In March 2012 Donna Morris, Adobe's senior vice president of global people, had just landed in India after a long flight. She agreed to an interview with the Economic Times. The reporter asked what she would do to disrupt HR. Still jet-lagged, Morris answered that Adobe planned to abolish the annual performance review format3.
The idea was real enough, but she had not yet put it to the leadership team. The article ran, the news reached headquarters in California, and Adobe had to turn an interview answer into an actual plan3.
Three years later a much larger firm announced something similar in Harvard Business Review, this time with full numbers.
Deloitte counted the hours, then the bias
Deloitte had more than 65,000 people at the time. When it added up the time spent filling in forms, holding review meetings and agreeing ratings, it came to close to 2 million hours a year1. Most of that time went on discussing last year's results rather than helping people do better next year.
The second finding was more troubling. Deloitte cited a study of 4,492 managers, each rated by several others: 62 percent of the variation in ratings came from the peculiarities of the raters, while the actual performance of the person rated explained only about 21 percent1. Researchers call it the idiosyncratic rater effect. Put simply, a rating tells you more about the rater than about the person being rated.
Deloitte's answer was not to stop assessing people. It changed the questions. Instead of asking team leaders to rate other people's skills, which humans do very inconsistently, it asked about the leader's own intentions, which humans report quite consistently. At the end of each project, or each quarter for long projects, the team leader answers four questions, paraphrased here1:
- If it were my own money, would I give this person the largest raise and bonus available? (five-point scale)
- Would I always want this person on my team? (five-point scale)
- Is this person at risk of low performance? (yes or no)
- Is this person ready for promotion today? (yes or no)
Alongside that come short weekly check-ins between the leader and each team member, focused on upcoming work rather than past work1. Before rolling it out, Deloitte studied 60 high-performing teams totalling 1,287 people against a representative sample of 1,954 to find which conditions really set strong teams apart1.
Adobe measured it in attrition
Back to Adobe. When it did the sums, it found some 2,000 managers spending 80,000 hours a year on annual reviews, the effort of 40 people doing nothing else full time2. Employees described the old process in unflattering terms.
Adobe replaced it with Check-in: regular conversations between managers and employees about expectations, feedback and development, with no mandatory form, no forced ranking, and managers given discretion over pay allocation within budget2.
The results were clear. Voluntary attrition fell by 30 percent. Involuntary departures rose by 50 percent, because difficult conversations now happened throughout the year instead of piling up at year end, and most of those who left were people the company did not regret losing2. That detail is often dropped when Adobe is told as a "more humane" story: Check-in did not make performance management softer, it made it more honest.
GE: from the curve to a phone app
GE is the most symbolic case, because it gave birth to the "vitality curve" under Jack Welch: each year employees were sorted into the top 20 percent, the middle 70 and the bottom 10, and the bottom group was typically managed out4. Forced ranking was relaxed from the mid-2000s. By 2015 GE had moved to an app called PD@GE for frequent feedback between employees and managers, phasing out the old annual review for a workforce of around 300,0004,5.
Not everyone should drop ratings
The three stories are easily boiled down to a slogan: scrap the annual review. That is a misreading. All three companies still had to decide who got what raise and who was promoted. What they dropped was a once-a-year ritual that ate time and relied on unreliable scores. What they kept, and invested in, was frequent conversation and data that managers could supply honestly.
For Vietnamese companies, where the year-end review is usually tied to the 13th-month salary and the post-Tet pay round, dropping ratings altogether may create more confusion than it solves. But two things can be done straight away without replacing the system: shorten the form, and change the question from "how good is this person?" to "what would I do with this person?"
