Two interviews are enough to hire
My most provocative opinion about hiring in IT is that there should only be one technical interview and one general interview, and that each should last no more than an hour. I believe that making the hiring process longer than that is not only useless, but counterproductive.
Of course, such a laconic hiring scheme is much easier to implement and takes less time, but there are other, less obvious advantages.
More effective interviews
“When everyone is responsible, no one is responsible.”
Interviewers are much more careful in choosing the right questions when they know that no one else will be evaluating the candidate on the same parameters. Their decisions carry more weight because if they can't gather enough information to make a judgment about the candidate, no one will back them up or fill in the gaps.
Introducing additional interviews, whether technical or general, only reduces the likelihood that the necessary information will be obtained, since no one is specifically responsible for collecting the information on which the decision will be made.
Attracting good senior specialists
When looking for people to fill senior positions, the best candidates are unlikely to have the patience to go through a long, drawn-out selection process. A cumbersome hiring process turns off the very people the company would most like to recruit (i.e., those who can afford to be picky about who they hire).
Many people think that by delaying the hiring process they will ensure that they retain the highest caliber candidates, but in reality they are inadvertently choosing the most desperate ones, with a high tolerance for stupid demands and red tape. Do you really need such programmers to grow your organization?
Bias and Preferences
In my experience, people form certain attitudes toward candidates early in the interview process (and sometimes even before the interview process begins). By adopting a shortened hiring process, we officially acknowledge the existence of this unofficial phenomenon.
This is especially true in large companies: the hiring manager has a few candidates in mind in advance (say, someone on the team recommended them, maybe even the manager himself, or they have a good portfolio) and has a strong bias towards choosing those who caught his attention before the interview process even began. Delaying the hiring process is a poorly disguised attempt to cover up this bias with an “objective” selection scheme, even though the manager is likely to ignore or challenge the result if it contradicts his personal preferences.
This is not to say that I think this approach to selecting candidates is reasonable or acceptable. I just doubt that adding more steps will do anything to eliminate bias. If anything, these extra steps will only further distort the picture, since the selection will be narrowed to candidates who can afford to deviate from their usual routine for a long time and spend a lot of time in a series of interviews with the selection committee (and these are usually candidates from privileged backgrounds).
How we got here
What prompted me to take this position was my experience as a hiring manager at a previous company. At first, we had a long series of interviews for permanent employees, and a much shorter version (one technical interview and one general interview) for interns. The original rationale for this arrangement was that when hiring interns, the “stakes are not as high,” meaning the hiring process can be structured with less “scrupulousness.”
But in reality, we found that the internship scheme was producing exceptional candidates, even though we were supposed to have lowered the standards for it. So we thought: why not try this for all positions, not just internships?
This transition did not happen overnight. We changed the hiring process gradually, eliminating one step from the interview series with each new vacancy. Eventually, we arrived at one technical interview and one general interview (that is, what we were already practicing with interns). With each new step in simplifying the process, we realized that the additional step we eliminated was not necessary at all.