20th May 2024

The potential for Generative AI in ad ops and programmatic media

User Photo
Ed Perry
Ad Ops Manager
Read time: 6min
Left Angle Mask

Back when I was a university student, there was a major road near the campus that used to get very congested. A combination of commuters, the local school run, and students meant that for a good portion of the day, this road was almost completely jammed with cars. I spent far too many hours of my life inching my car along that stretch of tarmac, waiting for my turn at the lights. 

But, on the plus side, one of the best ads I’ve ever seen was placed beside that road. It was a large, brightly coloured billboard outside a mini dealership, bearing the slogan “Wouldn’t you rather be queuing in a Mini?”. 

That ad did a lot of stuff well, but there are a few things I want to draw your attention to:  

  • The message was clear without being too on-the-nose (“Minis are comfortable and fashionable.”) 

  • The ad was hyper-relevant to the context I was seeing it in, without feeling invasive (“We know you’re in a queue right outside our dealership.”) 

  • The tone of voice was perfect for the context (“How about a bit of humour to lift your spirits while you wait?”) 

Partly because of these qualities, I still remember that simple billboard today, over a decade later. So, one question a digital marketing expert might ask is, how can I replicate those qualities on a large scale? 

Enter Generative AI 

Flash forward to today, and the programmatic advertising industry is at an historic inflection point. There’s no understating the impact the shift towards a privacy-first web will likely have on programmatic media. And another massive curveball is coming, in the form of generative AI. 

For those who don’t know, generative AI (often called GenAI or GAI) is basically a very advanced version of the predictive text function on your messaging app. It uses vast amounts of processing power and complex models to sort through a corpus of data and generate unique text, images, and sounds. Essentially, GenAI can be used to make lots and lots of ‘digital stuff’. Right now, I’d argue the quality of that ‘digital stuff’ is still in a place where it requires quite a lot of human intervention before it’s at a decent standard, but there’s no knowing what the future might bring. 

There is a big conversation happening about the societal impact of GenAI, both regarding its benefits and its drawbacks. I won’t be addressing that wider discussion today, but I do want to touch on some of the specific impacts that I expect GenAI to have on Ad Ops and programmatic media in the not-too-distant future. 

Benefits of GenAI 

Ad Design and Creation 

The most obvious use for GenAI in the programmatic space is designing concepts for ads and generating copy. GenAI can produce huge numbers of basic design concepts, or variations on a chosen concept, which would take a normal human designer a considerable amount of time. 

Enhanced Ad Personalisation 

The main advantage of having a large number of ad variants at your disposal is extremely granular ad personalisation. Ad creative and copy can be altered depending on contextual signals, or behavioral signals if the advertiser has the relevant first party-data available. The benefit of this is improved engagement and brand perception. Research from Salesforce shows that 53% of all customers expect offers from companies to always be personalised, with this figure rising to 67% for Gen Z.   

A/B Test at Scale 

Another benefit of having large numbers of creative variants for your campaign is the possibility of large-scale A/B testing. Programmatic advertising can purchase large numbers of cheap impressions, allowing advertisers to gather data about the effectiveness of ad creative and focus media spend on the best-performing variants. 

Streamline Workflows and Automate Time Consuming Tasks 

GenAI can be used to streamline certain workflows which rely on time-consuming data entry tasks. Trafficking large numbers of creative assets, for example, often necessitates huge spreadsheets which it may be possible to populate automatically with a sufficiently advanced generative AI. 

Assist with Custom Tagging 

Generative AI can be used to produce code as well as text, which can speed up the process of creating custom HTML tags and JavaScript variables in GTM. 

Drawbacks of GenAI 

Reduced Oversight 

One major hurdle for a display campaign using large numbers (potentially hundreds) of different creative variants is ensuring brand reputation and tone-of-voice is maintained. Every single creative variant needs to be checked, considered, and approved before it is displayed to the public – doubly so if the advertiser in question is operating in a sensitive area like healthcare or finance. If all the creative variants were entirely generated by AI, the number of required amendments could be huge, which in turn would require significant resources to implement. 

Backsliding Expertise  

Over the longer term, a possible drawback to letting generative AI complete tasks like creating custom HTML tags is that Ad Ops professionals will no longer get to write the code for those tags themselves. This means that if there is a problem with the code, an inexperienced Ad Ops professional might not notice, or know how to fix the issue, because they would never have needed to learn how to write code for tags themselves. 

Environmental Impact 

Generative AI uses more power than typical cloud computing, which in turn requires the burning of more fossil fuels. Businesses and organisations who have made environmental commitments may want to consider how to offset the environmental impact of GenAI before integrating it into their digital marketing. 

Back to that Mini ad...

Generative AI promises to have a real impact on the scale at which creative assets can be produced for display campaigns. This in turn may lead to programmatic campaigns which leverage large numbers of ad variants to deliver an audience experience less like traditional display, and more like the Mini billboard I described. Highly contextually relevant ads, with a clear message and a tone-of-voice that feels personal without invading your privacy. 

While there are some hurdles to overcome to deliver this kind of campaign, mainly centered around oversight and protecting brand reputation, they are not unsolvable problems. In my opinion, the fact GenAI can help to deliver this personalised ad experience without the use of third-party cookies means that the question is when, not if, the AI revolution will come to the programmatic space. 

If you’d like help planning your route through these seismic changes, please get in touch to talk to our experts. We'd love to help you!



Newsletter sign up

Author Ed Perry
Channel Media