Being marketers in the land of black boxes

Back in the days when I had just joined Facebook, I remember standing in a room full of agency people and presenting what back then was the cutting edge of marketing - audience targeting. The idea of targeting people by interests, demographics and social connections was shocking to some. "We have our research tools", "We don't need Facebook to tell us if someone is into Luxury goods", or "So basically, my competitors can now target people who like my brand."

Later, in similar rooms, I learned to spot the enthusiastic and the early adopters. Those who understood why custom audience targeting is much better, why automatic bidding always wins, why they should embrace first-party data, and why they should connect their CRMs to media platforms. But I also met the sceptics, who said it wouldn't work, it's too far and too automatic. I learned to turn off my arrogance and be more empathetic. It is intimidating to hand over your strategy, audience knowledge, tone of voice and control to black boxes without knowing what's happening inside. It's not easy to strip off the agency's glory and the years of customer expertise.

Where the big Platforms are heading

And I could also understand why the big tech platform had to shift strategy and do it quickly, almost as a survival act. Privacy regulations and restrictions like iOS14 forced them to reinvent their wheel and give up their interest and keyword-targeting goldmine. Over the past years, platforms like Google and Meta have invested much of their R&D to build algorithms that will switch interests and online footsteps from 3rd party cookies with sophisticated predictions. Without external data, they had to create a self-relaying mechanism that collects signals internally and tries to understand what consumers want, even before they think about it.

With this philosophy, it's easy to understand why any human intervention  - manual targeting, placement selection and bidding  -  can mass the machine learning process. Google was the first to announce that it was slowly removing the targeting option in 2018 when it rebranded Google Keywords as Google Ads. It might look like a minor change, but it indicated where Google was heading: from manually choosing keywords to broad contextual targeting at first and later to a fully automated audience selection, creative and platform matching. Google's current offering, Performance Max, will make itself mandatory for marketers in 2024, removing most of the controls, targeting and selections from the media buyers' hands. It will optimise for conversions and ROAS. How? That's the Google's secret sauce.

Meta, which had a massive hit from iOS14 that resulted in a few gloomy earning calls, had to join this trend quickly and, in 2022, introduced Advantage Plus. Like Google, Meta is also asking advertisers to ditch demographics and interest targeting and even to neglect the much-loved Audiences and Lookalikes. Meta's current message to marketers is: "Run as broadly as possible. Decide on the budget limit, and Meta's AI will work for you across Facebook, Instagram, and Messenger, even deciding between prospecting and remarketing".

The Big Tech platforms are asking you to trust them. Completely.

The most significant change in 2024 is that Google, Meta and others are asking marketers to trust their platforms completely. That was always the case to some extent, but it wasn't so extreme. Until recently, we could mix this trust with our ways of granular controlling: Breaking into niche targeting, A/B testing creatives and audience, and playing with the depth of funnel optimisation.

2024 will mark the symbolic handing over of power to the platforms. The promise will be that the platforms will become increasingly sophisticated and incorporate more machine learning, 3rd-party will be gone, and marketing teams won't have any choice but to align.

The Pros and the Cons

Imagine, only six years from now, running a Google Ads campaign would require using specific keywords—the same is true for Facebook and audience interests. Back then, marketing teams would have to invest in keyword research, understanding which narrow niches would work best with their products,  manually uploading different combinations to the platforms and hoping for the best.

From manual labour, POV and algorithm efficiency,  Meta switching to Advantage Plus and Goolgle to Performance Max is good news. Fewer expenses and lengthy processes, fewer pre-existing assumptions and more freedom for AI to investigate and understand which creative works well, for which audience and predict who is more likely to convert. And all without asking too much from advertisers. Give us a landing page and payment method; save time on targeting; we will do the rest. Google and Meta still need to gain the trust of more traditional brands and agencies, but they will get there.

But the other side is that while the leading tech giants are developing their AI, marketers are standing in front of black boxes that don't require much but don't give back knowledge about their assets. Marketers no longer understand what's working and what's not—leaving experienced professionals to figure out what drove success, which elements of the targeting, the creative and the product. They can't crystalise these learnings and ship them to other platforms or plan how the upcoming campaign will be more successful. Platform AI efficacy comes with a prize and a price, and the price starts to feel like a total loss of control of marketing strategy.

The future: learn to manage the black boxes

Being a marketer in the 60s was an easy task—a creative mind who understood consumers from their gut and decided on the message. In the 2000s, marketers started to be spoiled with endless data and narrow targeting with matched creatives. In the 2020s, it seems like everything is taken away: the cookies, the keywords, the micro-targeting and shifting into a big machine that makes the decisions for you. AI even decided on creative.

And this is where sophisticated marketers will thrive. They will need to search for solutions to help them manage the black boxes and harness their power to their side, not only for the big tech platform. This is why we are building Sphera Networks: to give marketers the power back and the macro understanding of connecting people to products in a platform-agnostic way.

Once again, I can't wait to return to these rooms, talk about AI, neural networks, and closed gardens, and identify marketers brave enough to control the black boxes.

‍