(FORTUNE Magazine) – How do you hire and keep the best? "Topgrade," says consultant Bradford Smart, author of Topgrading: How Leading Companies Win by Hiring, Coaching and Keeping the Best People (Prentice Hall, $26). That's not as simple as it sounds. Smart's method takes time and diligence. To decide who is an A player, for instance, he asks job candidates to fill out a form with more than 100 questions, called the chronological in-depth structured interview, or CIDS. Sounds wearying, but Smart notes that hiring the wrong person is more expensive than hiring right. If you want to turn C players into A players, for instance, you'll have to invest many hours in coaching. Smart, who has consulted at General Electric, AlliedSignal, and Gateway, recently talked to FORTUNE's Geoffrey Colvin about how to topgrade your corporation--and yourself.

How not to hire. The way most companies approach [hiring] is for the CEO to delegate too much to human resources professionals who contract with search firms. [Those firms] get a marginally clear understanding of the position, don't truly perform a thorough job analysis, have a partial understanding of what the job really should be and a partial understanding of what the competencies should be.

After a couple of months, candidates are generated by the search firm and they're put through a series of shallow, redundant, superficial, one-hour interviews. But you add up a lot of insurance physicals and it does not come close to approximating a three-day executive physical at the Mayo Clinic. The CIDS interview is really the Mayo Clinic thorough examination of the candidate.

The Mayo Clinic of interviews. Instead of the typical approach of investigating the most recent position--"Tell me about your present job and what you're looking for in your next job"--the CIDS interview has a chronological approach that begins very early in a person's career, perhaps even during the educational days, and very thoroughly covers what the person did and how the person did it. Questions are asked about specific accomplishments and successes, but also about failures and mistakes. And it includes a powerful TORC [threat of reference check] technique that essentially asks, "If I were to ask you to arrange for an interview with that boss, and if the boss were very candid with me now, what's your best guess as to what he or she would say were your strengths, weaker points, and overall performance?" If a person has had ten jobs, the patterns that emerge provide penetrating insight into what makes this person tick and how this person is apt to function in the next job.

Care and handling of C players. If people see some of their co-workers unfairly treated, even A players might be inclined to leave. So C players should be handled in steps taken very, very properly. If I had a C player reporting to me, does the C player truly understand what the short-term and long-term accountabilities are, and what he or she is being measured against? There should be a very thorough action plan--sometimes it's called a performance-improvement plan--for someone who's regarded as an underperformer. It's oral and it's written, and it lays out specific activities that the C player agrees to with his or her boss in order to move up to the level of performance of an A player.

So imagine meeting with an underperformer on a weekly basis for three months, four months, five months. What typically occurs is that the individual will say, "Boss, we've been at this for three months now, and I'm still not performing at the level I agreed was the level of performance necessary for me to hold this position. I don't like getting up every day, looking at myself in the mirror, and saying, 'I'm going to probably fail again today.' And I don't like dragging the team down and disappointing you, so perhaps I should look for a position elsewhere."

Salvaging C players requires coaching skills that, frankly, not many managers have. It's smart for anyone in management to try to learn how to coach, but if it doesn't occur, it is better to bite the bullet and nudge the C player out.

I beg to differ, Mr. Drucker. In the March-April issue of The Harvard Business Review, Peter Drucker, for whom I have the utmost respect, says, "Don't try to fix your weaknesses--it doesn't work." I disagree vociferously with Mr. Drucker on that point, because I know that fixing weaknesses is, in most managers' cases, time best spent.

For example, [at General Electric] if a person is not exhibiting the GE values--meaning teamwork, treating people with respect, and so forth--and that is measured and documented several times, then the person could lose his job or her job, even though the performance is outstanding. Sharp people--resourceful people--will figure out how to do damage control on their weaknesses, and it really, really does help their overall performance. Tom Brock [who was a GE division general manager] was considered deficient on the interpersonal aspects of leadership. And during one year he progressed from the interpersonal weaknesses representing a career breaker to a career enabler. My experience is that you don't have to be as good as Tom Brock at fixing weaknesses. Fixing one's weaknesses does not mean going from a two on a ten-point scale to a nine. Sometimes it means going from a three to a 4 1/2. Where a three on interpersonal flexibility is just not good enough, a 4 1/2 means a person who is, at times, kind of a pain in the neck to work with, but whom A players will still go to work for. The teamwork will be good enough, and so the person can succeed.