Prediction vs reality: What can 2025 tell us about the art of prediction?

2025 is a bit of a milestone year. It’s halfway through the 2020s, and it marks the end of the 21st century’s first quarter.

Whenever you have a milestone year, you tend to get a lot of people trying to imagine what life will be like when it comes. OK, 2025 might not have attracted as many as the year 2000 (hey, a new millennium doesn’t come along too often), but if you dig deep enough, there are hundreds of predictions put on the record imagining what life would be like in 2025 - with the most famous perhaps being a prophecy of women opting for silicon partners over human ones. 

The video below gives something of a flavor to how many visions of 2025 there have been:

Now that we’re here, we can do something quite unusual for predictions, and actually see if they came true or not. We scoured the archive to find all the predictions about 2025 we could, then used a combination of our own surveys and third-party data to sense-check them.

Here are the biggest things we learned. 

1. Most predictions are horribly wrong.

The old quote has it that “making predictions is hard, especially about the future”. This couldn’t be more true.

When we sense-checked every list of 2025 predictions we could find through the years - across books, magazines, and articles - only a couple got the majority of their predictions correct, and even then just barely. 

If you see any list of predictions, you can safely assume most of them will be wrong. 

This has practical implications. Some advocate that as the future is so inherently unknowable, businesses should have an “emergent” rather than a “deliberate” strategy, responding to events as they happen, rather than planning in advance. 

2. Experts are the worst predictors.

One of the foremost authorities in sense-checking predictions, professor Philip Tetlock, famously showed that experts were often no better at predicting future events than random choices - as he put it, experts were worse forecasters than dart-throwing chimps.

We saw a similar pattern in our own research. Groups of experts and CEOs had a terrible track record at predicting 2025. On the other hand, two of the most successful were Thomas Harper (a scientist with a blog) and the respondents of a Gallup poll from 1998, asking ordinary Americans what they thought would happen in 2025. 

Ironically, futurists are some of the worst people at predicting the future. 

3. The closer you are to a year, the worse you tend to be at predicting it.

Before carrying out this project, one of our initial hypotheses was that newer predictions would be better, and older predictions would be worse. This would make intuitive sense - the closer you are to 2025, the less chance there is of things changing.

In fact, the opposite is true. Predictions in the 90s tended to perform above-average, while one of the most recent sets of predictions, from 2020, got nothing correct. 

The reason behind this is a common flaw of the predictive arts - to focus too much on what’s currently happening, and project outwards from there. So many predictions from the early 2020s had their visions of the future consumed by Covid-19, which has had important lasting effects on the world, but not as many as originally thought. 

4. You might be able to predict a trend happening, but you probably won’t predict how it happens.

The textbook case here is working from home, one of the true lasting impacts Covid has had on the world. Increased levels of this were actually one of the more common predictions we found. In 2000 BT estimated that 25% would work from home, while a more recent forecast from Tata Consultancy Services put the figure at 40%. Our most recent research puts the figure at 37% working remotely in a typical week, so TCS was pretty much bang on. 

Its forecast came in late 2020, but ones made prior to that didn’t anticipate that widespread working from home would come from a shock event like a global pandemic - most felt it would be a natural byproduct of catering to millennials in the workforce. It’s a reminder of one of Nassim Nicholas Taleb’s most salient points in his classic treatise on unexpected events, “The Black Swan”: history doesn’t crawl, it jumps. 

5. Just because you foresee a trend, doesn’t mean you’ll act on it.

One of the few things you can predict in advance with accuracy is demographics. Birth rates give you a solid idea of how many people will be alive at a point in the future, and the average age of the population.

So many people correctly foresaw a 2025 with aging populations across the world, and it was a natural progression to think marketers would surely cater their efforts to this growing market, one that would control an awful lot of spending power.

This hasn’t really happened. As we see in our research, older consumers are massively overlooked in marketing. In a chapter of his book “Why We Buy”, looking forward to an imagined 2025, Paco Underhill predicted that with more older consumers in the market, and thus a higher level of visual impairments, any text “smaller than 13 point type will be a form of commercial suicide”. 

Anyone that’s been to a dimly-lit restaurant recently would surely disagree. 

6. Headsets have not replaced smartphones.

One of the most trivial observations you can make about life in 2025 is that almost everyone has a smartphone, and they use it a lot. But this flies in the face of lots of predictions, even some made relatively recently. 

The futurist Ian Pearson said “if you have a smartphone [in 2025], people will laugh at you”. Jon Peddie said “it is entirely possible we’ll have no need for a smartphone”, while Robert Scoble and Shel Israel more explicitly named the cause underpinning all these predictions - “headsets will replace handsets as the primary device for most people” in their book “The Fourth Transformation”. 

All of these predictions were made in 2016 or 2017, but now 2025 has rolled around, only 1 in 20 consumers globally own a VR headset, to say nothing of AR devices. 

Just after the breakout success of Pokémon GO, many predicted some kind of AR device would take over from the smartphone as people’s go-to. But this vastly overrated how quickly such a change could happen, and underestimated the fact that the smartphone market was still growing.

7. The internal combustion engine is sticking around longer than we thought.

If you’re in the market for a new car in 2025, you’re just as likely to plan on buying a diesel car as an electric one (both 21%). To repeat, in 2025 cars powered by diesel - a polluting fuel with huge reputational issues - are as popular as EVs. 

The UK government’s recent relaxation of its mandate around petrol and diesel vehicles is a good example of how the transition in the car market has been slower than many forecast (one Stanford professor thought that no ICE vehicles would be sold this year at all). 

Even further from the mark were many predictions around self-driving cars. While Waymo is expanding, we’re far from seeing them be “commonplace in society”, as predicted in the 2017 book “What Will Your Grandchildren See When They Look Up?”.

8. We focus too much on technological trends, and overlook social trends.

Most futurists’ vision of the future is inspired by technology. Products that could be made based on current trends, or that currently exist in patents and prototypes.

This tends to miss two big things. One is that the future is defined by consumers, not producers. Just because a technology can be made, doesn’t mean that people will want to buy or use it. A great example is Concorde. We have the technical capability for supersonic flight, but at the time of writing there’s no market for it. This is the central insight of Steven Schnaars’ classic prediction-checking text Megamistakes: by falling in love with the possibilities of technology, we forget the market it’s intended to serve. 

The other thing this emphasis on tech misses is the importance of social trends. In his book “Welcome To The Future Cloud”, a compendium of predictions for the year 2025, the Dutch futurist Marcel Bullinga made some overambitious calls, such as:

 “[illnesses will be] discovered long before the symptoms arise. Showing your personal DNA chip to your doctor [will be] standard practice”. 

But unlike many of his futurist peers, he astutely observed the social tensions that would arise from increased immigration to Western countries. 

Sometimes these trends actually manifest as backlash aimed at technology. People who anticipated 5G to have a big impact on the world probably didn’t factor in people opposing masts being built in their neighborhoods. 

9. Nothing ever really dies.

People often imagine the future in terms of endings. Many people have a view of technology where the new thing “kills” the old, but this isn’t usually what happens. Those old things can still persist in niches or among laggards (like fax machines), or be revived to counterbalance new tech (like vinyl and film cameras). 

A better way to analyze technological transitions isn’t asking how the new invention will replace the old, but what use-cases each will be best at. 

10. Sometimes 550-year-old technology still wins.

When the Amazon Kindle became established as a consumer product in the early 2010s, it sparked a host of predictions about the future of publishing, many of which were very pessimistic about print. John Biggs felt that “[by 2025] all publishing will exist digitally”, while The Bookseller magazine, analyzing the results of their Digital Census in 2014, thought 2025 would mark the point when digital sales of books would overtake print.

But paper books are in rude health. And not only does our research show UK consumers prefer reading print books to ebooks (or audiobooks for that matter), that holds true even among people who buy ebooks. People who read books electronically still actually prefer the paper versions, on the whole. 

This is another often forgotten lesson about transitions in technology - sometimes the “old” version is better in important ways. Print books are easier to lend to a friend, they don’t need battery power, and they’re less expensive to lose. When a new invention comes along, it’s always important to ask how it might be worse than the current alternative, rather than assuming it’s better in every way. 

11. Not everything has gone digital.

The anticipated shift to ebooks was one part of a bigger tendency among 2025 predictions - a sense that with more of the world’s infrastructure becoming digitized, everything else would follow suit. But there have been two big exceptions to this - currency, and voting.

Many forecasters anticipated that physical currency would cease to exist, and in some cases expected a single, global currency would take over. While contactless and mobile payments have rapidly expanded, 4 in 10 consumers globally still prefer paying with cash. 

Another very common prediction was that all voting would become digital. While there’s been movement in both areas, we’ve collectively learned that the old-fashioned, physical version is often superior, and more secure.

12. As veganism shows, we don’t always do “the right thing”.

In a 2019 report looking at the future of food (fresh off the back of the launch of Greggs’ vegan sausage roll, of course), British supermarket Sainsbury’s predicted that in 2025, a quarter of the UK population would be vegetarian (including vegan), and half would be flexitarian.

Our research shows that those figures are actually 6% and 9% respectively, and the number of UK vegans has actually decreased in recent years.

As well as once again overestimating the speed of change, this highlights another common flaw with predictions. Often we’re guilty of conflating our idea of a virtuous future with what people will actually do. As the veganism case study shows, focusing too much on a moral imperative makes you overlook the things that really matter to consumers, like price (especially given this decade’s issues with food inflation), culture, and health (with vegan foods being targeted for their use of ultra-processed ingredients). 

13. We tend to predict utopias or dystopias, but what happens is the messy middle between the two.

Part of the UK Government Office for Science’s Futures Toolkit says the following:

 “A common challenge with Futures thinking workshops is focusing on overly utopian or dystopian perspectives.

The world in the future will most likely be a mix of appealing and disagreeable things – as it is today. If the discussion seems to be heading to either extreme, then asking “what is the upside/downside of this situation” is a way of changing the dynamic.”

And this is an insight not taken into account by enough forecasters. A handy way to read many future predictions isn’t actually as a set of predictions, but a reflection of things the author likes and dislikes.

In his book “The Future”, depicting a world 100 years from his own time, Archibald Low made some uncannily accurate calls, but also fell victim to this. He couldn’t understand the popularity of football, still the world’s most popular sport, saying:

“Imagine a really intelligent, thoughtful man - and future education will make men thoughtful - kicking a ball about in a field for a living!”.

This is a valuable lesson for marketers in particular - you have to understand the future in terms of the aggregate preferences of the market, not your own worldview. Just because you don’t like something, doesn’t mean it’ll become less popular. 

14. We carry the past with us into the future. 

The media theorist Marshall McLuhan once said “we walk backwards into the future”. Unlike many others who’ve written about the future, he recognized a key human tendency, to look back as we move forwards. 

One of the best examples of this tendency being called correctly lies in an otherwise throwaway comment in a letter to The New York Times from 1925 looking at the etymology of the word “Beefeater”: 

“Do not hurry, those beefeaters will be there on guard in 2025 A.D.”

Granted, this is not an official prediction per se. But unlike many futurist forecasts, it recognizes that many people see value in tradition and cultural continuity; it recognizes that when we enter the future, we carry the past with us. 

15. Satire is often more realistic than sci-fi.

Science fiction is often written with a mindset of describing a technologically plausible view of the future. Within that there are some hits, but a lot of misses too. 

But in capturing the essential flavor of human life in the future, satire is often more accurate, because it’s grounded in the daily frustrations and oddities of life. These are unlikely to go away in the future, yet many people treat forecasts as a kind of tidying-up exercise. It becomes a projection of things they want to disappear or be made more convenient - like traffic, cooking, or disease. 

A good example is Dan Erck imagining a day in his 2025 life all the way back in 2001. Fair enough, moon holidays and holograms haven’t taken off, but there’s a sense of life’s essential friction he captures better than most, including around technology itself.

16. A lot of future visions are actually quite old.

One of the most eye-opening insights of this project was just how similar many visions of 2025 looked, whether they came in the 1980s or the 2010s. Common tropes included: touchscreens, body implants and wearable technology, self-driving cars, augmented reality, virtual reality, holograms, 3D printing, and genetic engineering. 

The future allows us to imagine anything, but we often imagine the same things. In practice, “the future” is a genre with a set of standardized signifiers. 

17. We forget the fundamentals of consumer behavior. 

In our Consumer Tech data set, we ask consumers in the market for particular tech products what would most motivate them to buy one. Some options, like graphics quality for games consoles, are distinct to the category. But others are fairly universal - like price or processing speed.

If we take these options that exist across multiple categories and calculate their averages, we come up with a useful rule-of-thumb list for what matters most to consumers when adopting new technology in general. And none of these should come as a surprise, but in the context of predictions, they’re often forgotten. 

Three elements are key: price, speed, and comfort (which can include form factor). Battery life is also really important. You might invent the most impressive VR headset in the world, but if it costs $10,000, takes a minute to boot up, feels heavy on the face, and runs out of juice after half an hour, no-one’s going to buy it. 

18. Some things never change.

To some degree, making any set of predictions is a futile exercise. Not just because it’s so hard to anticipate how the world will change, but also because the number of things that change are vastly outnumbered by the things that don’t. 

If you make a set of predictions that only focus on what will change, you’re almost guaranteed to overestimate how much change will happen. Some of the most convincing depictions of 2025 in the world of fiction come in books that make no attempt to describe speculative technology at all. 

As Philip Tetlock found in his classic research, a simple algorithm of predicting no change beats many experts a lot of the time. And Thomas Harper, the most successful forecaster we identified, was notable for the number of correct calls he made about things that wouldn’t change. Inertia is a powerful force. 

Just look at The Fast and The Furious film franchise, which is a contender for the most consistent data point in our research. Since we began tracking it in 2018, the share of UK consumers who call themselves fans of it has never risen above 19%, or fallen below 18%. No vision of the future would be complete without acknowledging this fandom, it seems. 

The more things change, they more they stay the same

Ultimately, the main conclusion to draw from sense-checking a pile of predictions is that less changes than we think, and it happens more slowly. This insight is always relevant, but it’s extra important to remember when at the time of writing, many are currently making some far-reaching predictions about the disruption that will be caused by AI. And what’s more, the biggest changes are the ones no one saw coming. 

The next big thing is often the thing that happens when you think you’re looking at the next big thing. 

If anything, we could do with more marketers as part of the broader conversation that imagines our collective future. Futurists have their place, but we need people who specialize in understanding all the inherent weirdness of human behavior, and its continuity through time. 

A (curated) 2025 reading list

Mike Walker - 2025 (radio play)

Verner Vinge - Rainbows End

TIME Magazine - Visions of the 21st Century

Stephen King - The Running Man

Archibald Low - The Future

Le Temps - A high-tech day in 2025

Erwin K. Thomas & Brown H. Carpenter (ed.) - Mass Media In 2025

Paul Fairie X thread of 2025 predictions

Thomas Harper - Predictions 2015-2025

BBC recap of Tomorrow’s World predictions

World Economic Forum - Deep Shift

Kenneth Sylvan Guthrie - A Romance Of Two Centuries: A Tale Of The Year 2025

Marcel Bullinga - Welcome To The Future Cloud: 2025 In 100 Predictions

Joseph F. Coates - The Highly Probable Future: 83 Assumptions About The Year 2025