Interview Transcript

Disclaimer: This interview is for informational purposes only and should not be relied upon as a basis for investment decisions. In Practise is an independent publisher and all opinions expressed by guests are solely their own opinions and do not reflect the opinion of In Practise.

Julian, can you briefly share an introduction to your background, please?

I’ve always worked for media agencies, across my career. I started in London, working on some of the big retail brands, such as Dixons Store Group, when they were the largest retailer in the UK. Then I moved into international media planning and buying. I started off as a planner and a buyer, working with lots of clients, taking in their briefs, doing the media plans for them and then actually going, negotiating and buying those media plans. I moved into the international department and then I started having some leadership roles in media agencies in Vietnam, in China and then came back to the UK to run some European accounts.

I then decided to have a quick look at the other side of the fence and spent three years as a marketing director, in a large start-up, in Abu Dhabi. I’ve been on the other side of the fence and seen how clients and the marketing teams work. Then I came back to working at a digital consultancy. More recently, I’ve spent over four years at Mediacom, as a global data strategist where, rather than looking at planning, specifically, or buying, it’s looking at how you can apply data across the planning, buying and measurement phases of an agency to, effectively, increase efficiency and effectiveness of media campaigns.

My current job is the global agency sector lead for media owners and agencies, at YouGov. I’ve now moved to the third part of the relationship; having done the agency and the client, I’m now the supply.

Can we just take a step back and actually walk through the planning phase of a typical campaign? Let’s say that I’m a big FMCG brand, with a new campaign. I want to market it on TV but also through my digital channels. What is the first planning stage?

It’s a good question because it will very much change by type of brand. If you are an FMCG you will have a very specific way of doing things and a specific set of goals and targets. If you’re a digital retailer or a luxury brand, you will do things in different ways. The planning has morphed into very specific sector-led approaches.

If you’re an FMCG brand, most FMCG brands that I’ve worked on in the past are tied into the bar and chart way to help brands grow. They are really focusing on sales and they’re focusing on reach. They’re looking at broad target audiences for their brands and they are looking at maximizing reach. Generally, that tends to lead them to the large reach channels, such as TV. It is a very different way of approaching it versus a luxury brand, which would be looking at specific audiences – very niche, small audiences – and looking at the best way you can engage with those audiences. Most FMCG brands think, how do I maximize my reach and that’s all I need to do.

In recent years, we’ve started saying, not all reach is equal because you can find reach channels where you can engage more with an audience. If you can find that emotional connection and engage with them, that’s going to drive more sales than just shouting at them.

What data do they use, typically, to build the audience in the planning phase?

Again, it very much depends. It depends on what sector you are in. If I am a retailer, with a large website that has huge amounts of traffic, then I’m going to plug a DMP (Data Management Platform) onto the top of that. I’m going to be looking at my customers coming in, looking at the customer journey across my website and where else they’ve been and I’m going to build audiences off that and go and target similar people, through lookalike modelling. If I’m a luxury clothing brand, I might have a large data set of customers. If I’ve got that data set of customers, I know that those are people who are engaging with my brand, I know they are people who buy my brand. I can start looking at them and say, right, how do I understand more about them and how do I go out and find those people and other people similar to them, and I can build audiences off those.

If you’re an FMCG brand, then it becomes a bit more difficult. If I’m Ariel washing powder and I have a website and people come to that website, that’s not really going to help me build an audience. Those aren’t the sort of people that you want to be going out and finding. Predominantly, you want to find housewives with kids; people who have washing machines and are washing their clothes. It’s a much broader audience and those are easier to find through general capabilities in the marketplace. A TV channel can easily find housewives with kids for you and so can digital channels, because it’s quite a generic audience. Actually, probably housewives with kids is outdated now; you probably just need to look at large families; people who are in large families and making the buying decisions. They’re easier to find than more niche audiences, for specific brands, that hold a specific space.

For the FMCG brand, how has the approach to building audiences changed?

I worked on a project, recently, which was pet food. Generally, we would be going out and looking for people with a pet. If it’s a cat food, people with cats; if it’s dog food, people with dogs. You would go out and try and find those audiences. They are relatively hard to find, accurately. Actually, we took a step back and started thinking, right, how do we go out and find people, specifically, with a dog or a cat? Is it possible for us to build a data set that can then allow us to target more effectively and target more efficiently? Can we bring that audience in, as a first-party data set? Would it be better than a third-party data set, a commercially available data set? A first-party data set is a data set that the brand, itself, owns. Could we bring that in, cost effectively and use that as a backbone of delivering our communications?

Actually, we found a few mechanics where we could go out and, at scale, bring in mobile ad IDs, effectively, at scale and hold those in a DMP and then use that as the route to market.

As first-party data?

Yes, as first-party data.

How did you go about finding those mobile IDs of potential pet owners?

It was through a mobile game where we, effectively, said to the people, play this game; here’s a discount voucher. Would you like to give us your mobile ad ID so that we can communicate with you in the future? It was an exchange of a discount voucher and a game – a bit of engagement – in exchange for permission to use and target them in the future. It proved effective at reasonably large scale.

This is brands that are trying to innovate in gathering more first-party data to amplify their audiences or build better segmentation of customers?

Yes and I think you can do that, in some instances. With pet owners, you are talking about relatively smaller and specific groups. It allowed us to know whether they were a cat or a dog owner. The next stage would have been to then understand how old their pet was, so you can start targeting them with food for a specific age; whether it’s puppies or more mature dogs. You’re getting into more niche audiences so you can stop wastage. There’s no point communicating pet food to someone who doesn’t own a pet.

When you are looking at something like chocolate, the amount of people in a market who don’t eat chocolate is insignificant in terms of trying to target them. To build an audience, of first-party data, around people who eat chocolate is never really going to be cost-effective. You have to look at the brand and the marketplace you need to understand when building an audience of first-party data, collecting it and then using that as a way of communicating, whether the collection outweighs the efficiencies you drive in your communications.

With these larger categories, such as chocolate or snacks, for the Mondelēz of the world, can they really innovate and keep improving the audience-building phase?

The way they’re doing it at the moment is working out ways to get rid of that inefficiency. So they are always looking for the cheapest reach possible. But a lot of these brands are now going into adjacent markets, so they will be looking at slightly different product extensions. For example, Snickers going into protein bars and, therefore, targeting people around gyms. The way that they are driving innovation is, really, to look at those adjacent markets that they are going into – which are more niche audiences – and looking at how they can build reach quickly amongst those niche audiences using, potentially, non-TV, is really the way it works out; so digital channels.

If we go through that process and, let’s say, I’m a brand and I’ve got a rich first-party data set; maybe I buy something off the shelf. Do I then feed it into my DMP?

It depends what format your first-party data is. This is where you need to do a full audit, understand exactly what you’ve got and understand whether you’ve got the right permissions in place, because GDPR has come in recently, in the last few years, and that’s really had an impact on how you hold data and what you can do with it. What sort of data have you got? Does it have the right permissions? Then you need to look at where you are storing it. A lot of brands that I’ve worked with in the past, because of historical reasons, will have multiple data sets, in different databases, around the world. Trying to combine those data sets into one single view of the customer can really be very difficult, very time-consuming and very expensive.

When a client looks at first-party data, they need to understand what they’ve got. I think they need to understand what format it is and format is really important. That can either be held as a form of personal, identifiable information, PII. With that, we are talking about email addresses, telephone numbers, mobile ad IDs. Actually, mobile ad IDs are PII under GDPR, but they are not like traditional email address, home address information. Those sort of data sets, you tend to put in a CDP, a customer data platform. They have their own routes to the market. They will be more about things like newsletters, postal letters and things like that. More what was traditionally called below the line activities. If you are storing your data in cookie, anonymous formats, you need a DMP and you go to market via a DMP.

That is how it was three years ago but everything is changing now. GDPR has come in. Are things like mobile ad IDs PII? How do you hold them? CDPs and DMPs are merging and becoming one system. Cookies are going, potentially, which is a whole other area that we won’t get into today; potentially they will go at the end of this year into next year, although we will do workarounds and other things. But that has been cut off as a solution. When I say cookies, I mean third-party cookies.

Just on that point, what do you think the impact of that could actually be on advertisers collecting data and really leveraging this data in digital advertising?

The impact of third-party cookies going? Other solutions will pop up; we’ll have workarounds. I don’t see, in the short term, it having a massive impact. It’s going to be inconvenient, but companies will work out workarounds, they will work out different solutions. In the market, we are already seeing solutions such as ID5; InfoSum coming in and providing a way of delivering audiences without third-party cookies. It’s just going to be a bit messy for a while.

How do you see the impact on companies like Facebook and Google?

Google are the ones that have brought this in, predominantly. Generally, Google, Facebook and Amazon are the walled gardens. You can put data in, but you can’t take data out. They have huge amounts of data sitting in that walled garden. They also have huge dominance. I think it’s about 80%, at the moment, of all ad digital spend, goes into the walled gardens.

It causes huge amounts of problems for advertisers; we’ll talk about this later, when we get into things like measurements, as well. The challenge is that you can build a really good audience in Google, you can execute it in Google and you can measure it in Google. But you can’t connect that to any other activity that you are doing outside of Google. The same applies to Facebook and Amazon. They’ve become their own ecosystems. As long as advertisers realize that is the case, they need to really build understanding in each of those platforms then they have the more complicated effort of trying to understand the overlap between Facebook, Google and Amazon, which is the more difficult thing to do. If you can understand that overlap, you can then understand how much activity you need to be pushing into each of those walled gardens.

Do you not think that IDFA and these new regulations coming in put Facebook, for example, at any risk?

Specifically with the Apple solution, my understanding is that it stops Facebook’s capability of targeting off their platform and off their assets, into their advertiser network. It’s going to hurt them, financially but, in terms of their capability of delivering advertising on their core platform, my understanding is that it doesn’t impact that. I might be wrong – I haven’t looked at it for a while – but that’s how I read it.

In terms of building audiences, I would have my first-party data set, maybe the mobile ad IDs that I collect, or my own website, supplemented by third-party. Then I would build an audience and, potentially, replicate that on the Googles or the Facebooks of the world, to build a lookalike audience in their walled ecosystem, to then target, via ads, on that platform?

You might need to make a strategic decision, early on. That strategic decision is, am I going to go into the open web or am I going to go into Google and Amazon? Facebook is slightly different because that’s social. Generally, a lot of clients will make a decision. Do I go on the Google assets or the Amazon assets or do I go into the open web, with DPSs like The Trade Desk?

Do they see it very differently, then? Do they see Facebook, normally, as a different category versus Google and Amazon?

Yes. Facebook is seen as paid social, so within an agency, it is separate to buying advertising in a DSP, in an open exchange, as it were. As a client, and as an agency of that client, they need to work out, with the assets they have and they integrations they have, whether they need to go fully just all in with Google or whether they need to use the open web. A lot of brands will make that decision. For me, I don’t think it’s worth going a bit Google and a bit open web. I think you just need to make that decision. A lot of brands will just go all in with the Google stack and, once you do that, if you’ve got a website, you can start pulling in data from the website. You can build really strong audiences using Google data. You can also onboard your first-party, so you can use Brand Match. Therefore, you can take your first-party data, put it into Google’s ecosystem and then use it to build audiences and, potentially, enrich it with other Google assets, such as Maps, Search and that sort of thing. That’s all starting to open up where, within Google’s ecosystem, you can use their other data sets and pull in, to enrich your data and to build more specific audiences.

The other solution is, you use the open web. I’m being quite black and white about it but I do think that’s the most sensible way to go. One of the jobs of a data strategist is to look at the assets that a client has; what audiences, understanding and data they have. Do you need to go and find other data sets to either enrich your understanding or to grow your audiences and then go and buy those? Have you got it in the right format so cookies, email, mobile ad IDs? That will dictate what channel you go and buy.

How have you seen brands changing the way they’ve approached that question of open web or Google only ecosystem, for example?

A lot of brands still haven’t really looked at that question. What I saw happen, over the last four or five years, is that brands started, specifically, at a brand level, moving contracts out of agencies, particularly for Google and Facebook, and doing their own direct deals. Particularly on Google, they are doing this because they were already having deals with all of the Google Analytics on their websites and, potentially, with a lot of them, the search was out of another specialist. They were using more than just the Google DSP – DV 360, that element – and they were using other elements.

Naturally, they started to look at single Google contracts. A lot of the clients pulled that out of the agencies, to do the contracts, but then the agencies still do the execution. But the contract is held up at the client level.

What did that mean for the actual commercial terms? The agency only deals with the working dollars, so it’s lower margin for the agency because they’re only buying or executing the marketing?

Quite often, they’re still rolled up to the agency contracts so it didn’t impact them in that way. But one of the really important things that I saw happening was that these were specific commitments that the client had made directly. When you got to a situation – say, for example, the pandemic – clients needed to stop spending or cut their budgets significantly and they cut those budgets from areas where they didn’t have specific commitments; but they had specific commitments to Google and Facebook, so you saw Google and Facebook’s revenue protected. That was one of the larger implications and you have to just say, clever Google and Facebook. They are able to do that by tying up these large deals. It was just an interesting move that has happened.

What does a deal look like, typically? You mentioned the commitment? So the brand would actually commit a certain amount of spend, per year, on that inventory?


No matter what? It’s a fixed commitment to spending X amount, per year?

The deals are relatively flexible so it depends on how you structure them. A lot of them will have a minimum spend commitment and then, with that minimum spend, it opens up a whole load of potential innovation, lower rates and if you spend more, then those rates can go down further. Then you get specific innovation, as well, and opportunity to pilot and do interesting things.

Previously then, the brands were almost just leaving it up to the agency and the agency would deal with Facebook and Google and execute, plan and build the audiences on the platform? But now, what part of the work is the brand actually bringing in-house?

It depends. For a lot of brands, there has been a recent move to bring the actual planning and buying on those platforms into brands; out of the agencies and into the brands. A while back, there was a report from the A&A, in America, that talked about transparency and how the agencies weren’t transparent. Off the back of that, a lot of brands decided to start pulling this capability – predominantly digital programmatic buying – in-house.

It does drive some efficiencies, early on. I’m not that close to it now, but my understanding is that a lot of brands have struggled in doing this and they’ve struggled for a number of reasons. One is that they are bringing in a lot of people to do programmatic buying, but there is no room for those people to grow and to be promoted. Because there is only a small group of people doing a very specific role, within a company, there is nowhere for them to be promoted, to grow and to move.

They are literally a team that execute the buying of the campaigns on DV 360 or Facebook’s platform?

Yes, so that causes an issue. The second issue is that, one of the great things about an agency is that you’ve got thousands of brands, under one roof and all are doing interesting stuff with lots of different suppliers. You can start to look and innovation is much more rapid within an agency because, potentially, you have got lots of different things happening and it moves much more quickly with the market, because you’ve got lots of different types of businesses, types of brands that they are working for, so they can see all facets of the market and how it is changing. You can see that when you are sat in isolation, in a small buying unit, at a client’s. That’s been a troublesome area for them. A lot of the clients have actually come back to the agencies and engaged with them for strategic advice and to help them with innovation.

What do you think the endgame is here, for advertisers, in this in-housing trend?

As with all things in the world, you move to that halfway between the two. I think that, certainly, the client brands would like to keep some control and understanding but I think they’d also like not to have large departments that they have to manage. I think it will move to a hybrid model.

That’s more the planning stage and having control of the data and knowing what the data is and managing that, then leaving the execution to the agency, maybe?

Yes. What I saw working well and the trend that a lot of people were leaning towards was the strategically important assets of businesses were held in the business. Predominantly, that would be the CRM systems, the data, the strategy around the data and the contracts with the technology vendors. Then the actual execution is done at an agency level.

How do you think agencies, like S4 Capital, the newer, smaller digital focused agencies, can compete with the larger agencies of record, given this in-housing trend?

They are able to do it. At the moment, everything is quite siloed. You’ve got a creative media agency, potentially, your direct agency and, in the big holding groups, those are very specific. In the media agency, you then have very different kinds of silos in there, such as search, programmatic, TV buying. Then you’ve got data agencies, as well. The holding groups hold all of those types of agencies. They are separate companies and they try and blend teams by doing a multi-functional team for a client but, at the end of the day, everyone still has their own P&Ls everyone, at the end of the day, still has their own allegiances to the company. At the end of the day, if they want to be promoted, within the company or organization, they need to stay loyal to that company.

What you’re seeing with S4 is a much more blended, smaller, nimble approach. It’s not quite going back full circle from the days when the advertising agency and the media department were all in the same company. It is going back to much more blended companies, that offer a whole range of services. You’re seeing big companies, like WPP, trying to respond by merging creative – such Wunderman and JWT merging, to Wunderman Thompson – and trying to respond to that.

You have to look at it slightly ironically, in that Martin Sorrell has done, with S4 Capital, what he was trying to do with WPP but because it’s such a large organization, it was structurally difficult to do. He can rebuild that thinking, with S4, with smaller companies that are more flexible.

S4 recently won the contract for the Mini but, but is it really on the creative side that they have an advantage, in creating thousands of pieces of content for digitally focused campaigns? Or do they actually have an advantage in executing and buying programmatically with their deep expertise?

I don’t necessarily see them having the advantage at the media agency level. The media agency level is still, very much, dictated by size. Particularly if you are a brand that is on TV, you need to go with an agency that has that volume, that can get you those cheaper.

What about on digital? Does scale matter, buying on Google and Facebook?

Yes, absolutely. It’s only the big companies, that spend over a certain amount, that do the direct deals with Google and Facebook. If you are a smaller company, that isn’t spending hundreds of millions globally, you will be doing it through the agency volume and you will need that to get cheaper rates. Also, the sophistication of the planning tools and the understanding is there, in the bigger agencies. It has become less so in digital, certainly, but there is still the volume argument. But certainly, if you are on TV, or outdoor, you need that volume.

Will Google and Facebook agree some kind of discounted rates, on a CPM level, or do they just give you a kickback, for the agency?

It’s on discounted rates. Again, not my area, but my understanding is that it’s discounted rates and then, if there are volume rebates, they are allocated back across the clients.

Which brands do you see moving most effectively in-house and why?

That’s a really difficult question for me to answer and the reason why is that I’m going to say that I’ve seen brands move in-house; I have read people talking about how brilliant it is and how much money they saved. I don’t believe them. I don’t see anyone who has done it absolutely fantastically. I had a close look at Heineken and I thought they were doing it in a really interesting way. Potentially, they could be getting back value from it. But at the moment, it’s very difficult to say because a lot of it is puff pieces. So far, no one has come out and said, it was a disaster, because their reputations are on the line, so people are sticking with it. There is no one where I think, oh my god, that’s really worked well for them.

Potentially, what worked well, when you looked at them?

When I looked at Heineken, what was really interesting was that they had bought their insights, their digital programmatic buying and their data strategy and data people all into one group, in the UK. That was a really interesting group that was sitting there, as one single unit. They were then using and pulling in a lot of help, still, from the agency strategically and buying in the other areas. It was a real hybrid, collaborative model. What I liked about it was how they internally organized.

Are there any categories, such as beverages or the beer companies, given the data and the access that they have with first-party and third-party, that are better placed to bring things in-house or to be more effective at that? Clearly, the digital players are best placed, given the first-party data they have. But with traditional advertisers, auto companies, FMCG, is there any category that sticks out to you that would have an advantage in moving in-house?

People who have got very large digital assets. One that I looked at, that I thought was interesting, was one of the telecom companies, because they had so much information and they were doing such complicated segmentation. Putting out thousands of messages, different messages, it made sense to hold that in-house and hold that very sophisticated operation about exactly where customers were in the lifecycle and what were the right deals to offer to them, being very specific. It made sense to hold all of that in-house. Then they were able to put those audiences through a DMP, lookalike model them and get them out as bigger audiences.

When I looked at that I thought, I can see why because it is incredibly sophisticated; some may say overly so. They had 3,000 different audiences in their DMP and you’re thinking, you might have gone slightly too much in one direction. But you can understand that, with that level of sophistication, you do need a very dedicated team that would probably benefit from sitting in-house.

How does YouGov help with the planning phase?

YouGov has a number of tools that can really open up planning for an agency. The first one is Profiles, which is a massive data set of media consumption; behavioral, attitudinal and demographic information. That can help you to create audiences. It also has something called BrandIndex, where we track thousands of brands in the market and in sectors, as well. You can start to understand a brand, understand, from BrandIndex, where the problems are, from general awareness, ad awareness, down through to consideration and purchase intent. You can start to see a brand’s performance against other brands. You can see how that is trending, over time. It allows you to create audiences – which a media agency tries desperately hard to create all the time – which are going to drive growth.

Is this audience an area that is going to drive growth? Can you get them to buy more? Is it large enough to do that? You can do that with BrandIndex and Profiles, together. You can create those sort of audiences. Because of our audience data capability, you can put that into a DSP and buy it programmatically. Then, as it is all one data set, you can measure it on that data set. You can see, in BrandIndex, whether your audiences are performing and awareness is going up and whether that is improving whatever you are trying to improve, whether it be consideration or purchase intent.

In terms of the workflow or the process, would the agency be working with the brand, collecting first-party data, collecting cookies and third-party but also supplementing that with YouGov’s unique data set, to build their own audiences on that? How do you merge?

It depends on what a brand is trying to do. YouGov will be a third-party data set in this instance, but a lot of brands won’t build specific audiences out of a first-party data set. They might want to look at a growth opportunity say, for example, win backs, and looking at people who have recently left the brand; you might build an audience out of those people and you can do that with YouGov.

The other way of looking at it is, if you do have a first-party data set, you can also use the YouGov data set and enrich your first-party audience. You can start to append further understanding and that allows you better segmentation, so you might want to do something like that.

You almost feed the YouGov’s audience into your DMP, which somehow integrates on top of your other data?

Not all clients are using DMPs or CDPs. A lot of clients will have a data lake that has some connectivity and you can match YouGov to that data lake. At the first stage of data strategy and media communications strategy, what you’re trying to do is to look at your data set and say, is it enough? What other data sets can I add to it? You look at integrations; you look at enriching data. Then you say, am I able to build suitable audiences from it? If you can, you can move that to the next stage. If you can’t, you look at what you can do or you look at maybe saying, this isn’t suitable. Let’s go and find something that is suitable, to find my growth audiences. Then you move it on to the next stage where you start to look at what the best channels are to communicate with and, if they are into digital, then it’s a lot easier because you don’t have to do the proxies if you can get it into an addressable audience and you don’t have to move to a proxy audience. The problem with outdoor, TV and those sort of things is that you have to move to a proxy audience; specifically for TV. You then start saying, right, what is the nearest audience?

YouGov, for example, has an integration with BARB, in the UK. So what you can do is to look at your complicated audience and then match that to BARB’s, to see the TV programs they are viewing and then you can pull that back and put it into a TV planning tool and understand what the best TV audience to buy is, to reach your audience. It works and it drives efficiency but it’s not that specific audience.

In terms of the offerings YouGov have, they have the custom research side and then supplement that with the syndicated data?

A lot of the agencies are using both because, as YouGov is one single data set and a large-scale panel, if you do custom data, you can match it back to the syndicated. It enriches that custom data, automatically. But a lot of the agencies take our data at respondent level data, which allows them to push it into their planning tools and it allows them to do modelling with it. BrandIndex, particularly, is time series so you can start to use econometric modelling and use it in forecasting. A lot of the agencies use our data, into their planning tools.

How important is YouGov, in that planning process, for an agency?

It depends on your perspective. From my perspective, it is massively important. It is incredibly important because what you need to do is build an audience that has growth. If you can show that that is an audience that has potential then that becomes important. When you’ve got first-party data, a lot of the challenge is that you don’t really know much about those people. If you can enrich YouGov with first-party data, you can start to understand about those people, which then tells you the best media channels to go out and reach them.

How is YouGov moving to that second activation phase of the process?

We have integrations with a number of DSPs and a number of the agency trading desks. With those integrations, you can pick up the audiences in the DSPs. We’re also integrated with LiveRamp in InfoSum, so you will then be able to pick up the audiences in those capabilities.

How important is that, for the agency?

It depends on the type of audience they’ve built and it depends on whether that audience can be lookalike modelled and that determines how important it is. I guess the other one is, how different it is from some of the free of charge audiences that you might get from a DSP, or incredibly cheap data. You need to weigh those up. Lookalike modelling works well for some audiences and less well for others. You won’t be able to tell the difference between a Pepsi and a Coke drinker through lookalike modelling. But you could certainly tell the difference between a BMW driver and a Fiat driver.

You mentioned how the data is really important and can be useful, if you can replicate it as a lookalike audience. Do you need to replicate the lookalike audience on Facebook or Google?

Currently, we’re not in Facebook and Google because there are very few third-party data sets in there because of GDPR. I see that changing because we are fully permissioned. In the future, I see us being able, hopefully, to get our audiences in there.

That could be hugely beneficial for clients.


How do you see this activation phase evolving, over the next five years, for brands and how they approach building lookalikes and actually activating them?

It’s an interesting one because I absolutely cannot tell you what’s going to happen in five years. At the moment, you’ve got GDPR, which is kind of being implemented, but not really. A lot of the regulators are not acting, for whatever reason. You’ve got another one which is that Google and Facebook – particularly in America and the EU – they’re talking about anti-trust, so being too large. You’ve got third-party cookies going from the marketplace. You’ve got Apple and, potentially, Google, who have mentioned scrambling mobile ad IDs, so that’s up there as another thing. With all of that happening, I wouldn’t like to start predicting what the landscape will be like in two years, never mind five.

I think they’re in a state of flux. I think there are some things you can pull out. The market is in a state of flux; you will see a lot of workarounds. You will see digital marketing prevailing and this ecosystem is too big to be dismantled overnight. I don’t see us going back to contextual audiences. I see us carrying on with the ability to use targeted digital advertising; I think that will carry on happening. I think, with companies like YouGov, where we’ve got fully-permissioned data sets, we have the capability to carry on, it is just about what happens in the marketplace.

The last question is around the challenges in measurement and how that’s evolved over the last five years.

I touched on this earlier. I’m going to stick my neck out and say measurement hasn’t really evolved much in the last five years. When I joined Mediacom, five years ago, as a data strategist, I looked at lots of different measurement frameworks. At that point, everyone was very excited about attribution and there were a lot of companies looking at attribution. None of those companies have really come on and dominated the market. I think Google bought one and shut it down. Other people have bought them and nothing has happened with them.

The problem with attribution is, you can’t look at the marketplace and attribute because there are too many holes in it. When you look at the dominance of Facebook and Google, you just can’t do it. What you have to do is to use the measurement systems within Google and Facebook and understand your delivery within a walled garden. I think, the next step for an agency or a client is to try and understand that as a whole and that’s the tricky part. Trying to model the walled garden’s delivery as a single; to understand your duplication and, therefore, understand what a more efficient approach would look like.

You can do it really well, within a walled garden. You can do it really in the open web, as long as you’ve got no walled gardens there. A lot of measurement pixels certainly don’t go through Google and Facebook anymore because it just becomes really difficult. I would say, in the last five years, I don’t think measurement has moved on particularly. Certainly, when I was at Mediacom, we had a really innovative measurement capability which we were unable to pursue because of the changes in GDPR.

What’s the answer to these walled gardens? Obviously, you can’t measure across them but is there a solution, potentially, that has to come around one day? Or are the Facebooks and the Googles just too powerful?

Their solution would be, you can use our measurement outside of our channels but no one else will accept them. They hold all that data within their system. I would hope that some of the big measurement companies are able to start putting their pixels through the walled gardens. To be honest, it’s one for the advertisers; it’s one for the large clients to use their weight to talk to Google and Facebook and dictate that this happens. But they haven’t been able to use their weight on anything in the past, really. It will be interesting to see; I think it all depends on what direction Google and Facebook end up taking.