Using Machine Learning to Fuel an AI-Powered Renaissance for Agencies
Artificial intelligence can help agencies be more effective and profitable, but they can only do it when agency networks are willing to plug AI into creative development. In my ad design business, we use AI to capitalize on machine learning and big data, bringing agencies predictive performance insights, such as predicting the future performance of an ad on Facebook. This requires a fundamental reworking of the creative process.
For an outsider, the agency landscape can seem like feudal Italy, with different fiefdoms, duchies, and principalities, constantly shifting loyalties and changing fortunes. Amongst agency networks, there are so many pockets of talent, skill, data and opportunity, but it’s a struggle to see a coherent proposition to advertisers. But like Italy in the Middle Ages, I believe agency networks could be the birthplace of a Renaissance, a creative one that is driven by AI.
What Are the Current Challenges Of AI Adoption?
1. Decision making lag time. Media agencies have taken ownership of data. But for data to make sense and to make decisions using it, humans need to spend long hours developing pivot tables, staring at dashboards and preparing presentations. And as we all know, sometimes all this hard work is stuck in a decision maker’s inbox for too long. Oftentimes, the real insight is reached too late, when the campaign is nearing its end. A further reality is that this data seldom spurs any future creative decision making.
2. Creative development centered on instinct, not data. Of course, creative concepts are rigorously tested in focus groups, but the truth of the matter is that what is finally presented to clients is made after a judgment call by a senior creative. And what is ultimately run is decided after a similar call by a senior marketer. The prominent illustrator Christoph Niemann makes a potent point in his latest book, Sunday Sketching: ‘The small but distinct downside of focusing on your craft is that you become blindsided. What if I’m spending all my energy getting good at the wrong thing? Like a chef who spends night after night perfecting the burger, without realizing that everyone’s become a vegetarian.” Niemann articulates exactly how agencies have become so focused on instinct and “creative talent” that they’ve missed the fast-growing technology that could deliver much-needed added value to their clients. Merging your creative talent outputs with artificial intelligence and machine learning algorithms that predict future performance of your campaigns is what could you apart from your competitors and bring your clients better results (and confidence).
3. Mid-campaign decisions.The distance between brand, content, media and data has not been reduced. The impact of that is that tweaking campaigns to address shortfalls and seize opportunities is still a slow process. One of our clients saw dramatic improvements to its Facebook advertising campaigns by simply increasing the number of ad creatives mid-campaign.
These challenges are increasing in intensity, not decreasing. There is only more data, more channels, more content, more decisions and more players involved.
Which Agency or Marketer Roles Can AI Enhance?
1. Media buyers. AI can predict how well creatives will perform before the ads are run. For example, our platform and ad analyzer uses a proprietary AI where clients can input their Facebook and Instagram ad creative to see its predicted future performance. This mathematical model takes thousands of design points from each ad and extrapolates from past performance data and targeting information. Running ads on Snapchat, Pinterest, Google, Facebook and Instagram generates so much data that the media buyer can make decisions on which creatives to use in traditional media like print and out-of-home.
2. Strategic planners. Based on a mix of hot and cold factors, AI can effectively predict creative performance. By taking thousands of design points from each past campaign or ad and extrapolating from past performance data and targeting information, machine learning is now able to output vast amounts of insight, including predicted future ad performance. Strategic planners can supplement that guidance, using focus groups to tweak creatives.
3. Creative directors. With performance prediction based on past ad creative’s performance, creative directors can free up their teams from a lot of grunt work and focus on finessing creative directions that are already predicted to work. AI cannot only tell you if your creative will perform in the future; it will also give suggestions on how to improve the creative by altering colors, adding or removing copy, changing location images or the sex of the model. With this level of insight, the creative team can refine their ideation in a much shorter period of time. They also can be confident that a particular creative will drive clicks, conversions and revenue for their clients.
AI can truly pull the disparate agency roles together – churning through vast amounts of data, identifying gaps, and deciding on creative direction – enabling us to focus on what we do best: being creative. Allowing machine learning to output insights that benefit each team brings them together in achieving one goal: ensuring their clients succeed.
This article was originally published in Forbes.