boatload of attention. If a pre-roll ad is viewable, then it is almost guaranteed to be viewed, which is highly unusual. And it is likely to get viewed for much longer than other media and formats that we track. Roughly 40% of viewable print ads get 1 second or more attention; roughly 80% of viewable pre-roll ads get at least a second, and over half of them get over 4 seconds of attention.  By way of comparison, only about 6% of desktop digital display ads get a second or more attention.

This superior performance is driven by a mix of factors. Creative plays a part: advertisers try harder to make pre-roll advertising interesting and engaging, whereas display ads are often churned out willy-nilly. If more effort went into creative development of display ads, then the results might be better. Until then, junk in, junk out.

But there are two other factors that drive the results.

Firstly, size matters. As we have said before, bigger ads get more attention than smaller ads. Pre-roll ads just take up more pixels than almost any other form of advertising we track, and, crucially, take up a far greater proportion of the screen – often 100% of the screen. There’s simply nowhere else to look. Of course, some people still find a way to avoid the ads: they look away from the screen or open up a new tab or look at the comments below the video. But it’s less of a problem than you might think. It seems that we have to make more of an active effort to look away from a pre-roll ad than you do from other formats, and while few people actively want to watch the ads, far fewer can be arsed to actively avoid them.

Which brings us to the second reason that pre-roll ads do so well: we are forced to watch them. It seems that if the content is worth it, we’ll sit through quite a substantial portion of an ad to get to it. Below, you can see the percentage of people who are looking at each second of a pre-roll ad as it plays. This means that it’s different from the aggregate attention curve that is shown above. It’s also worth noting that this data is exclusively from desktop computers, and not from mobile, which has different attention patterns.


What this shows is that at any given time, about 60% of viewers are actively engaging with pre-roll ads – and that this is just as true for 20” non-skippable ads as for 6” non-skippable ads. If people are forced to watch the ads, then watch them they will.

Of course, if they are not forced to watch them, they won’t. If you look at the line for the skippable ads, then that declines quickly over time. But the drop off is less precipitous than you might expect. 60% of viewers stick with the ads for 5 seconds or so, which is the sort of result that we only ever see for double page spreads in Print. This may be down to the fact that it may take people a bit of time to find the skip button on desktop – and modern ‘ecstasy of fumbling’. Or it could be down to the fact that pre-roll creative is, relatively speaking, quite good, and people don’t mind giving the brands a few seconds of their time. Creative is really important.

A quick side note on the skip button: yes, everyone looks there, but not for very long. We did an analysis of over 1000 desktop pre-roll ads YouTube ads recently, to create an aggregate heat map of where attention goes in general. Below, you can see a visualisation of the percentage chance that different areas of the screen will get look at at all:


Check out the mad heat in the bottom right hand corner – which is where the skip button is. Pretty much everyone looks there. And there’s some secondary heat in the bottom left hand corner, where people are obviously looking at the clock to check how long they have to endure the ad for.

But that’s not the whole story. Sure, people check the skip button, but no one is staring at it rather than the action of the ad. The second chart visualises the proportion of time spent looking at the different areas:


As you can see, by far the majority of the attention of the ad goes to the centre of the screen – on the action of the ad. People don’t hate-watch the skip button: they glance at it, and then watch the ad until it’s over. They may prefer not to watch an ad, but hey, what are you going to do? You’re getting free football highlights or cat videos once it’s done.

What does this mean for advertisers and publishers? Firstly, that that they should treat pre-roll very differently from display. Pre-roll is what you might call a ‘sequential’ media, where ads follow content which follows ads, whereas print, OOH and most digital formats are fundamentally ‘simultaneous’ media, where the ads and the content compete for attention at the same time. Given that we can only focus on one thing at a time, and that ads are usually less interesting than the content they accompany, ‘simultaneous’ media are at an inherent disadvantage to ‘sequential’ media. In response, ‘simultaneous’ media owners, such as newsbrands or social media platforms, should strive to simplify their sites to reduce clutter and increase standout of their ads. The Guardian’s commitment to ‘fewer, better ads’, served in prominent locations, makes it far closer to a ‘sequential’ media owner than many of its competitors. Facebook’s in-line ad formats take up much of the prime real estate of the desktop screen, and consequently are more ‘sequential’ in nature. This is part of the reason that desktop ads on Facebook get 50% more attention than the norm. 

Secondly, we should re-evaluate skippable ad time lengths. There’s a move in advertising towards 6” video ads across both TV and digital at the moment. In their place, I am sure that they are quite useful. At the same conference in Athens, the Yan Liu of TVision gave a fascinating and compelling presentation about their utility on TV. But Yan was quick to point out that while 6” ads get looked at and can be well remembered, they are best used to trigger established memories, rather than building them in the first place.

If you are in the business of making memories, however, you may need more than 6” to get the job done. And it seems that pre-roll advertising may be actually quite a good way of delivering these longer messages. Many consumers are prepared to stick around for 20” to get to the content that they want to watch. This may seem like a long time, but it’s less than the 2-3 minutes we have to endure on terrestrial TV in the UK (and 4-5 minutes you might have sit through in the States). Further work will have to be done to calibrate the optimal timelength in this ‘attention exchange’, but at first look, pre-roll looks like an ‘attention bargain’ for the advertiser, the publisher and, crucially, the viewer.

Initiative’s contributor panel

Manny BHM.jpg

In the closing weeks of Black History Month, one of Lumen’s Attention Consultants, Emmanuel Agu, was invited by IPG Mediabrands (via Initiative), to speak to a room full of creatives, eager to expand their understanding, and interrogate the representation of black culture in the media. Here’s his account of the event.

I was grateful to be joined by such wonderful thinkers Adébayo Bolaji, Damola Oladapo , and Nelson Abbey, co-author of upcoming novel ‘Think Like a White Man’. To initiate the evening of discussion we were posed with one key question:

“in the internet age, do people have enough influence to make change on the representation of black people in the media, or is it only politics that can make this change?”, steered by our wonderful host, Mimi Okorie.

We began by casting our minds back to press coverage in the wake of the London riots following the death of Mark Duggan. There was complete unanimity on the awful media discourse of this time; papers and news reels were littered with racialised depictions of “opportunistic thugs” embroiled in gang warfare, as opposed to those illustrating disenfranchised and angry youth. Though we had conceded some progress had been made since days past, it feels as if expecting anything more than the stereotypical facsimiles of thugs, gangs and social disobedience from an industry dominated by the white and structurally privileged,  is an unrealistic goal.  

We turned our eyes instead to brilliant community owned projects such as gal-dem and media diversified, providing spaces for exploration of talent for Black (and minority) people, as well as pushing back against tired narratives dominating the zeitgeist.

Similarly, whilst we lament(ed) the need for such channels, we celebrated the power of #blacklivesmatter hashtags across social media, drawing a spotlight on injustices faced by our communities, whilst managing to somehow lower the socioeconomic bar of entry required to contribute to these discussions. Yet still, we worried about the accessibility of such platforms – who’s voices dominate these discussions? What measures are put in place to defend those from the attacks of trolls? Why do we see the plight of Americans’ used to diminish those in the UK? What are the long-term effects for the marginalized, waking up day after day to seeing black and brown bodies brutalized by the state again and again?

Next, we turned our eyes to focus on creative and advertising strategy. As we navigated this discussion,  we brought case studies and examples, citing the fallout from such campaigns like Heineken, and perhaps “tone deaf” ad placements from H&M. next we contrasted the consequences and success of Pepsi’s ad featuring Kendall Jenner and Nike’s ad featuring Kaepernick. We were led to question the racial composition of the production teams in the first three instances – could such creative really be produced if black people were present in these spaces? And if they are, do they feel empowered enough to speak up in their workplaces?

Whilst the pursuit of justice and equality in all that we do is admirable, we acknowledged that we still navigate a capitalist society – profit is king. Nike’s ad showcases the potential to beyond the ordinary, disrupting the status quo in a political manner, whilst still remaining entirely profitable.   

The question posed at the start of discussion will, of course, be debated for many more black history months to come, but, I feel as if the fundamental answer is clear. Progress cannot happen if there are simply no black people in these spaces. for some larger corporations, perhaps positive discrimination programs might be seen as a stop-gap measure, but ensuring those we take into the industry are well supported in equal professional discourses, whilst guaranteeing them agency to assemble and network with peers – we will be sure to continue to head in the right direction.

The future of Facebook is Lumen (according to Mark Ritson)

festival of marketing Lumen

The star of this year’s Festival of Marketing, like the star of last year’s Festival of Marketing, was Professor Mark Ritson. His conversation with Steve Hatch of Facebook was as entertaining and sweary as we have come to expect. One particular highlight was this exchange (as reported by Marketing Week):

"Keeping with the effectiveness theme, Hatch moved the conversation onto viewability and attention, asking for Ritson’s point of view on the future and how best to measure it. The answer to that, it turns out, is Lumen.

Too right. Totally agree. Mark went on:

[Lumen] measurements…found that news media garners the best attention, yet no one seems to know about this data because the news media “doesn’t have a f-ing clue what they’re doing in promoting their advertising.”

He's right to highlight how good ads on newsbrands are - and wrong about how good the newsbrands are at promoting themselves. More and more advertisers are coming round to the view that newsbrands are great at gaining and holding attention, and their inventory is undervalued at the moment. So they must be pretty good at getting the story out. 

But the message we want to get out today is this:  Facebook is f-ing good too.

A brief look at our most recent data for Facebook shows that it is an excellent way of getting people to look at ads. All the subsequent data comes from Lumen’s UK-based desktop eye tracking panel, analysis of 176,000 viewable display ad impressions and 23,000 Facebook in-feed ad impressions, 2016-present.

On average, 30% of viewable in-feed ads get looked at, compared to around 20% for display ads in general – this is very good, though as Mark would point out, not as good as newsbrands. Viewable ads on The Times, for example, have a 54% chance of actually getting looked at.

graph viewable seen.png

Where Facebook distinguishes itself is in how long people look at the ads. Whereas most display ads get viewed for around 1.3 seconds, Facebook ads get an average of 2.1 seconds of attention on average. This is an important threshold to get over: we have noticed that people are far more likely to remember a brand if they look at an ad for more than 2 seconds.

graph average dwell.png

What is driving this success? Normally, we would say that attention to advertising is a function of attention to editorial: the longer you stay on a page and can see an ad, the more likely you are to notice an ad. But that isn’t the case for Facebook, where viewable times are significantly less than the display norm. Ads are usually available to be seen for around 20 seconds on most sites we track; on Facebook it’s just 8 seconds. So ads on Facebook have less time than half the time of most display ads to get noticed, but generate nearly twice as much attention.

viewable time graph.png

The answer probably has more to with three separate by complementary factors: video, targeting and site layout.

Since Facebook’s famous ‘pivot to video’, not only the content but the accompanying advertising has become dominated by video.

I don’t know if this is such a good thing for content, but certainly works for ads. People spend far more time engaging with video advertising than static advertising. Not engaging like a Cinema or TV ad, but more than static banners. On average, if a video display ad gets looked at, it will get looked at for 2.5 seconds, almost double the time length for a static ad. Not a great deal of time, but how much do you really need to say about your catfood to remind people about the TV ad they recently saw for the said catfood? And crucially, this time length breaks the 2 second barrier, meaning that if you have branded your ads properly, people should remember the brand of the ad.

graph multiple.png

(Note: ‘player instream’ data excludes YouTube. We have this data, but do you expect us to give all our data away for free?)

The second ingredient is targeting. People are far more likely to notice and spend time with ads that are relevant to them than ads that are not. Catfood advertising is more interesting to cat owners than the rest of us, and Facebook is very good at knowing whether or not you have a cat. At Lumen we do lots of tests for lots of clients to establish how important this targeting is. The effects differ from category to category, and the results are, of course, proprietary to our clients so I can’t share them here, but believe me: they’re big. So, while Bob Hoffman may lament the pernicious effects of hypertargeting on our creative industries and democratic institutions, he’s wrong to say that it doesn’t work. It’s the fact that it does work that makes it a moral and ethical choice whether or not to use or regulate these technologies.

Finally, the key to Facebook’s success may lie in design. Most of the ads on Facebook are in-feed. On Facebook, the reader’s eye doesn’t have to move from what they are really interested in (the content) to notice the ads. The ads are interspersed at regular intervals and sized to be of equal importance to the content. This also means that big enough to be seen.

If you look below, you will see a typical (anonymised) user session on Facebook. Sure, there’s some attention to the ads in the top right hand corner of the page, but it’s easy to see that most of the attention goes to the feed – and that’s where Facebook puts most of the ads, which gives them the best chance of being noticed. And that’s how you make the big bucks.

facebook hm.png

 So what should we all learn from this:

  • Advertisers should understand that Facebook is a jolly good place to advertise, but not totally sui generis. It’s just another website, albeit a very good one for advertisers.

  • Publishers should learn a thing or two from Facebook about how to layout a page and how to give advertising some space to breathe.

  • And Facebook should understand that it doesn’t matter how many impressions you serve – or even how many viewable impressions – but how much attention you create for advertisers that really drives business success. Viewability is table stakes. Attention is the real deal. 

The Scooby Difference


Advertisers spend a lot of money creating brand characters or licencing the rights to use them in ads. But how much is a character worth? And how much should you be paying to have a celeb in your ads?

This was one of the questions that Ipsos set themselves to answer in launching their new Ipsos:Connect digital pre-testing tool. It helps brands optimise their creative to boost attention, recall and sales. And yes, it now includes Lumen eye tracking as standard.  

To help publicise the new product, we have been doing some digging into the extra attention that ads with well known characters receive.  

Take this example for the Halifax. We would expect a DMPU for a finance brand like this to get around 2 seconds attention. In the tests we ran, it got 7.5 seconds attention, almost 4 times the norm.

Scooby 2.png

The difference is down to Scooby-Doo. The character captures people’s attention quickly and easily, triggering people’s memories of the now-famous TV campaign.

Scooby does the business for Halifax too. The character directs the viewer’s gaze to the logo and key product points in the ad, meaning that the ‘emotional’ side of the ad is directly contributing to the ‘rational’ take out. This isn’t just borrowed interest: this is hardworking advertising.

Advertising go BANG!


Do ‘big’ ads amplify the effect of ‘little’ ads? We were speaking at Ad:Tech Londonwith our friends from Inskin on Wednesday to answer exactly this question.  

Inskin make highly engaging digital ad formats that really get noticed. But does the effect stop there, or does it reverberate throughout the rest of the media plan? We conducted an experiment to understand the effect of starting your campaign with a Big Bang (such as an Inskin ‘Pageskin’), and then following it up with more standard ordnance, such as MPUs.

The results are impressive. Not only do you get a lot of bang for your buck with Pageskins (about 15 times more aggregate attention than standard formats) but you also get a multiplier effect on attention to the subsequent MPUs. People were 39% more likely to notice MPUs if they had been preceded by a Pageskin, and 140% more likely to engage with the subsequent ads for more than a second. You can read about the effects in more detail here.

People are more likely to notice a second ad if they have actually looked at the first one. Digital media planners need to take a leaf out of the TV planners’ playbook, where campaigns are often commenced with a barrage of 30 or 60 second ads, and then followed up with 10 or 15 second cut downs, that are designed trigger memories of the initial ‘blockbuster’ execution. Going large with high impact ads doesn’t just create an impression initially but multiplies the effects of subsequent ads.

Just remember, once you have made it go bang, don’t look back.

Want to find out more?

The Weak and Strong in South Africa

We were invited by Spark Media to speak to an audience of 300 media types in South Africa this week, sharing the billing with Dr Virginia Beal, of the Ehrenburg Bass Institute, which is led by the famous Byron Sharp. I met Dr Beal the night before the conference: she had just come back from safari with our hosts and had come up close with a pride of lions.


The next day, Dr Beal was up first, and gave a compelling talk about how advertising really works. Many of us who work in the industry think that our job is to persuade people to buy a product. We want to get in front of our customers and really convince them why our cat food, say, is better than the other guy’s cat food. This is called the ‘Strong’ model of advertising.

This is also nonsense.

In reality, she said, most advertising works as a simple reminder of a brand’s existence. The aim is to build mental availability or ‘top of mind’ awareness, especially amongst infrequent buyers. In comparison with the ‘Strong’ model of advertising, this is has been called the ‘Weak’ model of advertising. But that makes it sound a bit weedy, so now people called it the ‘Salience’ model of advertising.

She went on to show that the Weak model is much stronger than the Strong model, across categories and countries.

Then Lumen got up on stage. And much to everyone’s relief, our data supported her argument neatly.

One of the reasons that the Persuasive model of advertising must be wrong is that people rarely hang around long enough to be persuaded. The average dwell time with advertising is vanishingly short: around 2 seconds for print, mobile and OOH ads, and less than a second for digital display ads. With this sort of ‘attention budget’ it’s very hard to communicate the precise details of an offer or give people a compelling ‘reason to believe’ in why your kitty dinner is better than the other guy’s.

What you can do, however, is remind people that you exist, and trigger an implicit emotional response. And this is pretty much exactly what happens. Most advertising gets looked at rather than read or deeply engaged with. No one really cares that ‘eight out of ten cats prefer Whiskas’, but everyone knows that ‘cats like Felix like Felix’ – and why simple, emotive advertising helped Felix become the number 1 cat food in Britain.

Want to find out more?