asiCast Special Edition III: What do we mean by cross-platform reach, engagement and targeting?
This is the third asiCast in our series looking at what we mean by the terms we use when addressing issues of vital significance to all involved in the media and communications industry. This time, our Research Director Richard Marks discusses with our group of industry experts definitions of the terms we use in relation to advertising. There will be a fourth asiCast to follow this which will relate specifically to mobile advertising. Here we look at what we mean by cross-platform reach, engagement and targeting.
This is a transcript of the discussion originally published on 7th January 2019. You can listen again here.
Richard speaks to Tess Alps, Julien Rosanvallon, Nigel Walley, Brian Jacobs, Tim Elkington, Josh Chasin, Eija Moisala and Alex North.
Research Director, asi, UK
Having already tackled the thorny issues of what is digital and what is television, in this third special edition asiCast our panel of industry experts and thought leaders will be focusing on three main areas related to our theme of definitions specifically related to advertising. We are going to be talking about cross-platform reach, engagement and targeting.
So, firstly, lets discuss cross-platform reach. It’s important to get our definitions right. By cross-platform, we mean measuring video across platforms and devices as opposed to cross-media measurement of video, audio and text. To measure media across different platforms, will we need a re-definition of reach criteria?
Chief Research Officer, comScore, USA
As far as cross-platform reach is concerned, the first thought that comes to mind is that we have combined television and print into multi-platform campaign metrics for a long time and magazine ads or newspaper ads and TV ads are certainly different experiences.
One of the things that is happening here in the US is that the Media Research Council (MRC) is looking into a requirement for duration weighting in GRP calculations across platforms. What they generally mean is not that you should weight by how long the ad is or how long the engagement is, but that you should weight by what percentage of the impression was complete – so that watching five seconds of a five second ad is a full exposure and catching five seconds of a fifteen second ad is a third of an exposure.
My perspective is that what we should do as an industry is move away from average minute or commercial minute to average second. Now a question arises – is our measurement technology sufficiently granular and robust enough to capture ‘second-specific’ exposure? Ideally, we would solve all this by measuring exposure in terms of average second. So, if I am talking about a thirty-second TV commercial versus a six-second pre-roll, let’s say, I will have six seconds of exposure for the pre-roll and thirty seconds of exposure, assuming the consumer sat through the whole spot, for the TV spot. Time spent is a lovely metric – advertising ultimately is like physics, it’s all about space and time, so time spent is universal.
Head of Solutions & Partner Development, Facebook, UK
Exposure, as a concept and the opportunity for value, is consistent across all platforms. The definitions of exposure are probably going to be different and I think we need to be okay with that but we need to be aligned on the fact that they might be different. Exposure to a TV ad might be different to exposure to an ad on a digital platform for example. But, in the way that a TV ad can still be valuable if you’re not watching the screen because you are listening to the audio, a mobile ad might be valuable if you’re not actually listening to any audio. A lot of the mobile consumption is with people with their headphones on or when they are reading subtitles on something because they are on a train and they don’t want the audio to be there.
And so, it has value in some form. You need to know how much of that behaviour is happening and then what the actual subsequent value is of that exposure. But it all falls within the broader bucket of exposure. A lot of those might not be defined as viewing in the conventional form but they are the opportunities for value and I think that is what is consistent across the platforms and what we strive for.
So, if we agree that not all reach is the same, how do we make comparisons across platforms?
Chief Digital Officer, Internet Advertising Bureau, UK
I think getting that wider agreement even within media and then across media is very difficult. So, say for example, we are just not comparing apples with apples, so you are basically into a conversation that says, ‘how many pears equal how many apples?’. I think this maybe a parallel that doesn’t work, but we eat all sorts of different foods, but we know what the calorie count is across all of those foods. For example, we know that a burger equals two chocolate bars or something like that, so whether you can say – and I honestly don’t know – if how many YouTube ads equal how many TV ads, because it probably depends on so many variables in each case, but again, that is not to say that it’s not something that we should stop working towards. Imagine if we did have that, imagine how complicated the advertiser’s life is now, but how much easier it would be if we came to some sort of understanding.
One thing the industry does seem to agree on is that measuring exposure – time spent with media – is the bedrock for media planning and trading. However, it is frequently claimed that the currencies need to move beyond that – in a crowded multi-tasking media world they need to provide a definitive measure of attention or even actual engagement.
What does engagement mean, and can it realistically be measured in an agreed, standard way? Is that achievable, or even desirable?
Tim Elkington: I suppose in some ways, if you’re thinking about reach, you’ve got a basic figure that no one can really debate so its actual reach – how many people did that ad reach? But then, I suppose, you have also underneath that, got a sort of ‘quality of reach’ which is maybe a different thing altogether. I think we are right to say all time isn’t equal. I’m not thinking about particular media at all here, but if I am sitting in a room where something is passively happening in the background for a minute, does that equal me being really engaged and leaning into something for a minute? Probably not. You have to think that the thing you are actively engaged with is more valuable than the thing you are passively absorbing.
Having said that, other people would argue exactly the opposite way around and say that there is a sort of non-conscious engagement and you don’t have to be focused on a particular message to absorb it. Whatever you do, there are going to be winners and losers and the losers just aren’t going to like it and they are going to say its not credible so, it’s a difficult problem to solve.
Josh Chasin: Now we have the opportunity to see if programme engagements translate into advertising engagement. I have always been able to make that argument both ways – ‘I’m so highly engaged in the programme that the ad comes on and I’m riveted’, or ‘I am so highly engaged in the programme that the ad comes on and now I can unplug for thirty seconds.’
These are things that we can measure and start to make informed decisions. I’ve seen some research that suggested that if the feel of the ad matches the feel of the show it resides in, that is one user experience, whereas if the ad is experienced as jarring against the programme content, that is another experience (and) that is more likely to push the consumer out of the ad. So, I don’t know necessarily that engagement is going to turn out to be a binary where ‘yes’ is good, and ‘no’ is bad. I think we are going to find out that the way that people cognitively experience content will inform the best ways to put advertising in front of them.
Founder and CEO, BJ&A, UK
For as long as I have been around in this business, which was almost before these things were invented, people have always argued about and discussed attention levels and engagement levels, however one defines those things. I think again, we get slightly hung up on these things because I believe that we are straying from a buying currency or a buying metric into something which I believe is much more to do with a planning type piece of data. I think that attention and ‘eye on the screen’ type work is great, and that can help you decide on all sorts of things about, for example, how you construct the message (and) indeed the context within which you place the message. These are sort of planning type considerations rather than currency type considerations.
I think quite often we make the mistake when it comes to JIC type conversations, and we’ve got an absolutely brilliant measure in BARB; we have got a brilliant thing there that has been proven over time. It is obviously highly reliable and all those things so, just adding more and more things to it and trying to bend it into a shape that answers every single question via one study is unwise. So, I think that advertisers find anything to do with attention/engagement, whatever you want to call it, ‘eyes on the screen’, extremely valuable. But I believe that is more to do with a planning part of the equation rather than the currency part of the equation.
Julien Rosanvallon of French measurement experts Médiamétrie, doubts whether a common measure engagement could be agreed:
Senior Vice-President – Television & Online, Médiamétrie, France
For the advertising industry there are big debates around how to take into account engagement and factor this engagement to demonstrate effectiveness. I think that when you enter this debate it is important to stress the fact that we believe that although engagement metrics, or efficiency metrics, are very important it is very difficult to define such a metric as comparable metrics for a whole industry.
Why? Because very often efficiency metrics are related to a given advertiser, so it is difficult to just have a standard, global definition. As for audience, it is much easier to answer the following question: who has seen what and when? This question is difficult, but it is much easier to answer than the question of the efficiency because the definition of efficiency will differ a lot between two advertisers.’
Tim Elkington: Quite often I would go to a conference and I would hear ‘the IAB definition of viewability is rubbish. How can an ad be effective if fifty percent of it has only been viewed for one second?’ The thing about viewability is not to confuse it with effectiveness.
What we are saying if an ad is viewable is that there is an opportunity to see the ad. We are not saying that it’s effective. Effectiveness is something else entirely. You would then need to know what the objective of a campaign is. If it is to increase brand awareness you would then do a brand study, but people seem to be hooked on this thing. ‘No wonder digital advertising is no good because they count it as effective if half of it is seen for a second.’ The thing about that is I think it is relatively easy to understand, there are some nuances when it comes to viewability that make it a little bit complicated, but I think that basic concept that actually viewability is an opportunity to see, is relatively straightforward. I do sometimes wonder if people are misunderstanding that deliberately to discredit digital.
Josh Chasin: On the one hand advertisers obviously need to understand the quality of the experience that the consumer is having with their creative because advertisers need to decide if the creative is working or if it needs to be changed. But another thing that we are using this data for is to negotiate payment and at what point has the media vehicle done its job? So, the way I often think about this is that it is the media vehicle’s job to bring the horse to water, but it’s the advertiser’s job to make them drink.
So, if I am a TV broadcaster and you run commercials on my network and the audience tunes out of the commercial – let’s think about a digital context where a viewable ad is fifty percent of the pixels on screen for two seconds or more – so if the audience tunes out of the ad before an impression is made, that is bad for the advertiser but does the media vehicle still deserve to get paid? And I think that that’s a fulcrum point when we think about measurement. When does the media vehicle get paid? – and I am bullishly pro media vehicle! – so, when the media vehicle provides that audience, they should get paid. So, when does the media vehicle get paid and then how effective was the ad for the advertiser? There are times when you have to make sure the media vehicle gets paid because they have done their job even if the creative hasn’t.
In the discussions around new forms of advertising and particularly ‘ad tech’ three words tend to crop up, not always used in the right way or even to mean precisely the same thing. Is there agreement about what we mean by ‘targeting’, by ‘addressable’ and by ‘programmatic’?
Chair, Thinkbox, UK
Obviously the word ‘targeting’ has been used forever; to try and focus your messages to certain people in certain ways and you’ve always been able to target by time of day or by channel when it comes to television. The newer words in my mind are ‘addressable’ and ‘personalised’ and I do think those are different.
For ‘addressability’ I think we would see that as being something that is sent to a particular address, a particular TV set, and therefore is generally to a shared screen. So, it is not personalised, you don’t know that a single person is going to view but you’re making a bet that a particular family might shop at Tesco’s more than they do at Sainsbury’s, so you want to send them a different message because they are not customers of the other retailer. So that goes to that house because you know that home is, say, a Barclays Bank customer or not.
‘Personalised’ is a very different kettle of fish and that tends to be to a personal device, and I know not all desktops are personal, they are sometimes shared but, increasingly, obviously tablets and mobile phones are personal devices so, to ‘personalise’ a message is very different to an addressable message.
Chairman, Decipher, UK
Are the phrases ‘addressable TV’ and ‘targeted TV’ or just addressable and targeted interchangeable? There is implication in addressable that you have information on who you are addressing, whereas with ‘targeted’ the implication has always been that you can target very widely, you can target whole populations. There has always been a kind of specificity between those two. Addressable TV is, in a way, a sub-set, I think, of targeting and we old TV hands would want to describe TV as a targeted medium, it has always been targeted but we just used segments and groups and time of day methods for targeting. The addressable brings data into the mix and so that probably is the distinction between the two. I most certainly wouldn’t think they were interchangeable.
Josh Chasin: When I think of ‘addressable’, what I’m thinking about is a context where I can serve or provide an impression to an individual consumer so, I would argue that addressable was the opposite of broadcast.
The word ‘programmatic’ was being used at conferences arguably way before there was agreement on what it actually meant or what most of the audience understood it to mean. Are we now any clearer?
Tess Alps: ‘Programmatic’ as a word has got a lot of connotations to it but all it means is a computerised, automated way of doing something and that seems very sensible. Why not automate processes if they can be made more efficient, less onerous, for people? Unfortunately, what it conveys to the market is something that is probably low quality, low demand and bottom-feeding online advertiser opportunities that basically have only got value where its bundled together with something else – something that you wouldn’t buy as a standalone piece of media inventory, but you might buy it programmatically because you’re not basically considering it in its own terms.
Josh Chasin: It’s a dangerous phrase because everybody knows generally what they are talking about, but I don’t think everybody has the same definition. I tend to think of programmatic as automated, very simple right? But you can step out of the media business and think about programmatic as a way to manage airline reservations, for example. It is just the automation or the programming of a function.
‘Programmatic’ and ‘addressability’ go together because when I want to distribute impressions one at a time to specific persons or devices with tags on them that tell me that the user of that device has a certain attribute, perhaps auto intender, the only way to do that is programmatically, right? I mean, the only way to do that is to have a user of content who is about to be presented an impression. There needs to be a handshake where the engine that puts the ad in front of that user has to, in a millisecond, look at the information about the user, assess it, figure out who wants to buy the user based on that information, quickly sell that impression to that user. The only way you can do that is ‘programmatically’. So, I guess I think of programmatic as a platform for executing addressability which ought to employ targeting.
Richard Marks: I do think that we have clarity around what we mean by targeted, addressable and programmatic. As ever, we just need to be careful that we use them correctly. To sum up what our experts told us, nearly all advertising tries to be targeted. Addressability is a very specific means of doing that, whilst programmatic is a means of automating the process.
When it comes to cross-platform reach it is clear to me that measurement is going to have to get more granular in order for us to truly understand the relative value of exposure across different platforms. Minute-level granularity just isn’t enough. There is agreement about time-based measures as being the best definition of exposure but not around the minimum levels needed to be effective.
When it comes to measures of attention and engagement, I would argue that these should be left to advertisers and their agencies to decide and implement. They may have a place in the planning and formulation of advertising, but they will never become an agreed part of industry trading currencies.
Although the focus of this podcast has rightly been on advertising, it is also important to remember that a growing proportion of TV income does not come from advertising. At an extreme in the UK, advertising accounts for just thirty percent of UK television revenue, according to Ofcom. So, when people complain about how currencies should be measuring outcomes and not head counting, remember that audience measurement serves a wide range of needs, including understanding content and viewers. It’s not just about selling cars.
Tess mentioned the importance of understanding the distinction between addressable advertising sent to a set top box or a television and personalised advertising which implies a specific person will see it and, as we’ll discuss in the next edition, it’s in the world of mobile that arguably personalised advertising is the more viable proposition. In the meantime, thanks again to all the participants in this edition of the asiCast.