Capability Tells You Who Can. Desire Tells You Who Cares.

Your systems know a scientist's title and whether their calendar is full. What they are good at lives in people's heads. What they love doing is nowhere at all. The last one decides the quality of outcomes.

Two scientists sit in the same dose-expansion meeting for the same first-in-class program. One is the clinical pharmacologist who owns the PK model. The other is the translational scientist who owns patient stratification. On paper, the program has exactly the expertise it needs in the room.

She, the translational scientist, fought to get onto this molecule. It targeted a disease that had taken someone in her family, and she had read every readout from the first-in-human cohort the way you read something that matters to you. He had landed on the program in a Tuesday resourcing session, because he was the pharmacologist who happened to be free that quarter. Competent, well reviewed, no complaints.

The protocol called for a single expansion dose across all cohorts, a reasonable plan the team had signed off months earlier. Working a weekend nobody asked her to, she had found that a biomarker-defined subgroup, the very patients the molecule was built for, would likely need a different dose than the plan assumed. She pushed to split the expansion. He saw no reason to reopen a settled question, and his model supported the single dose.

The dose that program carried into expansion did not come down to who held the right title. It came down to who cared enough to interrogate an assumption everyone else had already accepted.

Same meeting, same data, two seats at the table with different decision drivers. What separated these two was not their titles or their training. She wanted to be there. He was available to be there.

The resourcing system did not see that. It saw little. It held their titles and their calendars. What they were actually good at lived in a manager's memory. What they wanted to do lived nowhere at all. In life sciences, where the gap between competent and exceptional is paid for in months that patients spend waiting, the variables nobody captures are the ones that matter most.

Solved at the Top, Broken at the Bench

Ask a CEO whether her people are working on the right things and she will say yes. A talented organization, a portfolio review every quarter, an engagement survey that comes back fine. Alignment looks solved.

It looks solved because the problem never reaches her. The mismatch lives two or three levels down, at the study level, where a functional leader opens a spreadsheet on a Friday and staffs the next experiment from the names he knows and trusts. He works from the names he knows because no system can show him the ones he does not but should. Proximity and comfort, not merit and fit, determine who gets assigned to what. By the time the consequences surface, they read as a timeline slip, a hard program, a market that moved, never the staffing decision that started it, which no report ever tied to the outcome.

When R&D leaders do feel the friction, many file it under HR, or under org design. It is neither. It is an operations problem, and it bites hardest in research, where the assumption that one person in a role equals another breaks down fastest.

A medicinal chemist whose pattern recognition fits one chemotype is not interchangeable with an equally senior chemist who has never worked that structural class. A clinical pharmacologist owns a first-in-human dose decision. A translational scientist decides whether a biomarker predicts clinical benefit. These are chokepoints. The judgment is irreversible once acted on, invisible until it fails three steps downstream, and impossible to reconstruct from the documentation later. Put the wrong fit on one of them and the probability of success erodes quietly, surfacing years later as a program that cannot be recovered.

Skills, experience, expertise, bandwidth, and aspiration are unique to every scientist. The gap between what a role-level forecast promises, which is an estimate and never a calculation, and what the named person delivers shows up as missed milestones, last-minute replans, and forecast errors. In early research the bottleneck is rarely headcount. It is the availability of the one specialist the work actually needs.

Hiring selectivity is necessary. It is never sufficient. The match has to be remade with every assignment. No one is equally good at everything. Many poor performances are simply consequence of poor work-worker mismatch.

Skill Gets You Into the Room

From the beginning of this series, I have argued that you cannot manage headcount as if it were capability, and that the right person on the right task beats the merely available one. Even if we take that as settled, this facet sits underneath that. Desire is not a substitute for skill. It decides between the people who already have it.

Let's grant the resourcing system its best case for a moment: that it knew exactly who was qualified, who was free, who had done this kind of work before. Now, most organizations are nowhere close to this. But even that ideal system would be blind to the variable that decides how the work actually gets done.

Four decades of research point at the same finding. In a meta-analysis spanning forty years, Cerasoli and colleagues found that intrinsic motivation, the pull to do the work for its own sake, predicts the quality of performance, while external incentives mostly predict the quantity. Pay moves output. Wanting moves judgment.

Teresa Amabile's model of creative problem-solving separates two things most resource plans collapse into one. Domain skill is necessary. So is intrinsic task motivation, a distinct ingredient, not a byproduct of competence. A chemist can have every skill an experiment requires and still approach it the way anyone approaches a chore: carefully, correctly, without the spark that notices the anomaly nobody scheduled.

This is what those two scientists showed. The one who wanted the program worked the exposure data on a weekend because she could not help herself, and that is precisely why she saw the subgroup the plan would have missed. Desire is not enthusiasm. It is attention. The person who wants the work pays closer attention to it, and in research, attention is where the early signal lives.

The cost of missing this is lopsided. A competent, available scientist on a poor-fit assignment does not fail loudly. The program advances, milestones get hit, mostly. What you lose is the upside: the earlier catch, the braver dose call, the leap that turns a stalled series around. You never see the outcome you did not get.

Multiply that across hundreds of programs and the invisible delta becomes the line between an organization that executes and one that compounds.

Experience Can Be the Wrong Tiebreaker

When allocation logic reaches past availability, it reaches for experience. The most seasoned person is the safest match, so they get the work. Usually that is right. Sometimes it is exactly wrong.

A program moves into cell and gene therapy work. Two scientists are in frame. One is a senior expert who has run this play ten times and could do it in her sleep, and is bored by it, angling for something new. The other has never done it, but wants it badly, has the adjacent skills to learn fast, and has been asking for the modality for a year. The capability math picks the expert every time. The fit, once desire is in the equation, often points the other way.

Both assignments can be defended. The expert delivers a clean, predictable result. The learner brings the attention that comes with wanting it, and does the work with something the expert lost three programs ago. The learner does not always win. But he is usually never even considered, because the one fact that would surface him, what he is trying to become, has never lived anywhere the staffing decision can see it.

None of this means desire outranks competence. A scientist who wants the work but cannot do it is not a match. The claim is narrower: among the people who can credibly do the work, the one who wants it brings a different quality of attention. And where the skills are adjacent and the risk is bounded, desire is the signal that a supervised stretch is worth taking, the assignment that builds energy now and capability later.

Career aspiration gets treated as an annual-review topic, filed under development, disconnected from Tuesday's resourcing call. But it does belong in the resourcing call. Growth intent is allocation data: who wants to stretch into what, and who is ready.

An organization pours months into hiring the best people, then assigns them by who happens to be free. The aspiration that made someone worth hiring goes unused the moment they walk in the door.

And the cost lands twice. The bored expert is how you lose your most experienced people. The overlooked learner is how you fail to grow your next ones.

The Question You Could Never Afford to Ask

Asking every scientist, every quarter, what they want next and what they are tired of, then having that answer in hand when you staff a study, was never realistic in an organization of thousands. The cost of capturing it by hand exceeded the value. So the question went unasked. It stayed in people's heads, and in the occasional career conversation no resourcing decision ever reached.

That constraint is gone. Capturing intent lightly and continuously, then putting it in front of whoever makes the assignment, is now an ordinary thing a system can do. The learner who has been asking for the cell and gene program for a year is no longer invisible on Tuesday. His wanting is finally something the assignment can see.

This is the problem I have spent fifteen years working on. The technology only removes the excuse. What remains is the decision to treat what people want as real allocation information rather than a year-end-review topic.

Who Stays

This is the variable that matters most now. Voluntary turnover in biopharma rose from 13.7% to 15.9% in a single year, and on a five-point scale, how strongly professionals expect to stay fell from 3.51 to 2.75. People are not only leaving. They are quietly deciding they might.

As the industry reshores manufacturing and pulls development work back inside, the binding constraint shifts from capital to people, and the people you most need have the most options. You do not keep them with another engagement survey. You keep them by putting them on work they want to do, often enough that staying feels like growth instead of waiting.

That is what makes desire a retention moat, if the organization can operationalize it. Skills can be hired. Experience can be bought. But a system that repeatedly matches scientists to work they care about builds a reason to stay that compensation alone cannot copy.

I recently put a question to a CEO: if you asked your people privately, how many would say they love the work they are on right now? Most leaders assume the number is high. The ones who ask are usually surprised. The closer the match between a scientist and the work, the better the outcome. Humans are like that.

And the outcome is never only the scientist's. Every poorly matched assignment is a program moving slower than it could have, and in this business a slower program is a patient waiting longer for a medicine that already exists in a lab. The medicine does not care who is assigned to it. The timeline does. And the timeline is decided, one assignment at a time, by whether the person doing the work wanted it.

‍

References

Cerasoli, Nicklin & Ford, "Intrinsic Motivation and Extrinsic Incentives Jointly Predict Performance: A 40-Year Meta-Analysis," Psychological Bulletin (2014). pubmed.ncbi.nlm.nih.gov

Teresa M. Amabile, "Componential Theory of Creativity," Harvard Business School Working Paper 12-096 (2012). hbs.edu

Pharmaceutical Technology, "State of the Bio/Pharma Workforce: A Comparative Analysis of Employment Trends and Industry Sentiment." pharmtech.com

Further Reading

Headcount Is Not Capability: Why Pharma Keeps Hiring and Firing Its Way to the Same Problem. On why adding people does not solve a matching problem.

People Are Not Your Greatest Asset. On how allocation breaks down when fit gives way to availability.

Estimating Is Not Calculating: The Resource Precision Gap Between Planning and Execution. On the difference between a role-level estimate and a task-level calculation.

About Me

I'm Andy, pharmacist, MIT engineer, 25 years in life sciences operations. I started Unipr because I watched too many medicines get delayed by operational barriers, not science failures. We support 100+ pipeline programs at top pharma and biotech across 30+ countries.

Aligning talent to outcomes has been on my mind for fifteen years: what it would take to put the right person on the right work, every time, accounting not only for what they can do but for what they want to do. If that question is live for you too, I would welcome the conversation. Learn more at unipr.com.

‍