The anti-trend email for 2025
Trends are easy to write. Questions are harder to answer.
Happy 2025. It’s been a while. Life, eh.
My resolution: more regular emails.
Onwards
3 questions to ask of trends in 2025
Predictions and trends for the year ahead always seem slightly pointless in marketing and communications. Useful for understanding context, yes. Useful if you read a lot of them to draw a more macro level set of trends. But what do any of them really mean and can they be applied practically?
It’s why I prefer looking at questions. There’s no definitive answers but they do help frame why this might be important.
For 2025, this is the not-a-trend set of questions.
1. What happens when hope feels a long way off?
This is a big question. The election of Trump. Climate change. Continual cost of living pressures. Global conflict. AI.
All of this creates uncertainty. For the industry I work in (marketing), it creates uncertainty over advertising spend and if people will buy things. Certainly, disposable income and a perception of being worse off means the mood is dark, towards throwing out incumbent governments and taking the view that the next lot can’t be any worse.
That’s a broad overview. Zoom in and - while I’m not a believer that demographics or generations should be lumped in together - there’s a lot of schisms emerging that go beyond attitude.
Let’s take Australian Gen Zs and Millennials. This is a generation that is seeing home ownership - the Australian dream - rapidly accelerate out of view. House prices have skyrocketed due to a lot of bad policies that fail to tackle the basics of supply and demand.
The properties that are being built - largely apartments - aren’t what buyers want and have a lot of defects. The average property price in Sydney and Melbourne is over $1m. The banks’ preferred 20% deposits are out of reach for many, forcing over leveraging and being locked into hideously unaffordable interest rates. Early inheritance or loans from parents are the only guaranteed way to get onto the property ladder, which inevitably pushes prices up further.
That has some wide ranging impacts. The Baby Boomer and Gen X path for life of buying a home and raising a family seems distant if not impossible for many under 50s. Having done nothing wrong and still being reliant on inter generational wealth transfers is disenfranchising.
Couples are now making binary choices between buying a house or having children, often in their mid to late 30s. And the under 30s look at the world ahead and go “fuck it, what’s the point.” This isn’t true for all under 30s - some are very financially and family focused - but it’s very different from 20 years ago, where as somebody in my 20s home ownership and raising a family was very much a tangible dream.
And then staying with the younger demographic, the divide between younger men and women appears to have widened. Conservative versus progressive seems more split on gender than generational lines. The neo-liberalist approach appears to be very last decade or event century.
Broadly, people across the world are fearful, unhappy, disenfranchised and divided. If marketing often reflects culture, what happens when culture becomes more polarised and political? What happens when people don’t want the dream their parents are selling? Is it an issue that we’re having less babies? What does this mean? Will any of this be noticeable? Culture isn’t just what’s cool. And culture going into 2025 feels defeatist.
2. What does it actually mean to use AI?
AI is not - and was never - a bubble. It had been the technological version of science versus philosophy for years until it got an interface. This might sound confusing, but a lot of AI is a bit vague until it gets a specific use case, upon which it just becomes software, and ChatGPT's biggest achievement was adding an interface that made it easy to use and understand by an average non-AI specialist.
So two years on from ChatGPT's big leap forward, what actually is AI used for? Yes, it can create very realistic images, videos and text but how useful is this to the average person.
The broad-use AI tools are very much like taking a blank canvas and asking people to create something. Some people will have a clear idea before they start creating, some will take a bit of time to figure out what to do, and some have no idea and just muck about with colours and brush strokes.
Where AI seems to have evolved into is different disciplines or tools for different industries and tasks, with the better tools being created and used by those people who have a clear use case.
AI can do everything or anything, but there’s a huge difference to an AI engine trained to identify promising new drug combinations for pharmaceutical researchers, and a broad promise of doing sort of everything for the mass market.
Creating an AI tool with the aim of doing anything you can imagine really creates nothing, as where do you start? Is ChatGPT a tool, software or platform? And what’s its actual use case, when a lot of what it promises to do is done better with competitors.
Perplexity does search better, Midjourney is a better image generator than Dall-E. A video generation tool aimed at creating hyper-realistic stock footage for brands that don't have the budget to shoot something bespoke is very different from "Our AI creates realistic videos".
A synthetic data tool like Evidenzia is specifically helpful for quickly simulating long, expensive and hard to generate qualitative research from senior B2B decision makers. A software feature that helps dyslexic users compose better emails to colleagues is more useful than just a general text based LLM.
Using “AI” to replace creative industry professionals can be hit or miss, largely miss. Using a specific AI powered piece of software that analyses patterns in performance creative and makes suggestions for improvement is significantly better. But then all of these examples are software with AI rather than AI.
It’s also an oversight just to view AI through the prism of US, UK and Australian white collar work. It may be easy for the average person to understand, but software that tackles healthcare diagnositics, agricultural optimisation, or access to education is very different and potentially more impactful than office efficiency driving tools.
Ultimately, AI as a phrase is likely to rapidly become as meaningless as "digital marketing" (all marketing these days involves a lot of digital, even in traditional channels like OOH). But will still be used as a catch all term that doesn't really help anybody understand what it actually is.
Some AI will be time saving software features that remove tedious work that could already be done by a machine. Some AI will be very specific and smart for a specific job or company. Some AI will be game changing, but still requires the programmers and users to know what problem they're solving and what they want to achieve.
All told, AI is following a very well worn path for all new technology (combustion engines didn't just power cars). That involves new training, new specialisms, and new applications that haven't been thought of.
That won't stop the overall conflation of of the general or specific, or lazy applications of LLMs which would appear to be good enough but have significant flaws in the data models they’re trained on.
This may be a bigger challenge for AI than developing better software, and these challenges and problems are human issues not coding ones.
3. What happens when media fragmentation accelerates and resources decline?
It's a very easy narrative to say that the US election was a podcast election, but what it did show was the fragmentation of "legacy" media is accelerating, while at the same time traditional media organisations still hold some level of attention. In some respects, this is a logical development: the easier a media channel is to distribute, the lower the barrier to entry.
While traditional media attempted to cover everything with an attempt at objectivity (which was never truly objective), the media mix in 2024 was perfectly happy to be both niche and partisan, or so non-partisan that voices who wouldn't usually achieve a platform get more prominence. This can be a good and a bad thing. Marginalised voices can build a larger reach, but so can conspiracists, cranks and bad actors who wouldn't usually pass an editor.
There's a lot to unpick here. Legacy media isn't necessarily bad either at what it does or catering to an audience. A lot of discussion on social media is still driven by original reporting or culture, but to traction and relevance it requires to generate profits and funding.
But Joe Rogan or Netflix doesn't have the same operating model as network or cable TV, while VCs are less keen on media organisations, which are notoriously hard to monetise (Buzzfeed and Vice, for example, largely were built off somebody else's platform), and there are limited amount of advertising dollars to go around.
Take this rise of retail media. This isn't opening up new dollars, it's another channel for media planners to consider versus other media channels. Online channels - display, social, native and similar - have already directed dollars away from channels such as TV, print and cinema or reduced the level of profit that can be realised through advertising - while it's harder for legacy media to open up to new advertisers with the current model.
It's why Meta and Google were so effective in hoovering up ad dollars: their platforms made it easier for businesses who could never afford legacy channels to self-serve, while also courting traditional advertisers with the promise of higher reach and better targeting, and an ability to appeal to both brand building and performance (sales or acquisition) driven budgets.
Subscriptions, which have been another traditional revenue generation tool, are also fragmenting. When high profile writers with a following defect to Substack or other solo channels, their overheads are lower, so require fewer subscriptions (and other revenue streams) to be viable. The money that might have been spent on a Guardian or New York Times subscription may now go to Jim Waterson or Taylor Lorenz.
So, more media choice for audiences that appeals to niches or mindsets in a different way, combined with new or better reach channels for advertising dollars, and a limited household budget to spend on media creates a squeeze, which in turn creates an environment for layoffs, and with fewer staff, legacy outlets often can't achieve the cultural impact they once enjoyed.
But the above is more than just context. It shows the challenges of the media in 2025. Media production costs, so is more likely to base a large amount of their revenue model around subscriptions and pricing, which skews to a different demographic to free outlets that either have strong advertising revenue, like the Daily Mail & Mail Online, or benefactors such as rightwing outlets like the free press, or social platforms like TikTok, Instagram and YouTube that trade on rewarding attention, or freemium models like Spotify, where the information (podcast) element is part of a wider offering.
In essence, media consumption - and particularly news consumption - looks set to skew even further along affluence and attention-led lines. Those willing to pay for news will get a very different worldview from those who can't or won't.
The media that holds attention for certain demographics - social media, gaming, SVOD - means less time spent consuming a variety of sources. And naturally, if you pay for something, you're more inclined to use it more often.
Some legacy media may find a way to tackle this challenge, but 2025's media fragmentation isn't just about viewing habits, and that has implications beyond just media consumption.
And finally…
There are other questions I won’t go into for lack of space, but are worth considering:
How would a TikTok ban change the landscape? YouTube would be an obvious winner here.
Will Enshittification matter? Potentially for Google, where there are genuine, well backed challengers for the first time since it became the default search engine.
Will creativity / human work command a premium? Or will it be seen as outdated quickly? I’m on the fence on this one: good creatives and strategists spot leaps very few others do. But good prompts and model training can also look at challenges differently.
Summing up: easily answered questions or easily predicted trends probably aren’t hugely helpful, but are probably comforting. Tackling big questions is harder but makes more of an impact.
And now, music…
As you may have noticed, Dispashes have been few and far between in 2024. This should change in 2025. I hope. Please hit the subscribe button below if you’re here for the first time and like what you’ve read.
Playing us out from this edition: Kim Deal - Nobody Loves You More. Turns out all 2025 really needs to be a brighter place is a solo release from an indie legend where she serenades a flamingo with a lo-fi style old standard. We can all only aspire to be Kim Deal when we grow up.