AI for programmatic is great – but tragedy shows us why human judgement remains vital

AI for programmatic is great – but tragedy shows us why human judgement remains vital
Ryan Green is the senior director of client strategy at Coegi, an all-in-one premium marketing partner for media professionals seeking a streamlined way to leverage programmatic and social solutions. Coegi enables marketers to become digital heroes among co-workers and clients by empowering them with best-in-class strategies, technologies, and expertise through simplified partnerships. Ryan has more than 15 years of experience in analytics and consumer behaviour and is also an adjunct professor at the University of Missouri, where he is teaching the university’s first course on programmatic media buying.

Artificial intelligence has been an indispensable tool for programmatic ad buying. But algorithms can’t always tell the difference between positive and negative sentiment — you need a human for that.

This was brought home to me in the days following Kobe Bryant’s death in January. The tragic event spurred a tidal wave of activity on social media, which triggered ads for his jersey and other memorabilia items. This came across as tacky to many consumers. One of my friends, for example, took a screenshot of an ad he was served and expressed his distaste in a long LinkedIn post that received a lot of engagement. But what looked like opportunistic “event jacking” or piggybacking was more likely the result of AI merrily chugging along before anyone thought to stop it.

AI in programmatic advertising

Combining AI technology and programmatic ad buying can create a scalable ecosystem where consumers receive tailored advertisements and producers and publishers garner higher revenue. AI-driven algorithms are typically set up to optimise programmatic ads towards clicks, and they rely on keywords from trending topics to determine which copy or products to feature in an ad. More often than not, this results in consumers receiving relevant content that drives action. However, there are occasional fringe situations where this common approach can backfire.

When sports megastars like Kobe Bryant trend on social media, it’s usually because they’ve done something amazing during a game. It would make sense for a company to try to sell jerseys in those circumstances. However, when an ad follows a negative or tragic situation, consumers will get turned off. That’s why it’s pivotal for marketing technologists to monitor current events and be ready to intervene and adjust algorithms to stop these kinds of ads from being served.

Three tips for effective AI-driven programmatic ad buying

In addition to remaining vigilant, use these three strategies to ensure AI enhances — not sabotages — your programmatic ad buying strategy:

1. Measure what matters. AI provides a huge opportunity for businesses to put their consumer data to work and drive results. You can develop a clear and accurate understanding of what is and isn’t working. And over time, you can adjust your imagery, messaging, and overall strategy to achieve the best possible return on investment.

All of that said, keep in mind that AI and machine learning can only optimise using the signals you give. If you don’t measure the metrics that are most significant to your business goals, AI engines won’t be nearly as effective — and your ad results will reflect that. Like they say, garbage in, garbage out: If you tell the algorithm to optimise towards click-through rate when your biggest business goal is to drive awareness, AI will place ads in front of the wrong users.

2. Implement a smart-pixel strategy. To help you measure what matters, embed a smart pixel on your website that tracks where users come from and how they behave once they arrive. Be sure to compare that data to your most important business outcomes, whether that’s web traffic, sales conversions, or a different metric.

Smart pixels can also help you determine your most valuable audiences. After you identify them, you can utilise look-alike modelling to find even more members of your strongest audience wherever they go on the internet.

3. Encourage cross-team collaboration. Analysing AI-based campaigns shouldn’t be limited to the hands-on-keyboard and quantitative teams. To truly understand what works for your business, you need to bring the data people and the marketing people together.

Stakeholders from multiple departments — including creative, brand, sales, and the executive suite — should collaborate to analyse the top-level metrics. This often involves creating visual dashboards so that campaign data is easily digestible by people with different areas of expertise.

Conclusion

AI can and should play a central role in your programmatic ad buying process, but you need to put proper guardrails in place. Your brand’s reputation could take a hit if you serve up the wrong ad at the wrong time, and your marketing budget could go to waste if you optimise toward the wrong objectives. But all told, the pros of AI in programmatic advertising far outweigh the cons.

Photo by Bacila Vlad on Unsplash

Interested in hearing leading global brands discuss subjects like this in person?

Find out more about Digital Marketing World Forum (#DMWF) Europe, London, North America, and Singapore.  

View Comments
Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *