The Data That Feels Safe
Why historical outcomes are an unreliable basis for future career decisions
A parent called recently about her daughter's career. The direction was already decided - health sciences in Australia. When I asked what had led them there, the answer was immediate: an aunt, already settled in Australia, working in the health sector, earning well. That was the evidence. That was the entire basis for a decision that would shape the next decade of her daughter's life.
The aunt's outcome was real. Her earnings were real. Her settlement was real.
But her outcome was built in a different moment, under different visa conditions, in a different hiring environment, with a different skill demand profile. None of that was part of the conversation. The historical data felt sufficient. It was not.
This is Backward-Looking Risk Assessment. It is not a failure of intelligence or care. It is a failure of framework.
Every family making a career decision is conducting a risk assessment. The inputs they use are almost entirely historical - what worked for a sibling, what the neighbour's son earns, which stream produced stable employment in the past decade, which country was giving visas last year. This feels rational. Past outcomes are visible, verifiable, and trusted.
The problem is not the instinct to use data. The problem is that historical outcomes capture what the market rewarded then - not what it will reward when the student graduates, works, and builds a career over the next thirty years.
Corporate history makes this pattern visible at scale.
Nokia dominated global mobile phone markets with products that were, by every historical measure, successful. Blackberry owned enterprise communication. Kodak built the camera industry. Each organisation used its historical position as the primary input for forward decisions. The data said they were safe. The market had already moved.
None of them failed suddenly. The signals were present and readable for years before the collapse. The assessment framework was simply not designed to read them. It was designed to confirm what had already worked.
Career decisions follow the same pattern. A family that watched engineering produce stable employment for fifteen years applies that data to a decision being made in 2026 - without asking whether the conditions that made engineering stable in 2010 still hold in 2030.
The data shows how fast those conditions are actually shifting.
According to the World Economic Forum's Future of Jobs Report 2025, 39% of workers' core skills will change by 2030. A career decision made today on the basis of what was rewarded five years ago is being made on evidence that will be partially obsolete before the student graduates.
The India Skills Report 2026 makes this concrete. AI and ML engineers recorded 600% growth in jobs. In the same period, MBA graduate employability dropped from 78% to 72.76%. The historical safe choice is declining. The forward-looking choice is growing at a rate no backward-looking assessment would have predicted five years ago.
According to the Indeed-NASSCOM Report 2026, 40% of Indian employers now prefer demonstrable AI skills or certifications over degrees. The credential that historical risk assessment identifies as safe is losing ground faster than most families can track.
The market has moved. The assessment framework has not.
In India, this gap is structurally wider than in most comparable economies.
The dominant information channels in Indian family decision-making are backward-looking by design. Peer networks, family referrals, and community word-of-mouth transmit historical outcomes - what worked for someone else, in a different moment, in a different market context. These channels are trusted precisely because they carry lived proof. But lived proof is always historical. It cannot carry what is emerging.
This compounds what earlier essays in this series have examined - the conditioning that makes familiar paths feel safe and unfamiliar ones feel risky, and the structural absence of a neutral decision framework that helps families evaluate evidence independently. The backward-looking information channel and the conditioned preference for the familiar reinforce each other. The result is a decision environment where the past is everywhere and the future is absent.
In study abroad guidance, this pattern appears with particular clarity. When the UK was processing visas smoothly, families applied to the UK. When Canada opened, everyone moved to Canada - including students whose academic profiles pointed clearly toward Australia, New Zealand, or the United States. The decision was not based on fit. It was based on which country was currently producing the most visible positive outcomes in the immediate network. Canada has since tightened significantly. The families who followed the historical trend are the most exposed.
Forward-looking assessment is not prediction. It does not require certainty about what the future holds.
A family that waits for certainty before making a forward-looking decision will always default to historical data - because historical data is the only thing that feels certain. The first step in a better assessment is accepting that certainty is not available, and that the best decision possible under uncertainty is structurally superior to the most comfortable decision possible under false security.
It requires three inputs alongside historical data. Present signals - what skills and roles employers are demanding now, not five years ago. Forward probabilities - which industries, technologies, and roles are structurally growing versus declining over the student's career horizon. And individual profile - what the student's actual capability, interest, and temperament point toward, independent of what has historically worked for someone with a similar background.
Historical data belongs in this framework. It carries weight. It does not carry the decision.
The aunt's outcome in Australia is relevant information. It is one data point from one moment. It is not a guarantee, a blueprint, or a risk assessment.
The parent who called had already made the decision before the conversation began.
That is the cost of backward-looking risk assessment. Not a wrong answer, necessarily. But a decision made without the inputs that actually determine whether it will hold.
The market does not wait for assessments to catch up. It moves. The question is whether the framework used to make career decisions moves with it.
Related reading: Decisions Are Not Chosen. They Are Absorbed.
Related reading: The System Isn't Broken. It Was Never Built.