Profile picture for user hannah.porritt
Posted By
Hannah
Porritt
hannah.porritt@cognitomedia.com

Advertisers using Meta’s AI-enabled advertising product ‘Advantage Plus’ have recently faced much higher-than-normal advertising costs with meagre returns, found The Verge.  

Advantage Plus was described by Meta as a ‘set it and forget it’ tool to help advertisers create, launch and run online ad campaigns across the Meta ecosystem more efficiently. So much for that.  

This product targets e-commerce brands, but financial services aren’t far behind. LinkedIn, the most popular advertising platform for financial services brands, has had automated bidding strategies for years. Its algorithm optimizes ad delivery towards the objective you specify at the outset. New AI tools help generate ad copy and create graphics, while ‘Accelerate’ sets up the whole campaign and ads for you. 

The logic makes sense. There are parts of campaigns that arguably don’t require much thought but can be time-consuming. Duplicating ads across campaigns is tiresome and paying close attention to make tiny tweaks to ad variations and bid amounts takes a lot of time, especially across audiences and formats. Having a computer take on those responsibilities should be more reliable (they’ll never forget, be in a meeting, or be too busy to check the campaign at the same time every day) and more accurate (with much higher computational power than I could ever claim).  

That’s the idea. I did a quick search on LinkedIn and the consensus among advertisers seemed to be ‘don’t use it to launch campaigns’, ‘some things that need to be worked out’, and ‘unsure about these targeting options’.  

Are we humans just dinosaurs who can’t deal with change, or do we have a point?  

Here are two main reasons I’m still wary of handing over the reins completely to AI and a few use cases where I think it could be helpful: 

The AI doesn’t have the same goal that I do

At the end of the day, I want to run an efficient campaign that gets the most results at the lowest cost. That is not LinkedIn or Google or Meta’s goal. They want to make the most money from advertisers. Digital marketers on ad platforms interact with ad representatives and settings designed to get them to spend more money. Their AI and algorithms’ definition of what success looks like is going to be different to mine.  

Switching off automated campaign settings like audience expansion and the LinkedIn Audience Network is the first step for experienced LinkedIn advertisers. These settings spread ads beyond the specific audience you choose, creating competition where you didn’t want to compete in the first place. More competition increases costs, which is good for exactly one person – the platform.  

AI can’t understand my client or their audience as well as I do 

Copywriting for ads is an art, not a science. The idea of an AI-generated ad based on data from a specific target audience and what they’ve responded to previously makes sense. But brands have specific tones of voice, value propositions and priorities that you must incorporate into ad copy and creatives. We’ve written about the considerations of AI-generated copy, from inherent data biases and originality, and the same principles apply to ads. 

The most effective ad campaigns don’t run in a vacuum. Decisions on where to deploy ad dollars or pull back on spending must be informed not just the on-channel results but overall business impact. For B2B companies, that’s a lot more difficult because form fills and contact forms don’t often lead directly to revenue, and automatically optimizing towards these goals alone, rather than towards real business results, won’t drive more revenue at the end of the day. AI can only do the right predictive modelling effectively if it has good data specific to each customer – specifically MQL and opportunity contact lists. Finding this kind of data in the volume needed for the AI to really work effectively is often a challenge.  

Use cases that might be helpful 

We want to be open-minded about the potential for AI. A couple of places where AI could help in the campaign build stage could include: 

  • Inspiration and new ideas for creatives and offers – AI-generated ads can act as a drawing board for new ideas that can be tweaked and adapted into campaigns 

  • Research how landing pages will perform, by giving your URL to LinkedIn and assessing what the AI-generated ads tell you about who it thinks the page is relevant for 

  • Find new audiences and add them to A/B tests 

Maybe I’m just a control freak. I reserve the right to change my tune, especially the next time I’m working on a vast campaign with multiple channels, audiences, formats, and budgets to keep track of. If AI can help me work faster and run better campaigns, I’ll give it a chance. At the moment, it can only truly deliver against the first of those goals, and neither my clients nor I want me to be a busy fool.  

For now, I’m going to choose not to ‘set and forget’. 

Hannah Porritt is an account director in New York