.png)
“AI can offer prosperity and opportunity,” said Adam Ray, CEO and founder of adtech company 59A.
But Ray stressed that it is a complex subject and shouldn’t be thought of as a single technology. “AI is a massive umbrella and can be applied in lots of different ways,” he said at the recent Power of AI Summit, an event jointly hosted by Campaign and Performance Marketing World. “Programmatic took a decade to understand and I’d like for AI not to take that long.”
To help illustrate the point, Ray showcased nine of 59A’s case studies to show “what’s possible today” for brands in their deployment of AI.
Discovery: identifying the best targets
A large carpet and flooring retailer wanted to fine-tune their targeting on Meta so 59A helped them analyse more than 800,000 UK property transactions across a year and split the data at each stage to discover actionable opportunities across all 11,230 postcode sectors of the UK. Understanding not just when transactions take place, but in conjunction with knowing the type of property, its proximity to the retailer, and its local competition density, elevates targeting capability to a new level.
“Getting that valuable data into advanced systems is quite a hard thing to do,” explained Ray, “and for this one we were able to leverage a range of powerful, very large data sets encoded into a custom algorithm. This was information they didn’t have before and helped improve the delivery of the campaign.”
Development: unlocking new opportunities
All charities in the UK face a similar problem: the classic donor demographic is progressively ageing and they need to develop new ways to find donors. “Adults are getting generationally older,” said Ray, and “by 2043, 13.4m will be over the state pension age.”
The approach used here was to move away from the widely used household income targeting and into targeting by disposable income, a new capability possible through the creation of a custom algorithm. 59A analysed every privately-owned property in the country through open-source data sets. (“We believe AI should be completely transparent, accountable,” added Ray.) They identified around 7.6m households that were owned outright with no mortgage, which suggested there might be more disposable income in those particular areas.
Not only did they utilise mortgage debt and Land Registry data, but they also integrated the information with the locations of 7.5m households holding mortgages. This combination allowed them to establish a scale for assessing decreasing mortgage household debt at a hyperlocal level. Ray said: “This approach resulted in 3x higher cash donations in those areas.”
Build: vast knowledge base
“The potential of AI is incredible,” said Ray, “even if the answer isn’t a billboard on the moon.” 59A built a database of every single company (5.4m) that has existed in the UK since 1416. “This is an invaluable knowledge base, a creative enabler for things which have historically been impossible.” Integrating such vast amounts of historic data into an AI-powered custom algorithm creates countless unique opportunities as it can leverage every dimension within the data set, and explore all potential avenues, even ones yet to be conceived by a human.
Unlock: turning complexity into simplicity
Ray reiterated his desire for AI systems to be “open and transparent” and not be a “black box of claims”. 59A’s data science team told Ray that we need to think about data being scrutinised in 10,000 dimensions – the scale of AI is seemingly impossible and unimaginable. Its vastness, though, should be embraced. While humans are unable to conceive beyond four dimensions, AI has the capability to create simplicity from what “feels to us as the far side of complexity”.
Transform: educating the educators
More than 29,500 schools at every level were mapped and graded across 132 dimensions, including exam grades, Ofsted rating and socioeconomic factors. “If we’re able to understand things at this new scale, then you can be much more efficient and effective in how you execute media,” said Ray. This transformative capability can revolutionise strategic approaches not just for higher education providers but for many brands for which families are a key target demographic.
Create: use more data to fit unique marketing concepts
59A refined its algorithms to align with the elevated marketing goals of a drinks brand. They brought this to fruition by conceptualising the customisation of a master algorithm for continuous brand activations. Beyond that, they demonstrated how data sets could be employed and weighted differently in subsequent custom algorithms, each optimised for diverse objectives while upholding a uniform data approach.
Enable: view through your own lens
Each brand approaches their audience with unique needs and objectives, shaping their perspective through distinct lenses. 59A extracted 14m data points spanning 20 years of mental health data and distilled them into a visually engaging word cloud. This highlighted the varying significance of different towns and cities in the UK, challenging common expectations. This also revealed the prominence of rural areas, a pivotal factor contributing to the success of the campaign. “You’re thinking about things in a new and powerful way,” said Ray.
Operate: the vastness of the custom algorithm
59A’s custom algorithm technology operates via a list of instructions which undertake a billion different reviews and decisions being made 24 hours a day, seven days a week on a typical campaign. “The technology infrastructure we built facilitates this and allows you to operate at atomic levels of detail,” said Ray.
Empowerment: helping brands follow their audiences
59A has collected data relating to the 790,000 roads in the UK, where they are, the road type, the speed limit, and how safe they are based on collision and traffic density data. This developed into an idea of road consumption, a unique way of understanding how car brands and activities are consumed by the British public.
Building on this: “We know where all 44,040 electrical charging points are,” said Ray, adding that the target for 2035 is 300,000 points to service an estimated 8-11m electric vehicles. “We also know where every new-build property has to have an EV charging point.” These are all crucial data elements which would not typically form part of a media targeting strategy, so incorporating them alongside other data sources such as historic media performance and customer data, can empower brands with new thinking and ideas for targeting and engaging with their audiences.