The Fishbowl Problem: Why Talent Acquisition Predictions Are NOT a Thing

Every December and January, without fail, my LinkedIn feed floods with confident predictions about the future of talent acquisition.

By December following, most of them will be quietly forgotten, filed away in the internet's bin alongside all the other expert forecasts that never quite materialised.

I've read enough of these to recognise my own eye-roll when I see them.

Someone in some world "capital of innovation" publishes their 10 bold predictions for hiring in 2026. They cite AI adoption rates, skills-based hiring trends, and the death of the CV (again). They write with the certainty of someone who knows something the rest of us don't.

My issue is that… they don't.

Talent acquisition predictions are like weather forecasts — except way less accurate

Meteorologists have satellites, supercomputers, and decades of atmospheric data. They can tell you with reasonable confidence whether it will rain tomorrow in Frankfurt. Ask them about next week and the accuracy drops. Ask them about three months from now and you're essentially flipping a coin.

Philip Tetlock spent two decades studying how well experts predict political and economic events. Between 1984 and 2003, he collected roughly 28,000 forecasts from 284 experts across various fields (Tetlock, 2005). His finding: experts performed barely better than random chance. In fact, a simple extrapolation of recent trends often outperformed the confident predictions of seasoned professionals.

Weather systems are chaotic but they follow physical laws. Talent acquisition operates in an infinitely more complex environment. We're dealing with human psychology, economic shifts, regulatory changes, technological disruption, and geopolitical events, all interacting in ways no one can model.

If meteorologists with all their tools can't predict weather three months out, why do we think HR thought leaders can predict hiring trends 12 months ahead?

So my take? We do a lot of this for commercial reasons. Companies want certainty, consultancies sell the s**t out of it, and so the prediction industrial complex churns on, year after year, mostly wrong but always confident.

Let's talk 2020 — and how useless our predictions were

There is no more fitting recent example than the pandemic. In early 2020, talent acquisition predictions focused on the skills shortage, the gig economy, and how AI would streamline recruiting (SmartDreamers, 2020). Then March happened. Suddenly, 38% of companies put hiring on hold entirely (Alexander Mann Solutions, 2020). The predictions about talent shortages became irrelevant overnight.

As the pandemic wore on, new predictions emerged. Experts confidently proclaimed that remote work was here to stay, that companies had finally learned flexibility, that the office was dead. By May 2020, consultants were predicting permanent shifts to distributed teams (Cielo Talent, 2020).

Fast forward to 2023. Mass tech layoffs. Over 262,000 tech workers lost their jobs that year (TechCrunch, 2024) — many at the same companies frantically hiring remote workers just two years earlier. The narrative flipped from "we can't find enough talent" to "we overhired during the pandemic" in 18 months.

None of the 2020 predictions saw this coming. None of the 2021 predictions anticipated the 2023 bloodbath. The experts were as blindsided as everyone else.

The fishbowl problem: why most predictions are dangerously narrow

This is what really gets to me.

The people writing these predictions often operate in a bubble. They write about the companies they know, the people they interview, the markets they understand. Then they slap a global headline on it and call it a universal trend.

When you're swimming inside a fishbowl, you genuinely believe you're seeing the whole ocean.

A prediction about AI adoption in recruiting written from Silicon Valley assumes every company operates like a well-funded American tech firm. It ignores that most companies in Europe are navigating GDPR compliance, the EU AI Act, and pay-transparency directives that fundamentally change how hiring technology can even be deployed.

It ignores that risk tolerance varies wildly across cultures. German companies approach AI differently than Australian ones. Japanese hiring practices bear little resemblance to Brazilian ones.

The predictions treat hiring as if it's a universal experience. As if regulatory environments don't matter. As if cultural norms around risk, hierarchy, and employment relationships are irrelevant. As if the whole world is just San Francisco with different accents.

They aren't lying, exactly. They're just describing their own fishbowl and assuming it's the sea.

Same "predictions" — for over a decade

• Skills-based hiring? We've been talking about that since 2017 minimum.

• Recruiters need to be more strategic? Peak 2015.

• Talent advisory as the future of TA? I've read that one every January for 10+ years running.

Meanwhile, recruiters are still struggling with the basics. They don't know how to manage a difficult stakeholder. They can't run an effective intake meeting. They're still writing job descriptions that read like legal documents.

The fundamentals are broken, but we keep predicting the shiny future whilst ignoring the messy present.

• Real predictions would acknowledge that most companies are nowhere near skills-based hiring because their data infrastructure is rubbish 🗑️

• Real predictions would admit recruiters can't become strategic advisors when they're drowning in admin work with no process improvements in sight.

• Real predictions would say: most of you will still be doing the same things badly in 2027 because you haven't fixed the foundations.

We should be taking snapshots of what is actually happening right now. Not what thought leaders hope will happen, but what's measurably true across companies, geographies, and market conditions.

Stop asking "what will the future look like?". Ask:

• What patterns are emerging in pockets of the industry right now?

• What conditions enable those patterns?

• What barriers prevent their spread?

• Which of these patterns will scale and which will remain niche — given my unique set of rules and challenges?

The research on expert predictions should humble you

Tetlock's work offers another damning insight: the more famous an expert, the worse their predictions (Tetlock, 2005).

Media-friendly folks who make bold, confident predictions tend to be wrong more often than quieter, more cautious experts. The qualities that make someone good at prediction (nuance, hedging, admitting uncertainty) are exactly the qualities that make them bad at getting attention.

This creates a perverse incentive structure. To get noticed, you need to make bold predictions. To make bold predictions, you need to ignore uncertainty. And to ignore uncertainty, you need to either be overconfident or willing to be wrong in public, repeatedly.

The industry rewards the wrong people. Again.

The confident prognosticators get the conference stages and the LinkedIn engagement. The careful analysts who say "it depends on numerous factors including regulatory environment, macroeconomic conditions, and technological adoption rates" don't make for good headlines.

Kahneman and Tversky's research on the planning fallacy shows that we consistently underestimate how long things will take and overestimate how smoothly they will go (Kahneman, 2011).

What should we do instead of making talent acquisition predictions?

If predictions are useless, what's the alternative? Do we just throw up our hands and admit we have no idea what's coming?

Not quite. We move towards something more manageable and real: adaptability. From "here's what will happen" to "here's how we'll respond when something does happen."

Build for resilience

• Instead of trying to predict whether remote work will stick or AI will replace your recruiters, build hiring systems that can flex.

• Train your team in various methodologies.

• Create processes that can scale up or down quickly.

The goal shouldn't be to predict which scenario will unfold — it should be to be ready for several different scenarios.

Focus on fundamentals

Some things don't change regardless of trends. Treating candidates with respect works in any market. Clear job descriptions matter whether AI screens them or humans do. Fast, fair processes beat slow, opaque ones every time. These aren't sexy predictions, but they're reliably true.

Watch for weak signals

Pay attention to what's actually happening in your own market, your own industry, your own company.

• If your offer acceptance rates drop, that is your data. Question: do you even track your data? Is it clean? Usable?

• If candidates start asking different questions, also a signal.

These small, local indicators matter more than global trend reports.

Scenario planning over prediction

Instead of predicting "AI will automate 60% of recruiting by 2026," consider multiple scenarios. What if adoption is faster than expected? What if it's slower? What if regulation kneecaps it entirely? Sketch out how you'd respond to each.

Accept uncertainty as a feature, not a bug

We operate in a fundamentally uncertain environment. New technologies emerge. Regulations change. Economies shift. Pandemics happen. No amount of prediction will change this. The companies that thrive aren't the ones who predict the future perfectly; they're the ones who adapt quickly when the future arrives differently than expected.

I think Talent Acquisition deserves better

I'm not arguing that we should never think about the future or that all analysis is pointless. I'm arguing for intellectual humility. For recognising the limits of what we can know. For being honest about uncertainty rather than hiding it behind confident predictions.

When someone publishes their predictions for 2026 — next up: 2027 — ask yourself: what's their track record? Did their 2024 predictions come true? Their 2023 ones? If they were wrong before, why should we trust them now?

And crucially, do their predictions account for regulatory differences, cultural contexts, and the full spectrum of how hiring actually works outside their immediate circle?

The next time you read a headline that promises "10 Talent Acquisition Trends That Will Define 2027," remember that the person writing it has no more access to the future than you do. They're swimming in their fishbowl, extrapolating from their local data, and hoping for the best. The difference is they're confident about it.

Perhaps the best prediction I can make is this: most predictions about talent acquisition will be partially wrong, the confident ones will be the wrongest, and by this time next year, we'll all have forgotten what anyone predicted anyway.

The sooner we accept that, the sooner we can focus on building hiring functions that are flexible and resilient enough to survive the next round of predictions.

Some companies will build integrated talent ecosystems (where AI handles tactical work, recruiters design systems, and CFOs view TA as strategic investment). Others will still be using spreadsheets and hoping for the best.

By January 2027, you'll read about how skills-based hiring is finally happening, how recruiters need to be more strategic, and how AI will transform everything — just like you did in 2022, 2024, and 2025. And most recruiters will still be dealing with hiring managers who change requirements mid-process, candidates who ghost them, and systems that don't talk to each other. Because predictions don't fix fundamentals.

Process redesign does. Capability building does. Honest assessment of where you actually are does.

We keep predicting the future we want instead of preparing for the one we're likely to get, which looks a lot like the present, only with more expensive tools and the same unresolved problems.

References

• Alexander Mann Solutions (2020). New Research Reveals Impact of Covid-19 on Global Talent Acquisition. Globe Newswire.

• Cielo Talent (2020). 5 Predictions for Talent Markets After the Pandemic.

• Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.

• SmartDreamers (2020). 7 Emerging Trends in Talent Acquisition in 2020.

• TechCrunch (2024). A Comprehensive Archive of 2023 Tech Layoffs.

• Tetlock, P.E. (2005). Expert Political Judgment: How Good Is It? How Can We Know? Princeton University Press.

• Tetlock, P.E. & Gardner, D. (2015). Superforecasting: The Art and Science of Prediction. Crown Publishers.