Rule of four: Google implemented a “Rule of Four” to avoid needless interviews

 It’s no surprise that Google attracts a lot of talented applicants, and narrowing them down to the best of the best can take time. A lot of time, in fact. Former interviewees sharing their experiences on Quora report that the process can take anywhere from two to four months.

“We thought that the more Google employees who interviewed a single candidate, the better our hiring decision would be—which meant someone applying to Google could be subjected to over a dozen interviews,” writes Shannon Shaper, Google’s Hiring Innovation Manager. “This took a massive amount of employee time and also could make for a grueling candidate experience.”

Realizing that this model was not sustainable, the company looked at its interview data from over a five-year period. At every interview, the interviewer would assign the candidate an interview score. The theory was that the accuracy of the mean interview score would increase the more interviewers were involved.

To Google’s surprise, the data revealed this was not the case. Four interviews with four interviewers provided an accurate prediction of a new hire’s performance 86% of the time. Adding more interviewers produced rapidly diminishing returns, increasing the accuracy of the predictive score by less than 1% per additional interviewer.

This was consistent with the company’s data about panel interviews. The decisions made by panels with four people were almost identical to those made by busier panels in 95% of cases.

Using this data, Google implemented a “Rule of Four” standard for most of its interviews. Using a maximum of four interviews, the company cut its time to hire by about two weeks—making the process less stressful for candidates and saving, according to Shannon, “hundreds of thousands of hours” of employee time.

Actionable items

Here is the article. I felt much better after the onsite and decision not to move forward. I could not beat the statistics. I need to work on another 200 hard level algorithms, and learn a lot from top voted discussion posts, and learn one more programming language like Python.