Topgrading

"In a nutshell, topgrading is the practice of packing the team with A players and clearing out the C players."

— Bradford D. Smart

Smart holds a PhD in industrial psychology from Purdue and has run since the seventies a consultancy in executive assessment, with more than 6,500 in-depth interviews over his career. In the eighties and nineties he helped General Electric design its hiring practices, and he coined the term topgrading with his son Geoff, later coauthor of Who. This book is the complete manual of the method.

Hiring is usually a company's most important and worst-executed decision. Résumés exaggerate, the traditional interview rewards whoever interviews well and not whoever works well, and the references the candidate picks say nothing. According to Smart, the typical company gets about one in four hires right, and each mistake costs between five and 27 times the job's annual salary, depending on the level, counting lost time, wrong decisions, affected customers and the cost of replacing the person.

The goal is to fill the company with A players, whom Smart defines as the top 10% of the talent available for that job at that salary level. B's are adequate, and C's are chronically below the bar. A key idea is that the definition is relative to the salary: an A isn't the best in the world, but the best you can attract with what you offer.

The full method has twelve steps. The essential ones are:

• Job scorecard. Before searching, write what measurable results the person must achieve in their first year and what competencies the job requires. Without that, you're evaluating likability.
• Recruit from your A's network. The best candidates come through recommendations from your best people, not through ads.
• TORC, the threat of reference check. From the first contact the candidate is told that, if they reach the final round, they themselves will have to arrange the calls with their former bosses. Weak candidates withdraw on their own and strong ones become more candid.
• Chronological interview. Two interviewers walk through the whole career, from school, job by job, with the same questions: what were you hired to do, what did you achieve, what mistakes did you make, what was your boss like, how would they rate you and why did you leave.
• References with the bosses. Real calls with the direct bosses the candidate scheduled, not with HR, which usually confirms only dates.
• Report and decision. A written summary of patterns, strengths and risks before deciding, so you don't hire on the last impression.

The question that makes it all work is "if I talk to your boss, how would they rate your performance from 1 to 10?" Combined with TORC, it works like truth serum: the candidate knows the call is going to happen, so they prefer to explain themselves the job that went badly. The reason for leaving each job is also revealing; Smart asks you to distinguish whether the candidate left because they were pulled to something better or because they were pushed out.

The method doesn't end at hiring. Smart asks you to evaluate with the same logic the team already inside: develop the B's who can become A's, and reassign or let go the C's, because C's drive away A's. The common mistake is applying it only outward and tolerating inside what you would never hire today.

How to apply it? Before your next key hire, write the scorecard on one page with three to five measurable results. From the first call, tell the candidate that at the end they'll introduce you to their former bosses. And in the interview, walk through at least their last three jobs with the question of how each boss would rate them.

As a founder it will help you stop depending on intuition for the most expensive decision you make, and to review your current team honestly. You can also give yourself the chronological interview: it's how serious funds are going to evaluate you.