Hiring Tests People In Context, Then Performance Reviews Strip The Context Away

A hiring decision usually rests on forty-five minutes of conversation with someone who spent the previous week preparing for it. A performance decision, months later, rests on a rating assigned by a manager looking at an outcome after the fact. Somewhere between those two moments the organization changes what it is measuring, constructing a situation to see how a candidate works inside of it, then scoring that same person on a result with no record of the situation that produced it.

Partho Mukherjee was formerly an Human Resources Business Partner at Flipkart, supporting a workforce of more than 7,000 across industrial relations and logistics HR, with a background spanning HR transformation and cost optimization. He came into the field from psychology, and he has sat on the hiring side of hundreds of interviews. What he says conventional interviews actually produce is a performance, rehearsed in advance and delivered on cue.

"Candidates come with their own scripts, thoroughly practiced. They tell you only those things," Mukherjee says. Preparation can look like capability from across the table, which is what he designs his own process around.

The interview builds a situation on purpose

Mukherjee skips the fixed question set and puts a candidate inside a problem to watch what they do with it. "My questions are always situation-based and case-based. I allow a candidate to perform in a given situation and then observe," he says. Varying the pressure is part of the design. "How a candidate performs in a stressful situation and a stress-free situation holistically builds up your interview process."

Selection researchers have spent decades arriving at the same conclusion. Structured, job-related questions tighten the connection between what an interview measures and how someone performs on the job. "Fifteen or forty-five minutes is not sufficient to understand a human being," Mukherjee says. What he can get in that window is a read on how someone reasons under a condition he controls, rather than simply what they have prepared to say. Recruiters are backing away from the resume for the same reason, because a polished document has stopped separating one candidate from the next. An interview is an attempt to observe the situation and the person at the same time.

The organization builds conditions too

One place this surfaces is placement, where someone capable lands in a role asking for capabilities they do not have or do not want to use, and the mismatch shows up later as a performance problem with their name attached. "It often seems a superman is given the task of a fisherman," he says. Early in a career, candidates will accept almost any role for the compensation attached to it, and the conflict only emerges once they have enough experience to recognize the work does not suit them. By then, the person can bear the consequences twice for the same organizational decision, first placed into the wrong seat and then evaluated when the mismatch shows up in a rating.

He puts the responsibility on whoever is doing the selecting. "Understanding the human side is much more important before moving to an advanced assessment of talent," he says. "It's about understanding a person from all angles, reaching a conclusion, and placing them in the right position." None of that works without a clear definition of the role on the other side of the table, which is why defining good first does as much for fit as any interview technique.

Doing it properly takes time, and Mukherjee has had to defend that time to leaders watching a time-to-fill number. "My leaders questioned why I took so much time finalizing candidates. I told them, 'I'm saving you money. If I put the wrong talent in place, it will cost you more,'" he says. Time-to-fill shows up on a dashboard that week. The cost of a bad hire shows up months later, which is part of why recruiting is straining under automated volume. The pattern resembles what happens at appraisal time, where a measurable outcome gets optimized while the conditions around it disappear from the frame.

What the rating forgets

At appraisal time, the order flips. Instead of building a situation and watching someone work in it, the organization starts with the result and works backward to the person, skipping the conditions in between. "A person may be ready to deliver, but if the system is not in place, they are not able to take calls on time," Mukherjee says. Approvals stall while leaders sit in planning meetings, execution waits, and the employee gets marked down for missed timelines they could not have hit.

And the rating is a shaky measure to begin with. Handed the same score to work from, managers accepted it far more often when it recommended a high rating than when it recommended a low one, and the reluctance traced back to protecting the relationship rather than any doubt about the number. Mukherjee's own complaint is simpler, which is that the conditions around the work never enter the conversation at all. "Leaders often overlook the ground view, because the aerial view always looks like everything is settled," he says.

From above, an outcome is a number attached to a name. On the ground, the same outcome carries a history of approvals and handoffs that nobody recorded.

Write down where the work stopped

What Mukherjee asks for is the system itself in place before anyone is measured against it. A rating means something when a written process shows where work actually stopped, which lets a manager separate a person who could not deliver from a person whose approval sat in a queue for three weeks. Without it, the review defaults to the individual, because the individual is the only variable anyone wrote down.

The principle behind Mukherjee's approach does not change once the interview ends. In both settings, he wants the situation around the person to be visible before anyone interprets the result. The difference is that an interview deliberately constructs a situation in order to observe it, while a performance review inherits one nobody bothered to record.

"Who is going to ask what held you back?" Mukherjee says. Until someone asks, the rating can't distinguish between a performance problem and a system problem.