Digitisation of the media has made advertising a highly data-intensive industry. This has pushed media agencies, which advise businesses on which advertising channels to use and execute their campaigns, to the forefront of data analytics. MediaCom, the UK’s largest media agency is no exception, with CTO Nadine Thomson overseeing an agile working transformation to help the agency deliver insight to its clients faster and more effectively.
The most valuable data for MediaCom and its clients is anything that describes the reach and impact of an advertising campaign. “For example, if the advert is a video: is it getting watched to the end? If the ad is a call to action: is it getting clicks to the advertiser’s website? We basically measure if people are engaging with the ad,” explains Thomson.
However, every campaign includes a unique mix of channels, from outdoor advertising to social media. And clients all have their own “business logic”, Thomson says, meaning they each have particular objectives and metrics for tracking them.
This means the agency’s business intelligence (BI) team contends with a wide range of data sources and reporting requirements. Every instance of a digital advert carries its own string of metadata that allows its individual performance to be monitored. A typical campaign report combines at least 12 data sources (such as ad performance data from Google, Twitter or Instagram) – some have up to 50.
The team uses Microsoft’s Azure cloud services and a specialist data integration and analysis platform called Adverity to crunch the data.
“There’s a lot of work with gathering and transforming data, not only connecting to the different data sources that are external, but also building connections into different APIs – whether that’s Google, Amazon or any of those sources – and then also transforming that data,” Thomson explains.
Using AI to understand creative
Having accumulated many years of ad performance data, MediaCom is now looking to artificial intelligence (AI) to understand what really drives success. Thomson and her team are currently working on scaling a pilot that examines the impact that an ad’s creative elements – the idea, copy and design of an ad – have on its performance.
“We analyse what makes the content of an advert appealing,” explains Thomson. “What is the most successful way of organising the content for that ad? What are the best colours? Then we analyse historic data using machine learning and we can see that this ad got a 0.5% extra click rate compared to the other, where the brand logo was smaller, for example.”
This is only possible thanks to the wealth data the agency has accumulated. “Once you start getting volume, you can start saying there is something in those elements of the ad that made it successful: all had very similar elements running through them.”
MediaCom is currently moving that pilot system on to Google’s cloud infrastructure in preparation to a global roll-out to clients.
Introducing agile working
Thomson’s chief contribution to this ongoing data transformation has been to ensure the structure and processes are in place to translate technical expertise into value for clients, she says.
When she joined MediaCom in October 2019, the global technology team was disjointed and siloed, she says. Putting together the BI systems required to report on a campaign could take up to six months. But Covid-19 disruption meant clients needed faster access to insights about campaign performance.
“With the pandemic, there is a need to act fast, otherwise you miss out. Clients need to make fast decisions – if cinemas are closed and clients can’t advertise there or their customers are increasingly using digital channels, they need to be able to react quickly,” says Thomson.
“Because of pressure on their marketing budgets, our clients need to be able to explain how they are being spent and what the return on investment is on that kind of budget, whether that’s improving their brand or whether it’s growing sales in particular areas of the company.”
Before the pandemic, Thomson had already begun to address the siloed nature of the technology team. She established an agile project management office (PMO), a centre of excellence to advise the rest of the technology division on agile working methods. This meant recruiting a small number of agile experts who embed themselves in teams and coach colleagues on agile working methods.
The results were soon evident: from working in siloes, staff moved into running cross-functional teams, Thomson says. Project managers, business analysts and developers began working together, not only bringing teams closer within the business but also clients.
“Implementing an agile PMO has been a turnaround for the whole organisation. I look back 12 months and I think, ‘wow, what a change the team has made’. We have delivered projects that used to take six to 12 months in just 12 weeks. We have developed a stronger relationship, a partnership, with the business. We’re delivering some really cool technology and getting involved in amazing projects and I’d say the team has come a long way to be able to come to that point.”
But this, warns Thomson, is still an ongoing journey involving an important cultural change. “I think agile is a journey in any culture. Change is a journey.”