Channelling difference

x Sponsored content on Research Live and in Impact magazine is editorially independent.
Find out more about advertising and sponsorship.

There is something distinctly on-brand about walking into Channel 4’s Victoria headquarters, past the iconic ‘4’, to meet Martin Greenbank, the broadcaster’s head of advertising research and development. We are on comfy couches in a central, glass-walled office and it feels like we’re in the thick of it.

Greenbank leads the team in charge of advertising research and its stakeholders are the ad sales team, responsible for revenue generation. This is one of the four areas that make up Channel 4’s insight, the others being data science, audience research and insight, and strategy.

In an open media marketplace, Greenbank’s team needs to prove why Channel 4’s audience is valuable to brands and measure the effectiveness of media investment – from campaign effectiveness, sponsorship and tracking to econometric modelling. 

Research can open doors with advertisers and convince them that Channel 4 is a good use of their carefully guarded media budgets. “It leaves them with a sense that Channel 4, or any other media owner, knows their audience, knows how their products are developing, and is in touch with modern Britain – and, therefore, is a safe haven for where you put your ad money,” says Greenbank.

Trying to understand what differentiates Channel 4’s audience from that of its competitors is a key focus. “There’s a sense of commoditisation that happens in media planning and buying – so you could just bucket everything as ‘a young audience, the 16-34 audience’,” says Greenbank. However, “there are nuances and clear differences in how we deliver that audience”.

The team also creates new ad products for the sales team, one being a tool that identifies what’s being depicted in programmes to place ads around. While the data science team used machine algorithms and artificial intelligence (AI) to develop the tool (see panel, page 51 ), Greenbank’s team produced the advertising research on it. 

Though AI goes a long way to helping optimise processes, and the team is continually looking at how to automate, the human element will never be removed completely – something that’s particularly pertinent in media, with concerns over brand safety and transparency. “Every ad that’s played out from Channel 4 is being viewed by a human, and that’s the difference. We are governed by Ofcom regulations – we are responsible for that. We have a traffic and compliance team, and that’s all they do,” says Greenbank.

Automation “can always be better, but that’s part of the skill of the data science team, to point out where it can be better,” he adds.

Collaboration with other teams is further highlighted by the research community 4Youth, which has 1,200 members and stems from the broadcaster’s research, 10 years ago, to segment the youth audience.

Channel 4 works with LRWTonic on the community, and members of the sales team use it to answer questions from youth brand advertisers.

“It’s everything from posting our products to them, to asking them to take pictures of their fridge contents and mini ethnographic studies,” says Greenbank. “That kind of resource doesn’t just go back to the sales team – it bubbles up to commissioning teams. Part of the skill of the team we’ve got is to spot where something might be useful for the wider organisation.”

New perspectives

With Channel 4’s new national headquarters having opened in Leeds in October, and new offices in Glasgow and Bristol, the research team is branching out around the country. Part of its in-housing has involved training its own team to be invigilators, and one of its projects has them hosting focus groups around the UK. Not only is this cost-effective, but it also gets researchers away from PowerPoint and meeting “real people”, says Greenbank, collecting video evidence that can be turned into “mini sales stories”.

For the past few years, Channel 4 has run a diversity competition, offering brands £1m worth of airtime for a campaign idea that promotes inclusion. The latest competition, won by Starbucks, followed research – commissioned by the broadcaster with YouGov – that analysed the representation of minority groups in advertising. “Poor old YouGov had to watch 1,000 ads! It had to codify these to a quantitative dataset to work with,” says Greenbank.

First, the team worked with internal groups, such as 4Pride and the Collective – focused on black, Asian and minority ethnic (BAME) representation – and took counsel from organisations, including LGBTQ+ campaign group Stonewall, to determine the best approach.

“As we represent audiences –because we have things such as the Paralympics – we’ve done internal studies about how that plays out on screen, and we thought: ‘Around all these beautifully crafted programmes, there’s no real audit of TV advertising.’”

Greenbank acknowledges that it was a 2016 study from Lloyds Banking Group and Cog Research that got the ball rolling, but says Channel 4 wanted to look beyond the top 20 advertising spenders. Beyond diversity – what could be codified on screen – the researchers also worked with focus groups of minority audiences to understand the inclusion aspect – their feelings about how they were portrayed.

“There is a really interesting disconnect between people who aren’t in minority groups and people who are,” says Greenbank. “In some respects, for example, BAME representation [in ads] is in excess of the population size in the UK – more representative, you could argue. But we coded it to the level of what kind of role they played within the advertising, and you find that they’re much more likely to be in a support role than a lead role, or incidental to the plot.”

There is a perception in advertising that “we’re getting the job done” in terms of representation, says Greenbank, but working with the minority groups highlighted how far there is to go.

In the latest instalment of the competition, entrants were asked to create campaigns challenging stereotyping of the LGBTQ+ community, and this was borne out of the research, he adds. 

“We had an idea of another group, but because the challenges almost feel stronger, and there’s more ground to make up, we thought we’d go to LGBTQ+ first. That was where research completely steered us, for quite a major decision, and we have everything to back it up.”

Collective questions

The perennial issue of measurement is never far from the mind in today’s media landscape, and Greenbank’s team works with the Broadcasters Audience Research Board (Barb) to help guide its efforts.

One particular challenge plaguing advertisers and broadcasters is unifying TV effectiveness data with streaming and video on-demand. Broadcasters have been grappling with this in recent years, to furnish advertisers with the evidence they need to justify spend. This has been particularly keenly felt with increasing investment in Facebook and Google, despite well-documented issues with measurement and brand safety.

“There are some very simple questions that have very challenging answers. We know that Barb will get there, and we are helping put in place as much as possible to either help Barb or answer those questions,” says Greenbank. “That’s the biggest challenge, because what we have found is that the ad money has followed people who can give answers, even though some of those answers may be very questionable.”

In recent years, there has been a concerted joint effort from Channel 4, Sky and ITV to work together to solve the problem. One of the catalysts for this, says Greenbank, was the Project Firefly research, which brought together the broadcasters’ first-party datasets to determine ad exposure in broadcaster video on-demand.

“We thought: ‘We could do this on our own, but it would cost us an awful lot of money, and we might not get it right.’ So we asked ITV and Sky to do it with us. We took our sales team with us, and told them: ‘This is not about sales. Put your swords at the door, walk into the room, and listen to our proposition, about doing something that will answer our problems and benefit the industry.’

“From that moment, two years ago, it’s just accelerated – we’ve found a lot more collaboration is possible.”

The advertising research team benefits from having a cinema in Channel 4’s head office, where it can host lunchtime sessions to share research with up to 80 people at a time. It can also book space on screens around the building, which it did for an internal campaign called ‘WTF is Dovetail?’

While Greenbank can’t do every research project he’d like to, the team tries to make sure the output reflects Channel 4’s forward-thinking credentials – and, if it comes down to two methodologies, they’re more likely to trial a new one. 

“We do it because we know there’s revenue linked to it or it’s reputational – ideally, both. Because we have a positioning within the marketplace as quite innovative, we have distinctive content on screen, and our research does reflect that. It is a very considered process, and we bear in mind that our stakeholders are a commercial team.”

Academic view

Feeling jaded by research approaches based on cost efficiencies rather than the best insight, the broadcaster set up a PhD programme at Durham University, to explore video advertising and how the brain works. “When we were briefing out to media, to research agencies, we could almost predict what the responses were going to be,” says Greenbank. “The methodologies weren’t changing; the ways of thinking were quite formulaic.”

So they recruited Matt Thompson, a student who had a Master’s in experimental psychology, for the programme, and he has been involved in a number of projects, including its Contextual Moments work (see ‘People and machines in context’ below). Thompson helped executives explain the offer to the market, because, Greenbank admits: “We had no idea what was going on. He could articulate why certain results were happening and use previous experiments that were quite well known within neuroscience but completely unknown to us.”

Channel 4 chief executive Alex Mahon was originally a scientist (she has a PhD in medical physics), so this openness to new techniques and knowledge is fostered internally. It has contributed to a sense that research is not merely a support function, but a discipline in its own right, says Greenbank. “She [Mahon] knows the value of research and she’s enabling that through the business. She’s challenged everyone, saying ‘don’t show me the research – show me the insight’.”

Greenbank brings the perspective of someone who has been a user of research for most of his career; before joining Channel 4 in 2014, he held senior planning roles at media agencies. “The head of research should not be a brilliant researcher – our challenge is to liberate insight across the organisation.”

People and machines in context

In 2018, Channel 4 launched its Contextual Moments tool, which allows it to run advertising next to relevant scenes in a TV programme.

The technology was developed by the company’s data science team, and it applies AI to identify specific moments around which adverts can be targeted. If characters in a programme are having a beer together, for example, the next commercial break could present an opportunity for a beer brand to advertise.

Although the technology automatically identifies moments within a programme, humans are still needed to do the final check and validate the moment, to ensure they are positive.

The organisation is also exploring how it can use machine learning to forecast the demographics of its audience. Speaking at the Unbound festival in London, in July 2019, Greg Detre, former chief data scientist at Channel 4, said: “Channel 4 has an expert human team that does a pretty great job at this. The question we were interested in is ‘can we improve on that if we include machine-learning techniques?’”

Some human processes are very difficult for machines to replicate, while it is not cost-efficient for humans to spend a lot of time on others. So the focus is on what Detre called a “hybrid system that is the best of both worlds – that allows humans and machines to complement each other”.

This article was first published in the January 2020 issue of Impact.