Forget Pope’s fake puffer jackets and ChatGPT’s bad verses. AI is doing “magic” feats, according to people who work at media agencies. Here are just a few of them.
The internet is huge… really huge… It hosts about 1 billion websites, of which about 4 billion people use smartphones and social media, while Google uses about 8.5 billion a day. I am searching for times.
Meanwhile, the advertising happening around all this has grown to a scale beyond human comprehension, consuming as much energy as a small country. Recent outdoor in digital advertising, audio, games and television.
But with the help of AI, media agencies are starting to see some clarity. These companies have long relied on digital tools to gain a competitive advantage. For example, programmatic buying has accounted for over 50% of his UK display ads since 2015. Many of the early problems stem from media agencies not having the time and effort. Monitor where your campaigns are running. Now AI (more precisely, machine learning algorithms) can do the heavy lifting. Amidst the buzz, AI is truly changing business.
Kate Scott-Dawkins, Global President of Business Intelligence at GroupM, believes the proliferation of consumer AI tools like ChatGPT and Dall-E has inspired new ways of working. While people are using generative AI to create prose, images, videos, and more, agencies are using it to solve some of the biggest mysteries, from targeting to measurement to campaign planning and delivery. I’m here. Advances in machine learning models, cloud computing, and enhanced data capabilities in institutions put them in a strong position to find true efficiency.
How AI will change media planning
GroupM’s Nexus chief product officer Rich Astley says human media planners will remain, but acknowledges their roles are changing.
There will always be a place for people who can balance “a little art and science, good judgment and experience,” and AI needs to be freed up to do more of what they do best. Yes, he says. “A human might be able to optimize 10 or 12 line items, but an AI can optimize hundreds of small line items simultaneously.”
For years, the agency has claimed to have “smart” technology and “AI-driven” products, but Astley said this time it was “not BS” and “dramatic and surprising” with the campaign’s performance. This is due, in part, to the fact that we have been trained in what successful campaigns look like and can emulate these excellent decisions. It can make these decisions faster than a human can.
So what is the role of media executives in all of this? The technology is still in its very early stages, says Astley. “I don’t want to train a performance algorithm on the wrong dataset…One of his biggest challenges is getting a reliable source of information to drive an app’s store, sales revenue, or downloads.”
Many humans have run bad campaigns to pursue KPIs. Machines can be trained to pursue these as well, at the expense of all others. Planners should investigate the decisions made by the machine, where the latest innovations come from. The most skilled individuals can now investigate almost anyone.
Diageo uses AI to optimize creatives every three minutes to ensure they’re perfectly suited for each platform they run on using Creative Quality Scores. This cuts the CPM in half. All Nestlé creatives, on the other hand, go through a similar process to ensure that everything they put out meets a minimum level of consistency and does not deviate from historical norms. You can get creative.
AI does the magic of analytics
Third-party cookies are crumbling, raising concerns about targeting and attribution. The agency believes AI can reconnect loops using highly accurate probabilistic guesses.
Scott-Dawkins explains: Just looking like a population is enough. A machine trained on enough ad campaigns and with a sufficiently stable stream of analytical signals can make it work.
Alex Campbell of Publicis Groupe’s Performics technology and intelligence team jokes that they spent “fucking” money to enable AI to “fill in the gaps.” Campbell admits: You can use them in most situations where you would previously use statistical models or some kind of complex mathematical regression model. “
The likes of Meta and Google have been applying machine learning to their advertising business for years. AI creep in these businesses is so advanced that Google failed to launch Bard because it was already an AI business.
“There will be red flags flashing for some traditional agencies and holding companies now that AI allows clients to plan media directly on the platform,” jokes Campbell. However, someone needs to mark the platform’s homework, and he believes agencies can examine these results better than anyone. “We can never say to our clients, ‘Trust us.
Agencies have already run hundreds of thousands of campaigns through the tech giant’s ad networks. They have enough performance reports to form probabilistic benchmarks for everything they run going forward. “You get strong, accurate, and responsive attribution answers without raising privacy concerns,” Campbell says.
These AI projects sound too good to be true. But the very fact that the network is so secretive about innovation may indicate that there is a competitive edge to be gained. Campbell will start one of such projects with an unnamed client. The project is a large one with a series of non-transactional websites that rely on in-store sales.
He said the site could provide product details and direct people to the venue closest to them, with a large amount of advertising dollars being spent on it, and his team combined online analytics and brick-and-mortar sales. We’re working to better understand how to associate and plan an advertising campaign.
Record site-wide analytics events and match offline sales. An AI then finds correlations between two “essentially disconnected” tables, and a second AI assigns each event a score or “powerful business goal indicator” to determine which action on which page. Predict which is most likely to contribute to offline sales. This data feeds ad bidding to direct users to the right landing page at the right time and at the right price. It works really well, he boasts.
On the other hand, Astley warns:
Meanwhile, Dawkins adds: And ideally, they are often always researched to make sure they are free of bias. “
AI, like humans, can be misinformed by bad data and bad intentions. Someone will have to keep a close eye on it.
Questions about AI talent
Scott-Dawkins wants to make it clear that human involvement is still required. “I don’t think anyone has the opinion that the media will be completely dominated by drones and bots.”
It seems like every other week, a report comes out citing lack of media skills. All the experts we spoke to for this article share the view that AI has the potential to lower the barriers to entry for the industry.
Campbell believes coding is entering a new “tier” where machine learning systems can write code. “If you can express what you want to do in natural language, you can start coding right away because you understand the core concepts of coding languages,” he says. He calls this “the hell of democratization…it speeds things up a lot.”
And it goes beyond media campaigns. AI has the potential to transform institutions themselves. Astley wants WPP to be “more like an operating system than a set of companies trying to work together.”
Advertising agency employees are never shy when it comes to describing how busy and stressful their role can be. Hopefully, agencies will be smart enough to stop here and see AI as just another tool in the kit that requires skill to use properly.
