How the automation of digital advertising is hamstringing the industry.

When we think about the categories of marketing, like advertising, we think of the glamourous side – MadMen-style board rooms, creative strategy sessions, big-brand awareness campaigns on traditional media platforms, etc. Of course, the reality is far, far more boring, but what's often left out of this vision is what comes after ideation and execution – the management period.

Launching campaigns is one thing. Optimizing them over time is another. And when it comes to optimizing campaigns, it's a long, often never-ending period of management. How you respond to the factors that affect the performance of your campaign is what ultimately determines how successful it will be. Sure, strong ideation and creative strategies can help set you up for success, but without a solid post-launch plan and the skillset to make it count, all of your work up until launch was for nothing.

This is where I believe marketers are incorrectly thought of as simply lever pullers. In contrast, I believe we work much closer to the role of crisis management – understanding the tools at our disposal so deeply that we can react correctly and appropriately to any factor negatively affecting performance. When performance is good, we can put our strategy hats on to develop unique new ways to drive performance further, but when performance is weak, we put on our crisis management hats and get to work.

Marketers effectively work in crisis management and understanding this will help you better understand how to functionally manage a downturn in performance.


Footage of a Google rep discovering a tool that can be used to optimize a campaign (joking... kind of? maybe?)

The trouble with working in crisis management in marketing is that the platforms we use are becoming increasingly unsupportive of crisis management initiatives. Namely, the move to automation, among other things. What does this mean?

Let's use Google as an example. Some time ago, Google announced that they were forcing a migration from all Google Analytics accounts to GA4, the newer, far less functional iteration of the Google Analytics line. During this time, Google heard loud and clear from industry professionals that this was not a welcomed change. However, they decided to stick to their guns and continue regardless. Fair enough – they'd already developed it, worked on a migration plan, etc. I may as well see it through. This is unfortunate, of course, but only a glimpse of the real problem.

The problem begins to show its face in the compounding of changes such as these. Google recently announced that, in addition to many other things, they will be sunsetting Similar Audiences – an audience builder for Google Ads comparatively similar to Lookalike Audiences on Meta – in an effort to push closer to full automation.

Okay, maybe Similar Audiences weren't the best tool in the belt, but it seems needless to begin removing options such as these for marketers (crisis managers) to lean on. When managing a crisis (poor performance), it's essential to have many tools you can use to advise recovery strategies. With very few tools, the options for recovery become limited, as do the odds of recovery.

There are many other, more famous examples of these platforms moving towards automation as well, which have been better received: Performance Max on Google, and Advantage+ Shopping Campaigns on Meta Ads, to name a few. While these have generally turned out to be a 'good' thing (results may vary - looking at you, PMax), the principle remains the same – what happens when your PMax campaign dies? Optimize placements and your keywords, or segment your funnels? Unfortunately, that's not really an option any longer.

With the trend of a reduction in tools in favour of automation on ad platforms, I imagine an extremely binary future: either your ads do well, or they don't, and that's it. It makes you wonder whether these platforms genuinely understand their core users. Where they should be designed for those who desire complexity, they are instead designing for the average joe who doesn't have the time nor desire to manage or understand the platform. For the latter group, full automation is great. For the former (you and I), it's crippling.


Great question. So while the above section spells automation out to be a big, scary element designed only to hamstring us, there are some benefits. At the end of the day, humans are not smarter than machines. While good marketers try to use the data they have to make decisions, the fact is that, in most cases, the machine can make a better decision.

This is all to say that automation partially helps the algorithms do better. With less meddling, the algo can lock into a path using the data at its disposal. With automation, it has free rein to do this. Without automation, there are several variables it needs to adapt to – primarily human interaction.

So automation means better algos, which is excellent, but is it worth it? Currently, I think not. Perhaps down the road, when the platforms at our disposal become stable, with less frequent updating/etc., and the algos become truly infallible, then sure. But the current state of these platforms is far from this utopian (dystopian?) idea. Platforms in our present day are very clunky, constantly being updated, they're buggy, etc. You often end up with poor results on a campaign simply because the UI malfunctioned when building a campaign, and you end up with unintentional variables.

In this state, reducing the number of levers we have to improve performance becomes a problem. Again, we should think of every potential variable as a tool in the metaphoric belt of a crisis manager. With each part becoming incrementally more automated, the number of tools in the belt dwindles. Fewer tools, effectively, means fewer opportunities to solve problems.

This is a loss for the platforms as well. In a binary future where either your campaign does well or it doesn't, the outcome these platforms can expect in the latter group is that the campaigns get turned off. Without a clear path to optimization and, therefore, success, it becomes a lot easier to simply remove money from the platforms. Why spend on a campaign that has no hope of improving? Do I keep effectively gambling my company's dollars away on a one or a zero? It would be a poor business practice to spend your dollars in this way, which makes the decision of allocating budgets to platforms like Meta a much easier 'no.'

I hope that all of us can learn something from this thought exercise. Marketers: advocate for control. Push back against the slippery slope of automation because our ability to ensure our hard-thought campaigns see the light of day depends on it. For platforms, keep your users in mind. Automation is sexy, and so are strong algos. Still, it becomes a balancing act – do we need to forego customer satisfaction for what is effectively nothing more than a value add of the platform gone too far? I think not. If we work together, we can find a balance that keeps usability and performance in check.

Today's edition of The Marketer's Playbook is brought to you by Shoelace — Shoelace is a digital advertising agency that helps small-to-medium-sized eCommerce businesses build profitable digital advertising campaigns. Managing everything from Instagram, TikTok, and Google, to email marketing and creative development services, Shoelace has what it takes to help your business succeed.

Share this post