Machine learning and Coronation Street. The combination sounds like something from a hare-brained scheme that you might overhear Steve McDonald, Weatherfield’s resident dodgy dealer, dreaming up at The Rovers Return. But not only is it real; it’s a pointer to the future of broadcasting – and we can probably expect contextual advertising to find its way into the ad breaks served up by any commercial TV channel.
Having started in 1960, Corrie is the world’s longest-running TV soap, so the prospect of an AI-driven Weatherfield represents a curious juxtaposition of tradition and modernity. Not without reason might cynics suggest that such a model would herald a marketing-driven era in which computers are in creative control and on a singular mission to monetise at all costs.
But, according to ITV’s group chief data and AI officer, Sanjeevan Bala, that won’t be the case – at least on his network. He acknowledges that there is a debate about “whether machine algorithms are going to decide what we watch or not. Speaking from ITV’s perspective, it’s definitely not the direction we’re taking.”
The thing that grates with consumers the most is when advertising isn’t relevant
Having worked as a consultant and strategist across a range of industries, including a four-year stint in marketing with Tesco Clubcard pioneer Dunnhumby, Bala eventually settled into broadcasting in 2011, becoming head of planning and analytics at Channel 4. He was responsible for a huge growth in the amount of first-party data received by the network through viewer registrations.
He joined ITV in May 2020 – and not a moment too soon, as the UK’s first Covid lockdown was spurring a huge surge in television watching. He has since overseen a range of data-led pilot projects, including a drive to obtain more information on viewers via registrations, just as he had achieved at Channel 4. This first-party data is set to be crucial in bringing contextual advertising to life in a way that’s effective but subtle enough to avoid sullying the character of some of the nation’s classic programmes.
Bala recalls that ITV understood “very quickly” that it had to retain the best elements of its programming – the character of the shows, the creative components and the human stories – and combine these with the latest tech. While art and science can seem like polar opposites – as was the case when he arranged one disastrous meeting between a programmer and a data scientist in his previous job – it doesn’t have to be that way.
“We often talk about this augmented world where it’s the best of both,” he says. “It’s about how we create tools and understanding and combine these with editorial, commercial and distribution.”
Here’s how it works in the case of Coronation Street: internal commercial team ITV AdLabs, created in 2021, has built machine learning systems that, in effect, watch and listen to episodes. Their algorithms will then recognise contextually relevant moments. So if, say, food features prominently in a particular episode, then relevant adverts could more easily be served up – by a food-box subscription service, perhaps.
“From an advertising perspective, the metrics are really strong,” says Bala, who holds AdLabs’ work as a “great example” of where ITV is bringing together art, science and creative storytelling. “The thing that grates with consumers the most is when advertising isn’t relevant,” he argues.
This is just one example, of course. Under Bala’s guidance, ITV is taking an “iterative” approach where it demonstrates the value of data to one organisational function and then uses its successes as a way to make others more receptive to the concept. For instance, soon after he joined, his team started targeting new viewers on social media with promos for shows – people who may not traditionally have tuned into ITV at all – and then showed the results of this initiative to other departments.
The trick to gaining their support for the new approach, Bala reckons, is to avoid talking about tech wherever possible. Businesses that focus too much on technology risk falling into a trap where they “end up with a faster horse” but fail to secure “real transformation”, he says.
When Bala joined ITV, for example, the broadcaster already had in place technology components such as Crocus, an analytics tool developed in association with Google Cloud Platform. But what it still needed, he felt, was an articulation of the group-wide mission, which required him to explain the desired outcomes in the language of the people he was talking to.
“At our core, we broadcasters are creative producer-distributors,” Bala says. “Our value chain is about commissioning, acquiring and distributing content. It’s then about promoting that content through marketing and, lastly, monetising the content through advertising. When you talk to business leaders in each of those areas, you start to have conversations about how data might inform what you commission or acquire, or ask ourselves, say, what if we could use data to think about how to distribute content differently?”
We avoid the term “data literacy” – I think it can be slightly contentious
Over the next year to 18 months, ITV hopes to expand data initiatives into its marketing, product and commercial teams, including the wider roll-out of contextual advertising. With this transformation already under way, the network has some bold ideas for what’s next. Last year, it invested £2.5m in Metavision, enabling it to expand its intellectual property into the metaverse. In July it made its first foray into this virtual world by launching a version of its gameshow The Void in the Fortnite Creative sandbox video game, for example.
To realise such ideas, Bala and his team are focusing on “fixing the last mile first”, which means ensuring that both business leaders and end users understand the value in any projects that are under way, and then connecting everything up to the outcome, rather than worrying about all the technical details from the get-go.
Similarly, he feels that organisations occasionally concentrate too closely on achieving data literacy as opposed to building business literacy. “We avoid the term ‘data literacy’, because I think it can be slightly contentious, with its implication of illiteracy,” Bala says. “We do talk a lot about how we can upskill our colleagues and improve their capabilities. And we’re making sure that we do it one team at a time.”
This has to be a two-way process, he stresses. Data professionals must understand how an organisation operates, what its culture is, how information is used and what choices can be made. “This is about how you mesh all those things together to then roll out a more relevant programme.”
