New technologies in advertising often promise to replace the high-priced labor performed by ad agencies with more efficient automated processes – reducing the need for workers and putting an increasing squeeze on agency revenues. Throughout the past decades, new ways of working and new technologies have continuously threatened to replace the role of ad agencies. And no technology has had more disruptive potential than Generative AI.
Paradoxically, the adoption of time saving and efficient technologies in the past has led to an increased need for more specialized, educated employees, and generated increased margins paid by advertisers for access to new technology. Look no further than the adoption of programmatic, ad exchanges and demand side platforms. Salaries of individuals who manage these programs exceed those of other agency staffers – and the margins charged often soared manyfold over standard agency commissions. Increased returns on lower cost ad placements – even with higher commissions – eventually made this a win-win for both agencies and advertisers. However the promise of automation of advertising tasks was not realized as specialized labor needs actually grew. And only with many growing pains – and considerable economic and human capital – was the promised efficiency turned into actual positive results.
We are likely to see similar effects with Generative AI, at least in the near future. First let’s look at some of the tasks that AI may be increasingly able to perform.
On the creative side, AI will have immediate impacts in a wide range of areas including but not limited to:
· To assist creative teams in producing more variations and visualizations of ad concepts – less expensively – and in a shorter timeframe.
· Producing ad creative targeted towards increasingly specific targets, allowing for better testing and optimization of creative assets.
· Speeding up the process of generating storyboards and cinematics for ad concepts.
· Video and image editing, increasing the speed at which we select scenes, crop, filter and apply special effects.
For media departments, some of the tasks AI will be used for include:
· Perform preliminary data analysis and insights – digging through past campaigns to find trends and correlations that may prove useful for upcoming ad campaigns.
· Assist in media planning, reviewing past ad plan performance, current pricing information, predicted response rates – producing a greater quantity of media plan options at different spend levels and with variations of targeting options.
· Assistance in performance tracking, optimization and alerts – increasing efficiency, noting in real time when assets are underperforming or problems arise in the campaign.
One might look at this list and calculate the many hours of agency work that will be saved – and in some way that is true. But likely the result is humans will be presented with even greater options and choices – that can only be made by humans. Turnaround times will be compressed. Our robot assistants may wind up giving us even more work to do in a shorter period of time.
The idea of time-saving technology making us work harder arises from the fact that as technology advances and tasks become more efficient, the expectation and demand for productivity increase. This is called Jevons’ paradox – named after the 19th-century economist William Stanley Jevons, and describes a counterintuitive phenomenon in which increases in the efficiency of resource usage lead to higher overall resource consumption, rather than conservation.
For an example: Imagine if we can produce many dozens of different creative ideas or media plan variations in less time it previously took to produce three. Who in this scenario would be satisfied with what was acceptable in the past due to human time constraints? The increased human tasks in evaluating and analyzing the vastly increased work product may completely negate the time savings generated by automated production. Who would be happy with A-B testing if for the same cost and timing one can perform A-Z testing?
Another paradox that may arise is that increased options and choices made possible by AI could actually create poorer results. This concept was explained by psychologist Barry Schwartz in his book titled “The Paradox of Choice: Why More Is Less.” Schwartz argues that while having some choice is beneficial, an excessive number of choices can overwhelm individuals and hinder their ability to make optimal decisions. Dating apps may have greatly increased people’s choices among potential suitors, but arguably have a deleterious effect on the quality of those dates.
When advertisers are faced with an abundance of options, decision-making can become fragmented and unfocused. Marketing directors may spend excessive time and effort evaluating numerous alternatives. The cognitive burden imposed by too many choices can also lead to decision fatigue, diminishing the quality of decisions made. Furthermore, an excessive array of choices may create internal conflicts, as different team members advocate for their preferred options, potentially causing delays and indecisiveness. As a result, the overall effectiveness and coherence of the ad campaign can suffer, leading to decreased impact, confused messaging, and ultimately, poorer results.
And if we let the machines make those decisions? Here we run up against the limitations of Both AI and the data that powers it. Despite decades of work in this area – digital attribution remains full of flaws, limitations and blindspots. A decision making process left to machines will likely be skewed more by attribution model algorithms than pursuit of actual success. Access to more data and more deep analytics has not always led to better results. Real live people remain complicated and often unexplained by data.
Despite these potential traps, the lure of using AI in advertising will likely be irresistible. One sure thing is that at the beginning, many tech companies will make fortunes selling AI ad tech solutions to clients that are ill equipped to use them. There will be sudden and severe inefficiencies among the earliest and heaviest adopters. This might be akin to early investments in DMPs that cost a fortune, took ages to make useful and demanded skilled (and high cost) human staff to salvage the technology from being a bust.
And what lies ahead for the next generation of ad agency employees? Some might argue that the removal of rote advertising work might be akin to taking the weights out of the weight room, leaving future advertising executives weaker for the lack of basic training. I suspect this concern is overblown. Creative departments long ago stopped making print ads using mechanicals cut and pasted by hand. This did not lead to weaker creative staff. The greater risk is that in an industry already characterized by high stress, tight turnaround times and long hours — we will increase the burden on those who work both agency and client-side. Wise adoption of AI with an emphasis towards providing benefits not just to corporations, but also the humans that work there will be critical to maximizing the returns on the adoption of AI.