Half baked, in the name of efficiency.

Look, it’s a bit late but I was super busy so please forgive me.

Before we get started, there needs to be a disclaimer, warning, caveat, whatever. The opinions expressed in this piece are my own. They are not reflective of my role or my current employer. Some of these may be jarring; if they are, I’d love to hear your alternative perspective to diversify my understanding further.

Over the end of year break, which feels like it was forever ago, I made an effort to disconnect from work for once. I’m usually the one that works through as skeleton staff but this year I needed to take a break. Like many others I had spoken to, I arrived at that decision because 2024 was a shit year in many ways.

For many years, I have been focused on quality first. As is done in sport, you perfect the technique first, then develop the speed. What I’ve seen over the past few years is the reversal of this order. Now I sound like the old bloke yelling at pigeons saying, “Back in my day, quality was important.”

This concept is jarring for many because we can’t see it. “Of course we care about quality! How can you say such things?” Do we? Do we actually care about quality ahead of cost? If trash was getting you a good CPA, would you keep it on the plan? Imagine that conversation.

With my time off, I spent some time thinking about the things I have seen emerging in the second half of the year, particularly in Q4, and how they are all part of the setup for 2025. This year, after taking the break, I came back refreshed and with a positive outlook on the year. I’ve told almost everyone I’ve spoken to that 2025 is going to be a good year and that I have a good feeling about it. Needless to say, that didn’t stop me from being a cynic, albeit an optimistic one.

So what I have this year is a different take on predictions. This time, I have an overall theme of what I think brands and agencies are going to be moving towards in 2025.

The theme of 2025, I predict, is “maximum efficiency.”?

Both cost and time efficiency.?

Quality has been decommissioned.

Trend #1: Cost efficiency over quality.?

Someone opened the cost window and quality flew out. ?

This year is all about maximising value returned and speed to activation at all costs to quality.?

If someone tells you quality is still there, and they’re doing it at speed and at a low cost, they’re lying.

All you have to do is ask for a site report from any total buying solution (what I’m calling advantage+, PMax, Demand Gen, and the rest) and hey presto, shit quality inventory. For the rest of the article, I’ll refer to these products as ‘Total BS' for short.

During the break, I had a moment of clarity. I finally understood the subtext of the “value” of Total BS. While helping a friend paint their apartment—working late into the night in a closed space—the fumes must have led to an epiphany that many others had reached before me.

It finally clicked why Total BS is so appealing to many. My personal desire for doing good work, investing in quality media placements, and focusing on long-term, meaningful results made it hard for me to see the current situation clearly. Budget constraints, squeezed retainers, and reduced control capabilities for end users from platforms have led us to a landscape that prioritises cost-effectiveness, speed, and low quality—the numbers game.

Years ago, at an ad week conference, the CEO of a large media agency discussed 'Mad Men vs. Math Men,' highlighting the divide between creative and technology. Creative is about what you show the person, while technology is about who you show it to, programmatically selected and targeted. I asked if the creative component would ever return, and the question was dismissed with a 'no.' Now, I understand why. In a way, he was right—no one seems to give a shit if the numbers are going in the right direction.

So where does that leave us? It achieves cost-efficient results effectively and rapidly, aligning with what the year ahead will be like. If people are content to invest in Total BS, why burst their bubble? Ride the wave, match your care to theirs. If they don’t want quality, why should you?

I hate doing this, and I die a little bit inside every time I’m forced to.

Trend #2: Short-Term Results & AI

Get short-term results in half the time!

Alright, I think you know how this is going. If you loved short-term results, how excited are you about short-term results 'With AI'?

Get short-term results faster with this new AI powered product. Why bother thinking about targeting when the machine can do it for you?

A few years ago, I wrote about the idea that the only way to out-optimise the machine is by having data it doesn't have access to. The reverse is also true: if AI has access to data unavailable to you, such as 'IP,' and that data provides a performance edge, AI will outperform you. Concealing metrics and dimensions for tactical manoeuvres mutes the ability for the human to drive incremental success.

Think of it as a competitive edge. If your competition has an insight you haven’t yet cracked, they will gain market share with it. The same applies here. AI will outperform you because it has access to data you lack, and you'll often hear 'Powered with AI' alongside 'Short-term results.' and a simplified, streamlined user interface that makes it harder to do analysis.

More often than not, these short-term results are presented as long-term solutions. This raises the question: Is Total BS truly a sustainable long-term solution, or is it merely a parasitic leech that makes it difficult to exit?

So on that note, the second part of this trend is:

The Illusion of Long-Term Solutions.

Short-term results presented as long-term solutions.

In my cynical circles, advertising on content is often seen as a slippery slope to addiction to advertising revenue. How do we generate more revenue? More ads. That's how websites end up with more ads than content. Once an ad delivers revenue, it's difficult to argue for its removal to improve user experience, given the direct and immediate cost to that revenue stream. It is literally short-term “small revenue now” versus long-term “big revenue later.”

Balancing user experience, long-term revenue per user, and direct advertising revenue can often be challenging to debate with a financially-minded revenue officer.

It’s like a Chinese finger trap. Once you insert a cylinder into an m&m’s mini tube, it becomes hard to get out.

I see the same issue with opaque Total BS media. While I can appreciate that transparency is slowly increasing, why are we even accepting a semi-transparent solution? The performance results can be tracked; it's not a reporting limitation but rather a deliberate opacity practice.

Once you’re in, the lack of transparency makes it very hard to transfer learnings and exit these solutions. Turn it off, and you’ll lose revenue because you’ve forgotten how to be tactical and use other channels masterfully. Therefore, it won’t be turned off, not without fierce debate and the odds stacked against you.

Expertise is being sacrificed, and the knowledge of the tactician is being eroded by Total BS. A Malthusian catastrophe of human skill, eroded by the overinvolvement of AI in business.

Due to this erosion, there will be a brief period where the craftsmanship of the master tactician is forgotten, and the simplistic approach of letting Total BS make decisions—where planning involves merely allocating funds—becomes the long-term solution.

It will appear that Total BS can deliver short-term results over an extended period, leading to the decline of the long-term perspective.

Trend #3: Sloppy Work & AI Integration

“Sorry about that, it was AI-generated”.

If you've experimented with AI and received subpar answers, you'd likely recognise them when someone else sends them without proofreading first. Congratulations, you've identified a member of the Sloppy Work Committee. They're committed to taking the least care with their work, completing it as quickly as possible with the minimum quality required to get it over the line.

It’s fine. They’re learning, and so on.

Just don't let them act highfalutin, claiming they've "created a GPT" or some shit.

If they do, you’ll know you’ve encountered a Grade A wanker.

Which naturally leads to:

AI Dependency & Addiction.?

I only use it sometimes.

If you're still learning the ropes yourself, that's fine; I am too. I’m trying to figure out how it fits into my workflow and how much involvement it should have.

What I wonder is whether, once AI is adequately integrated into someone's workflow, they'll realise the knock-on effects for themselves. My guess is they’ll become addicted to it as a necessity of life. They may become unable to reach the same conclusions or validate their preconceived notions without first consulting AI.

Is it a slippery slope fallacy? Probably. But it’s “fun” to think about.

Initially, this will start with sloppy work. People may think AI will immediately solve everything. Just do it with AI, just have the agent do it, etc., etc. Make your own GPT. The balance between AI doing the work and humans performing quality control hasn't been found. AI can handle a lot of volume, but quality comes from a good brief and the refinement of the request.

The other day, I had to tell AI to use a tilde symbol to search for a question mark in a string so it didn’t treat it as a wildcard. I found this information in a Google search knowledge panel. I didn't have to look far, and it's not a new feature. So if I shipped that solution out as it was, it would have been sloppy work on my behalf.

Prepare for a lot of this as people try to balance efficiency, shortcuts, laziness, and actual benefits for their capabilities. If they lack attention to detail, critical thinking, and quality improvement, you'll end up reviewing AI-generated slop.

Haha! Just kidding! Who gives a fuck about quality this year, right?

Trend #4: The Demand for Immediate Results.

I WANT IT NOW.

Shortcuts to the max, long-term isn’t a consideration and tomorrow is too late.

What I'm seeing more than before is this style of conversational skewering:

• Person 1: “That’s great. When?”
• Person 2: Responds with date
• Person 1: “Not good enough, can you do it faster?” / “Not good enough, what can you take out?”

This part is normal business, but the way I’m seeing it come up isn’t delivered in a polite, respectful, or professional style. This is what tells me things are different this time.

The mentality is: Get to the outcome NOW, regardless of its quality. Shipping something incomplete today is better than shipping it complete tomorrow. Reduce the time to deliver a product by reducing the quality of the product.?

This, sounds familiar… Enshittification, anyone? Not sure how that goes?

“Here is how platforms die: first, they are good to their users; then they abuse their users to make things better for their business customers; finally, they abuse those business customers to claw back all the value for themselves. Then, they die.” - Cory Doctorow

Yeah, it gets worse.

Trend #5: Existentialism arrives for many

“You’ll never replace me!” said the horse to the car.

A lot of people are still in their bubbles that their jobs are safe. If it wasn’t clear, Salesforce and Meta made it crystal.

I had a moment of this specific existentialism over the Christmas/New Year's break, even before this news hit my feed. It was a different feeling from the “normal” existentialism I’m most familiar with. The labour efficiency angle hit me when I saw this post on LinkedIn about creating an agent to save a salary.

I know AI will replace a lot of jobs. But now I feel it. It’s not a concept but a reality. It’s so close for many.

It might address inflation (as fewer people have money to buy things), but it creates many other significant problems. That existential dread will arrive for many who were living in their bubble, training their cloud-based AI GPT Agent to mimic and replicate them—answering emails, performing tasks, summarising notes, etc. They did this hoping the AI would be their assistant, not realising they were digging their own graves. They weren’t training their assistant; they were training their replacement. No longer a conspiracy theory, this one is landing for Meta and Salesforce in their engineering teams.

Salesforce:

When asked if Salesforce would have more or fewer employees in five years’ time, he said he thinks the company will “probably be larger”. But he went on to say:?“We’re not adding any more software engineers next year because we have increased the productivity this year with Agentforce and with other AI technology that we’re using for engineering teams by more than 30% – to the point where our engineering velocity is incredible. I can’t believe what we’re achieving in engineering.” “And then, we will have less support engineers next year because we have an agentic layer. We will have more salespeople next year because we really need to explain to people exactly the value that we can achieve with AI. So, we will probably add another 1,000 to 2,000 salespeople in the short term.”

Meta:

"Probably in 2025, we at Meta, as well as the other companies that are basically working on this, are going to have an AI that can effectively be a sort of midlevel engineer that you have at your company that can write code," Zuckerberg said in the interview.

The fear is even there for CEOs.

Last year I was early to predict a CEO would be replaced, but that existentialism is creeping in at all levels.

Klarna CEO Sebastian Siemiatkowski says AI has the power to take over all jobs, including his. Siemiatkowski’s company has already taken steps to replace human jobs with AI. Among the top concerns as artificial intelligence becomes more advanced is whether the technology has the power to take over jobs. One CEO firmly believes AI not only has the power to do menial or repetitive tasks, but also has the intelligence and reasoning to take over his own job as chief executive of a multibillion-dollar company.

Don’t feel it yet? I envy your hubris.

Trend #6: The disposable world reaches the corporate world.

Replacements are faster than repairs.

This approach creates an endless treadmill of stagnation with the mere appearance of progress. No actual progress is made; it is just another short-term solution that shows promise rapidly, with no sustainable longevity or plan baked in.

As soon as questions are asked, another solution supplants it without any foundational improvements. This continual can-kicking approach results in long-term damage.

It's no surprise that when you have to do something quickly, it often becomes a mess.

Imagine how much slower the MasterChef challenges would be if the chefs had to wash and clean everything they used as part of the challenge.?

I expect the last quarter of the decade will involve picking up the pieces from this mess or drawing a line in the sand. A similar line is being drawn, or at least I expect it will be, regarding AI image generation. We are certain these images are photos because they were taken before AI. Everything after cannot be considered part of the corpus of truth.

Like a black hole sucking everything in, there will be a trail of pure devastation left behind for those remaining to clean up.

AI won’t be able to clean this mess up for some reason, maybe because the models can be toxified. Is that how the uprising happens? Through code injection or model toxification?

Trend #7: Vulnerability to Scams

A Strategic Scammer’s delight. Act now, this offer can’t last.

This is a known tactic: when things are in turmoil, people often agree to something they typically wouldn’t. It’s known as the shock doctrine.

Shock Doctrine: This term, popularised by Naomi Klein in her book "The Shock Doctrine," refers to the strategy of exploiting crises to push through controversial policies while citizens are too distracted (emotionally and physically) by disasters or upheavals to mount an effective resistance.?

While this is more about policy and economic changes, the underlying principle can apply to individual decision-making during times of crisis.

Things are tumultuous, and people feel vulnerable about their livelihoods and career futures. This heightened fear and tension lead them to grasp any potential saviour or solution, at any cost.

Usually, the cost stems from a desperation for confidence. This is where a scammer enters the arena, promising something that might not be within their control but presenting it as though they are the one to guide them through the tunnel, through the dark times, and to safety.

There is no safety for the sheep; they become the shepherd's pie.

Trend #8: Approaching the Peak

We arrive at South Col.

There is more growth this year, but the peak is the next stop.

South Col is the final stop before you reach the top of Mount Everest. Efficiency is how we push this little spike to its peak.

According to this diagram, we’re nearing the mid-cycle trading peak: 2026.

What better way to accelerate to the mountain top than to go for maximum efficiency?

If you’ve ever run a marathon or a 100km trek, you know that when you see the finish line, you start sprinting, no matter how cooked you are, just to reach it as soon as possible. That’s where we’re at.?

Accelerating to the peak with cost efficiency in the tank. Cutting labor, cutting quality, and getting as much bang for the buck as quickly as possible before we reach the peak and it tumbles down again in its cycle.

We’re making the car faster by removing everything but the engine.

Trend #9: AI Powered Meat Bags

Smart human thinkers come head to head with AI-powered idiots.

There are still a lot of very smart people in the game who use knowledge to outwit others in a meeting room. They’re often very sharp, can see a few steps ahead, and can craft infallible responses to arguments, debating with sound, clear, and sharp reasoning. An AI-powered idiot is coming close to being able to do the same, provided they have the ability to respond with a delay.

• “Let me take this offline”
• “Let me come back to you”
• “Let me check with the team”

All valid phrases to consult the subject matter expert on the topic.

But what if this subject matter expert is, in fact, an AI?

The only benefit the quick human thinker has over the idiot is the ‘real-time’ thought process benefit. They’re important in the room but lose their power behind email. The AI-powered idiot can chat with their GPT, reach a conclusion, and achieve a better outcome, provided they give it enough QA. If they can’t yet, they will be able to soon.

What’s scary about this chart which shows the test scores of AIs surpassing human intelligence is:

1. It’s old.
2. How do we know if something is smarter than us?
3. How do we train it beyond our capability? Use AI to train it? Cut ourselves out of the loop entirely?

Trend #10: It’s automatic! It’s systematic! It’s AI-dromatic!?

Spaghetti hits the wall and the AI sees what sticks.

Something that has been on my mind for several months is that I haven’t been asked about the quality of media from Total BS yet. So long as the results are hockey-sticking, people don’t seem to give a shit.

I can understand why this might be the case, but I’m surprised and a little disappointed too.

It’s just brute force to get to the outcome—sticking an impression on everything until one converts, then going all out at that placement.

I don’t think I’ve heard much about reducing emissions for a while either, and for some reason, Total BS seems to have skirted it.

It seems like an industry secret that the best way to avoid questionable placements is to operate on an inclusion-only method for open web targeting.

For “some reason”, this is not possible and is at direct odds with Total BS, where the scale of placements and the ability to narrow them to inclusion-only is more difficult than the infinite time it would take to perpetually exclude trash placements.

The administrative burden has been harnessed to make it easier to step out of the spaghetti cannon’s firing line than to spend a few minutes being more precise and less wasteful.

Trend #11: AI as the Scapegoat

Someone will hide behind AI and claim the truth is misinformation.

Plausible deniability will rattle us a bit—that is, until we give up. There will be an image, text, speech, video, or something else in a medium we have always trusted and believed to be true, but it will be claimed as misinformation and generated by AI.

The person represented by the image will gaslight the world into questioning what is true.

Think back; maybe the pope did actually wear a white puffer jacket and just denied it.

You know this is fake; we all do now, but the current state is much further beyond the defects you thought you could see.

Trend #12: Monopolies get stronger.

Goliath Wins. David is on smoko.

Scale often plays a part in the capability to be efficient. This means the best businesses that can drive efficiency are... monopolistic or monopsonistic multinational organisations!

The competitive imbalance worsens between these two, so that between two large enterprises, cost and time efficiency are the currency they seek to sell and buy. That efficiency can be leveraged to drive lower costs and become more competitive to buyers, which in turn gains more market share, leading to greater size.

'Godfather of AI' explains how 'scary' AI will increase the wealth gap and 'make society worse'

"It's because we live in a capitalist society, and so what's going to happen is this huge increase in productivity is going to make much more money for the big companies and the rich, and it's going to increase the gap between the rich and the people who lose their jobs."

Watch this demonstrated with watermelons, provided your country still has access to TikTok.

Trend #13: Tool Fragmentation

There is no defrag for tools

Fragmentation creates individual efficiency at the potential cost of collective efficiency. There are so many tools out there now that everyone can use one of many to get what they need.

Yes, there is a natural heap forming with a few leaders that are most common to use, like most people using ChatGPT as their preferred LLM.

But what about more task- or industry-specific tools?

A conversation might go something like this:

• “Are you familiar with 10Web?”
• “No I use Wegic”
• “That’s a shame, we only use 10Web here.”

The scenario goes like this.

• Some people use tools ABC, which are not compatible with XYZ.
• They know ABC very well but not XYZ.
• The organisation is on XYZ but no one likes it.

The organisation forces everyone to sacrifice their personal efficiency for the collective inefficiency.

Aw heck, why am I being cryptic? ABC is ChatGPT and XYZ is Copilot.

Trend #14: A pivotal moment.

There are some things an AI can’t yet do for some reason, like my timesheets or expenses.?Maybe that will be the job of the future.

This pivot point is where a step change will happen.

People who use AI in their everyday lives, from therapy to reviewing emails, writing their meal plans, and being their confidant, will likely get ahead faster compared to those who are not using it to advance themselves.

Equally, the speed at which people can develop themselves, provided they have the attitude for personal development and improvement—not just achievement of outcomes—will outpace the intelligence we have ever had before and the value they can deliver for businesses. It’s one thing to get the answers to the multiple-choice exam, but another to learn and understand why those answers were correct. You’re better off asking the AI to give you a tutorial on how to do something than asking for just the output.

Many who are not native to AI will have to adapt, not resist. This won’t go away.?

Conclusion:

Businesses will focus on efficiency right now, but that will turn to growth when the first round of transition is done, where everything has been made so efficient that adding people who think the same way will increase potential. More people using AI have the potential to deliver greater business results than fewer people using AI tools, even if they are fragmented.

Quality advertising is always better than Total BS; it’s just that businesses, right now, from my personal perspective, are locked on efficiency for the short term and will be blinded by the results.

I am still optimistic that it will be a good year ahead, and one that I feel more closely aligned with the overall direction businesses are taking.

If this is the case, everything should be clearer this year if we view it through this prism and not the one of growth. 2024 was a tough year; it may have just been the transition from growth to efficiency for many, which many may not have been expecting.