*This post about how to apply Radical Candor in a crisis was originally written at…
This is a tricky question to answer because there’s a world of difference between development (helping people grow in their careers) and performance management (rewarding good results and penalizing bad).
Yet, the two are often conflated.
Giving a person a performance rating that translates to a bonus is very different from having a series of weekly impromptu guidance conversations that help a person develop the skills they need to succeed, help them know what to do more of and what to do less of.
And yet people often talk about feedback as if the impromptu conversation and the performance review conversations are one and the same.
They are two very different things. Impromptu conversations should be happening multiple times a week and should be focused on coaching a person to do more of what’s good and less of what’s not.
Performance reviews should happen once or twice a year and should help the person understand their compensation, and their promotion prospects or termination risks if any.
Most companies expend so much energy on performance management that they don’t focus as much on development.
Radical Candor is mostly centered on development because that was what I did as an operational leader. It was never my job to design HR systems like performance management processes.
But I’ve been on the receiving end of several different ones, and also have seen what happens when they are absent, so I do have some thoughts that may prove helpful.
One of the ways that these systems go most badly wrong is when the operating executives all have a gut instinct about what it should look like but are unwilling to take the time to help develop the system.
Collaboration Is Essential for Success
When operating executives over-delegate to HR, and then refuse to participate in the system that gets designed, it’s a terrible experience for everyone.
A good collaboration between operational leaders and HR is essential for developing a good system.
What I offer below are the thoughts of an operational leader. You’re going to need the help of your HR team, which has more experience than I do, to roll out a great system.
Of all the places I’ve worked, by far the best performance management systems were the ones that Shona Brown, who led business and people operations at Google, developed.
As she’ll be the first to admit, they weren’t perfect. And I’ll say what she is too humble to say: they were some of the best, most thoughtful, most fair systems ever developed in any company.
Shona, more than any other individual, figured out how to operationalize the kind of innovative culture that gave all employees agency to do the best work they were capable of and discouraged Dilbert-y management styles from getting entrenched at Google.
She provided some crucial guidance for this topic, though there is obviously much more to be said on the topic of performance management systems.
The way you think about developing the skills of the people in your organization and how you think about performance management must be aligned.
It is a manager’s job to both help each person on their team develop and grow in their career, and also to manage the performance of each person.
Balancing the intrinsic desire to improve and grow and the extrinsic desire for rewards like bonuses, equity, and promotion is one of the most difficult things about being a manager.
It also makes designing performance management systems and teaching development very difficult.
If you’re not careful, you’ll design a performance management system that makes people reluctant to invest in development or a development system that is so focused on helping people improve you never do proper performance management or let people know when they are simply in the wrong job.
The atomic building block of Radical Candor is the two-minute impromptu development conversation. You don’t want to try to operationalize impromptu chats. The motivations of both the feedback giver and the feedback receiver need to be intrinsic.
The motivation to solicit guidance and to act on it is the desire to improve, to grow, to do good work and then make it better, to build strong relationships and then make them stronger. “I’m listening to you because I want to develop the skills and the team I’ll need to succeed.”
The motivation to give guidance is mostly altruistic—to help another person and the team as a collective flourish. “I’m telling you this because I want to help you develop the skills you need to succeed and because it’s not fair to your peers if I don’t tell you.”
If you tie each two-minute conversation too explicitly to extrinsic motivators like bonuses, promotions, or termination, you can ruin the motivation for them.
If an employee thinks each two-minute impromptu conversation is going to impact their compensation or future career prospects, they will be less open to hearing what was said and more prone to fight it.
And, while managers must be held accountable for having these conversations, if HR were to be able to eavesdrop on all these private moments to judge if they were happening properly they would get corrupted.
Imagine being an employee and wondering if your manager was telling you about a problem to help you fix it, or to get a better rating for “developing people.” Development conversations need to be private to help the relationship deepen.
And yet at some point, performance management must kick in. If people fail to develop the skills they need to succeed, there must be consequences.
Bonuses, promotions and terminations must happen if work is to be fairly compensated and if the right people are to be promoted into the people who aren’t able to carry their weight moved to a job that’s a better fit for them or in the worst cases fired.
Leaders must be held accountable for investing in the development of the people who work for them, not just for doing proper performance management.
If development is never taught nor compensated, it won’t happen. Development takes real work and emotional discipline; it must be rewarded before a person gets to heaven.
All too often people use a performance review process as an excuse not to have the two-minute impromptu conversations that are an essential part of development.
And a performance review process without the development conversations is like capping a rotten tooth. It will just rot faster and more painfully.
That is why so many companies today are dispensing with performance review systems altogether.
Sometimes they do that as a way to build the development capability in their organization, and they have a plan to reintroduce a new and improved performance management system in the future.
That makes sense. However, all too often, they are simply dispensing with both development and performance management. In other words, they are dispensing with management. That never ends well and doesn’t make sense.
I predict that most companies eliminating their performance review systems will wind up re-instituting them in a few years.
If you don’t have a formal performance review system, many people won’t ever understand the implications of their behavior and work on their salaries, bonuses, and promotions.
The decisions about who gets salary increases, bonuses, and promotions become even more opaque and probably less fair.
The key is to design a system that does not interrupt the impromptu development conversations but rather reinforces their importance.
The information below can help you understand the impact of the performance review system that currently exists at your company.
If you’re responsible for designing your company’s review process or for your company’s culture, I hope you’ll find it especially relevant to your work—though I write humbly in the knowledge you’ve thought more deeply about this than I have.
My main value add below is to keep coming back to how a performance management system can reinforce or interrupt a culture of Radical Candor.
Elements of a formal performance review process
By no means do I think there is one “right” way to do performance management or performance reviews. Below are the aspects to consider, and an explanation for my point of view about how each will impact Radical Candor.
The key thing is to make conscious decisions that are right for your particular situation.
- Rating or no rating
- Categories of ratings
- Job Ladders
- Number of ratings
- Language matters
- Consequence of ratings
- Distribution of ratings
- Forced curve or no
- Calibration of ratings
- 360 or exclusively the manager’s POV
- Transparent or confidential
- Lightweight or heavyweight
1. Rating or no rating
There is a growing school of thought that adults should not be given a numerical rating; that feedback should be qualitative.
Moreover, there’s an argument that people tend to focus so intensely on their rating that the rating gets in the way of good development conversations.
So, the rating could actually inhibit progress. If you only have one development conversation a year and it’s tied to a rating, that is certainly true.
But if you separate development conversations, which should be happening weekly in two-minute impromptu conversations with a focus on coaching a person, and the performance management conversation, which should happen once or twice a year with a focus on helping a person understand their compensation and whether they are likely to be fired or promoted in the future, that is less true.
The reason is that they are simple, powerful forcing functions. There’s nothing better than having to rate people to force managers to think through how people are really doing and to ensure that employees know what their managers think.
There’s nothing simpler than a rating to interrupt “Ruinous Empathy.” A manager may think an employee is failing and may think they have told that employee clearly, but all too often the employee doesn’t really know until they get the rating.
Ratings are blunt instruments, and they do sometimes do unnecessary harm, but in the end, I believe they do more good than harm.
On balance, I’d recommend giving ratings.
2. Categories of ratings
Some companies just consider “overall performance” when having development conversations/giving ratings. Other companies like to break out the key performance drivers and discuss/give different ratings for each driver separately.
Breaking these drivers out correctly can take tremendous time, and sometimes feel more like a rathole than anything else. I watched in horror as debate swirled at one company around this question for months.
However, they landed in a really good place and I’d say the time was well spent. HR and operating leaders at the company collaborated to identify what mattered most to driving performance in a way that reinforced the culture in the long run.
Ten years later, they are still using the same categories to guide performance ratings.
Another resistance to choosing categories of performance drivers is that breaking them out makes the forms for 360s and for performance reviews more complex.
And I have seen some companies list so many competencies and ask people to rate on each of them that the whole process became a ridiculous joke.
The categories need to be abstract enough that they apply to everyone, and there should only be three or four of them.
Still, on the surface, one rating seems easier and faster than three or four.
Despite the potential for ratholing and the appearance of complexity, I like breaking out the key performance drivers.
The main reason is that invariably one of the drivers is “teamwork” and rating people explicitly on “teamwork” gives much-needed and very clear feedback to the people who get great results but leave a trail of dead bodies in their wake.
Explicitly calling out other drivers of long-term performance that may not get captured in annual results, things like innovation or efficiency or customer happiness, can help get people focused on the big picture, and not be so swayed by the stimulus-response of the day-to-day work or hitting quarterly targets.
You should choose the words that best reflect what your organization is trying to achieve. Four of the most common drivers of performance to consider.
Some may be more important to your mission, others less. And there may be a performance driver in your organization that is not included in the list below.
- Results: Is the person achieving the goals set, doing what they are supposed to do?
- Teamwork: How well does the person work with their team to get the results? Do they help others succeed, or leave a trail of dead bodies? Do they go out of their way to help people beyond their team succeed—how broadly do they define their team?
- Innovation: Does this person come up with new ideas that help the team do old work in a dramatically new, better way, or to start doing a whole new category of work that furthers its mission? Innovation can include step function ideas that change everything quickly and incremental change that improves operations dramatically over time. Both are important.
- Efficiency: Does this person work efficiently, and help others work efficiently as well? Does this person figure out how to eliminate unnecessary work and save the time of their colleagues?
Teamwork and results are pretty obvious choices. I separate out innovation and efficiency because it’s important to think about new things to do as well as how to do existing work more efficiently.
Often it requires great innovation to figure out how to be more efficient, and the time the efficiency gains you is where you find the mental space for step-function innovation.
Eventually, you want to identify the drivers that are more relevant and specific to your business. These may not be the right categories for your performance evaluation.
But yours should be at approximately the same level of abstraction. Don’t get too specific unless everyone in your organization is doing one specific thing.
If you decide to do ratings, I would not recommend an average to get to an overall rating. The only people who could get an overall rating of 4 or 5 have to get a 4 or 5 in all categories.
People who get any 1s or 2s get an overall rating of 1 or 2. This will help you get to a reasonable distribution of the overall rating.
3. Job Ladders
I hate job ladders. But they are a necessary evil. Obviously, a recent college graduate you just hired a few months ago should not be held to the same standards that your CEO is held to.
So you need to describe what teamwork means for an entry-level employee versus a manager, a director, a VP, and so on.
The key thing is to write at the right level of abstraction. If you get too specific, then people will try to use them like the instructions for an intricate lego set.
A job ladder shouldn’t be written like a false promise you can’t keep: if you do these things you’ll get promoted.
Also, if they are too specific you’ll have job ladder proliferation. Instead of having one for the major functions at a company, every sub-group will have its own ladder and then will get in a food fight with other sub-groups about why theirs isn’t fair.
However, if they are too abstract they feel irrelevant to people’s jobs and read like a list of bloviating BS.
It’s important that operating executives who have done the job at every level get involved in writing the job ladder descriptions. HR should be the editor not the author of them.
4. Number of ratings
Many studies have been done to figure out what the optimal number of ratings is. There’s debate in the literature, but a lot of evidence points to 5-7 ratings. However, that feels like too many to me.
The problems with too many ratings are manifold. One is false precision. Too many ratings make managers feel they should be able to achieve a level of granularity in their assessment of people’s performance that isn’t actually possible.
Another problem is that when managers have too many ratings at their disposal they tend to try to use them to manage emotions. “I want to show the person they are making progress over time so I’m going to nudge performance up as they stay in the role longer.”
This is not performance management, it’s calendar management. Just because one person has been in a role for longer than another doesn’t mean they are doing better work.
A third problem is that it takes longer but doesn’t add any value, and probably subtracts value.
The problem with 3 or 5 ratings is that they give almost no information to the vast majority of people. If you assume 5% get the lowest rating and 15% get the highest, which usually happens, that means 80% of people just know they are in the vast middle.
What they really want to know is whether they are in the bottom half or the top half. Not letting them know can result in a lot of unhappy surprises from people whose performance is toward the bottom, and a lot of dissatisfaction from employees whose performance is above average. Five ratings do much the same because the middle rating is generally over 60%.
Although the simplicity of 3 ratings is beguiling, I recommend four ratings. Many managers will resist four ratings because it forces them to separate cleanly above and below-average performance.
It does require some thought to divide the middle 80% of people between those who are above average and those who are below. But if you’re in that middle 80%, you really deserve to know what your manager thinks and what your team thinks.
Four ratings, like three ratings, push managers to warn the people whose performance is so bad they will be fired if they don’t improve dramatically; and to reward those whose performance is so amazing that the person must be retained.
Truly terrible or truly exceptional work deserved to be called out in a class of its own. Four ratings give clear feedback to employees about where their performance stands, without pushing for false precision of more ratings.
Separate ratings for each of the four categories you’ve chosen (see above) can help the performance review to serve as a diagnostic for improvement, not just a whip or a carrot (ie, “your results are amazing, but your teamwork will hold you back”) in a way that a single overall rating does not. In other words, it can give you some of the benefits of more ratings without the pitfall of false precision.
A downside of four ratings is that it doesn’t give you a rating for the people whose performance is in the top 1%. These are the people you want to retain at any cost.
However, there are probably better ways to identify these people than a performance rating. Since there won’t be many of them, they are probably best identified in a separate process.
On balance, I’d recommend four ratings, though there’s also an argument for 3, 5 or 6. Whatever you do, don’t go over six. A system with four ratings might look something like this (but better, with the benefit of some love from a designer, and comments explained later in this post).
5. Language matters
For the purposes of averaging, you will translate words to numbers. You could simply use numbers: 1, 2, 3, 4. You could use symbols, ✔-, ✔, ✔+, ✔++. However, I recommend language.
Words matter. “Not OK” to “OK for now” to “Good” to “Great” are not the words you should choose. Instead, choose language that reflects your culture, that gives the whole thing a little personality. But don’t make the words too charged.
This stuff can get pretty boring in the abstract but the emotions it stirs up in practice can’t be exaggerated. Make sure the words you choose demonstrate a desire to assess each person’s work in order to help them grow professionally, not to label or judge them as human beings.
6. Consequences of ratings
Ratings guide variable compensation, promotions, and terminations. Indeed, part of the reason to do ratings is to make the reasoning behind these decisions more transparent to employees and to force managers to calibrate.
Something is going to drive these decisions. Often it’s a half-formed and unexamined thought in the manager’s head. Doing ratings and explaining their impact on these decisions will force managers to think through and communicate the decisions more clearly and explicitly.
When ratings are given and it’s clear to everyone how they impact compensation, promotions, and firing, these decisions are more likely to be made on explicit, fair criteria rather than on subjective, fly-by-the-seat-of-the-pants instinct.
The most important consequence of ratings should be that managers work with employees to do more of what they are good at and less of what they are bad at. The stress should be on adapting each person’s role to their strengths, not punishing them for weaknesses.
However, a “Not OK” means there’s a risk the person will be terminated if they can’t achieve at least a basic level of proficiency in that area. Each person should be at least “OK for now” in each of the categories.
If a person is great in every category except teamwork where they are “OK for now” the person should be given more individual contributor work, rather than being put in situations where teamwork is essential. The person can be challenged to bring their teamwork up to a good or great level.
But the focus should be on giving them work where they can play to their strengths. If people “spike” in a certain kind of work, give them more of that work, rather than more work where they are just OK.
It sounds so obvious but many managers have a negativity bias and focus all their efforts on plugging weaknesses rather than helping people do more work they are great at and love.
There will be a large population who consistently do “OK for now” work, who never get any goods or greats. Over time, these people should be pushed to find a role where they can excel. “
OK for now” that never moves to good or great is a life of quiet desperation, and that is not OK. And nobody whose work is rated “OK for now” should be promoted. A condition for promotion should be a majority of great ratings consistently over time.
When a manager gives a “not OK” rating, HR should assign somebody to help the manager start documenting performance issues and ensuring that the person understands what the consequences will be if their performance doesn’t improve.
Too often managers don’t give a low rating because they don’t know how to or don’t want to take the time to deal with documentation.
If an HR person were assigned to help a manager with an employee who’s not performing, many managers would be less apt to resist the lowest ratings.
7. Distribution of ratings
There’s also a lot of debate about what the right distribution of the ratings should be. You are going to have to decide what numbers are right for your organization. Not OK means you’re at risk of being fired.
My experience is that at any given time there are at least 5% of people who probably need a strong wake-up call like that.
But you may feel that’s excessively harsh for your environment, or not harsh enough. Great means you want to give people outsized rewards because you want to retain them. In general, I’ve found that’s about 15% of a team. A
s stated above, there are usually 1-2% of a population whose work is truly exceptional, whom you want to retain at all costs.
Identifying these people is an important exercise, but there is no need to create a special rating for 1-2% of your population. That leaves about 40% of your population getting an “OK for now” and 40% getting a “good” rating.
I don’t know enough about your organization to make the distributions I’ve described above a recommendation. You’ll need to decide what the consequences of each rating are and therefore what the distribution ought to be.
However, I do recommend making the distribution you expect to see transparent to the organization. If you don’t, calibrations (see below) will become opaque and excessively frustrating.
8. Forced curve or not
There’s also a lot of debate about whether to force the curve or not. One of the benefits of having the categories (teamwork, innovation, efficiency, and results), and calculating the overall rating as I outlined above is that it takes some of the pressure off the distribution for managers.
If they feel it’s appropriate, they can tell every single person on their team they are great at something, and yet still be in line with the distribution guidance. The question is, does it make sense to force managers to hit that curve?
Volumes have been written about the evils of the forced curve. The unintended consequences and political gaming of a forced curve are terrible. I never experienced a forced curve, but I have heard crazy stories from organizations that do.
There was one VP I knew who found it too difficult to identify actual low performers, so he assigned the lowest rating Russian-roulette style.
He was transparent with his team about it, and thought he was being very clever—he never had to have hard conversations with employees on his team about why they were getting low ratings or his boss about why he didn’t hit the curve.
These kinds of stories convince me a forced curve is a bad idea.
Here is my recommendation. Don’t force a curve but do put pressure on it:
- State explicitly what the expected curve is.
- Count “nonregretted attrition” (people who were fired or managed out) in the stats for the lowest rating. Track non-regretted attrition over time. Otherwise, people keep their low performers around to hit a curve rather than firing them…
- Each manager’s curve should be visible to peer managers; each director’s ratings curve should be visible to peer directors; each VP’s ratings curve should be visible to peers, VP’s, and so on.
- Require managers who fall outside the expected curve to provide a written explanation for why; ask a lot of questions; allow peers to ask a lot of questions. When one team’s curve is notably skewed when compared to other teams, ask the question: “Did this team really contribute more/less than the other teams in this rating cycle?” If the answer is no, push the manager to reassess ratings. Similarly, if one team delivered something exceptional, make sure that the team’s manager is reflecting this in ratings. Most managers will stick to the middle, skewing to the positive; they give out too many “goods” and “OK’s” and too few “greats” and “not OK’s.” That’s why calibration sessions are so important.
9. Calibration of ratings
A calibration is a meeting in which a group of managers get together and make sure they are all rating people in the same way—that one is not a “hard” grader while another is an “easy” grader.
If you are a manager of managers, it’s important to make sure all the people who work for you get together and calibrate. If you are a manager and your boss is not doing this, check in with your peers.
Calibrations are invariably painful. A common scenario is this. Geoff manages six people. He thinks they are all fantastic, especially Tony, whose work he thinks is “great.” Geoff’s peer, Wilma, who also manages six people, does not think Tony is so great.
In fact, Wilma would probably rate Tony’s work as “OK.” If Wilma speaks up, Geoff will be furious. If she doesn’t speak up, she will be furious. Their boss, Ann, now has to navigate a hard conversation.
Ann probably feels tempted to skip the conversation altogether. But, if she doesn’t address this, several bad things will happen. First, she won’t hit the right distribution.
Second, Geoff will get a reputation as an easy grader. People will still love working for Wilma but will feel it’s unfair that Geoff’s team gets paid more and promoted faster because Geoff is an easy grader. This will create weird incentives.
Furthermore, because Geoff is actually overly positive about Tony’s work, if Ann allows this to stand, it will hurt Tony, Geoff, and the team’s reputation more broadly.
Ann could just fix it by talking to Geoff, but she knows that this is just one instance where she happens to know the details; there are plenty of others where she doesn’t know the details.
If she pushes everyone on her team to challenge each other, they are far more likely to get on the same page more quickly and more deeply than if she tries to tweak ratings herself.
If Ann had several “layers” under her, the calibration meeting is even more important because it’s important to make a conscious decision about what the distribution ought to be like at each level.
In too many organizations a disproportionate number of the “senior” people get the highest ratings, and the majority of the low ratings go to the most “junior” employees.
When I was at Google, I asked our HR person to slice the ratings by level. Although Google worked hard to be “flat” organization, there were eight different levels on the team I managed—from level 2 to level 9.
Level 2 was a customer service rep, level 9 was a Director II. Overall, our distribution looked pretty good. But when we looked at the distribution of ratings for level 2 employees versus the level 7 to 9 employees, I felt bad.
The bulk of the lowest ratings were given to the level 2s; the only top ratings went to level 8s. If I looked at how I rated the six people who worked directly for me, it was the most out of whack.
This felt like self-dealing and I knew it wasn’t fair. Before we went back through and re-calibrated by level, we had a theoretical conversation. Did we really feel that there should be such a difference by level?
There were some arguments in favor of a different distribution for entry-level employees. The first couple of years could be a kind of weeding-out process. Investment banks and consulting firms generally take people right out of college into a two-year program.
They invite about 10% to stay, but everyone else must leave after two years. That effectively means 90% of the people leave—they aren’t good enough to keep. They don’t call it missing expectations or poor performance—they just expect 90% to go.
The effect was like giving 90% of level 2 employees a bad rating. Should we do something like that? We decided we didn’t want to, that we wanted Google to be a place where people get a job after college and stay.
The skills they learned in the first two years would make them more valuable to the team and the company. Therefore, we agreed the distribution should be the same at every level.
My instinct is that it is fairer to impose the same distribution on all levels of employees. But this is an instinct, not a recommendation. In some situations, there may be a very good reason to vary the distribution by level.
However, I do have one strong recommendation: make a conscious decision. Having made it, all ratings must be calibrated by level as well as by team. In other words, if there is a group with a director, nine managers, and 90 individual contributors, the director almost certainly shouldn’t get the “fucking amazing” rating.
That director needs to be compared to other directors, not to the 9 managers and 90 individual contributors.
Out of every group of 100 people, there’s one who drives things forward to a far greater extent than anyone else. Identifying those people is one of the most important things bosses do.
Some companies do performance reviews every quarter, some twice a year, and a lot never do them. At Juice, I never did performance reviews and couldn’t understand why they were necessary. That was a disaster.
At Google, we did them twice a year, one lightweight and oral, one written and more intensive. That worked pretty well until the intensive review got too intense.
Never doing performance reviews is a bad idea. However, doing them every quarter puts a heavy tax on managers, and doesn’t give employees much time to show improvement from the last rating cycle.
So, my recommendation: do it twice a year; one can be lightweight, oral, and just between the manager and employee; the other should be written and include a lightweight 360 component.
11. 360 or exclusively the manager’s POV
I strongly recommend a 360 process. Requiring 360 reviews is one of the most important things any organization can do to make sure that a manager’s subjective point of view does not create favoritism or allow unfair/suboptimal allocation of resources.
When they are transparent, 360s can also reinforce a culture of direct, open feedback and debate.
However, the devil is in the details here. 360s are generally more time-consuming than they need to be. During the review season at Google, I would spend about 30-60 minutes for every peer review I had to write, and I usually got asked also write about 20 of them.
Everyone around me was in the same boat. There were benefits to having to write such extensive feedback—it did really make me think. However, I think 80% of the value of the reviews could have been met by simply letting people know how I thought they were doing along the four criteria.
I think that 360 peer reviews in particular can be much more lightweight for the reviewers than they are at most companies. I’d recommend simply asking employees to rate their peers on each of the four criteria; no need to write anything for “OK for now” or “good” ratings; ask them to write no more than 75 words for “not OK” and “great” ratings.
The expected distribution should be shown, and the employee’s actual distribution calculated for them so that employees can self-regulate when they see they are an “easy” or a “hard” rater.
12. Transparent or confidential
If you decide that you are going to do 360 feedback, you need to determine if it will be transparent (what people write about their colleagues is visible to those colleagues) or confidential (what people write about their colleagues is visible only to the colleagues’ managers.)
I strongly recommend transparency. Confidential feedback does not encourage people to Challenge Directly.
Confidential feedback does have a couple of benefits. One, it in theory allowed employees to write in shorthand—if you know that somebody is going to see what you’ll write about them, you’re likely to spend more time choosing your words carefully.
Two, sometimes confidentiality meant that people gave feedback to a manager that they wouldn’t otherwise have given.
But, there were several considerable downsides to confidential feedback. One, it didn’t force peers to have direct conversations with each other. Therefore, it was a missed opportunity to drive a culture of direct, open feedback. In the worst cases, it encouraged backstabbing behavior.
Two, managers had to read and synthesize all the peer feedback for all employees, since it wasn’t transparent to them. Unfortunately, this synthesis was sometimes a value-subtracting process.
Making 360 feedback transparent drives a culture of direct feedback, is more efficient for managers, and gives the employees higher quality feedback from peers.
Therefore, I would opt for transparent peer feedback. I am not opposed to putting in a field in which people can give confidential feedback, though.
13. Lightweight or heavyweight
A heavyweight review process requires everyone who writes reviews for employees and 360 feedback “up,” “down,” and “sideways” to spend tremendous amounts of time doing so.
Google’s process started out pretty lightweight but became more cumbersome as the company grew. By the time I left in 2010, I used to spend four hours writing the review for each of my direct reports.
Even though I only had six direct reports, that was still 24 hours twice a year; plus, I had all my peer feedback to write. It wasn’t as though the demands of the business slowed down just because it was review season.
That meant there were four weeks each year when everyone was in a foul mood. People stayed up too late doing reviews and got colds. The colds spread. We all started to dread review time, to call it “perfcrastination” because other things didn’t get done.
While I do think that forcing myself to spend this time did make me think more deeply about how I was managing the team, I am not sure that the toll this process took on the organization was in the end worth it.
Word to the wise: even if you put in place a good process, processes develop a life of their own if you don’t actively manage them. They’re like a fast-growing vine. They need to be trimmed.
Therefore, I would recommend instead a lightweight review process that happens twice a year, with active monitoring of how long the process takes, and trimming to make sure it doesn’t grow beyond what you intend.
Remember, the process will take less time and be more radically candid if all 360 feedback is transparent. Not having to synthesize and anonymize the 360 feedback will cut down on the amount of time managers have to spend writing feedback for each direct report.
A lightweight review tool would look much like the 360 tool I described above. It would ask managers to rate their employees for each category and ask them to input text when the employee is in the bottom or top two ratings.
The overall rating would be auto-calculated, and the manager’s rating distribution would be compared to the expected distribution. If the manager’s distribution falls outside of what’s expected, the manager must explain.
The manager should be able to fill in the review form for each employee in under 30 minutes.
If the organization stresses the importance of regular, impromptu feedback, much of the need for a heavyweight review process disappears.
Here are the elements it would include:
- The manager would get a summary of the peer feedback which they could take a look at before creating the final assessment. The report would show how many people gave which rating, and if the manager clicks on the rating, they’d see the comments.
- Then the manager could fill out their assessment of John Doe’s performance. If the manager’s assessment varies dramatically from the team, the manager would be prompted to explain.
- As a last step in the process, all managers who lead more than 30 people would get a message if their team’s peer assessments or their own assessment of their team is out of the expected bounds.
I hope this helps you think about how the kinds of development conversations I described in Radical Candor can work together with a performance management system to create a culture of Radical Candor.
However, I spent my career as an operating executive, not an HR professional, so I would love to get feedback on this from those who have more experience than I do. Please send your thoughts to RadicalCandor@RadicalCandor.com.
This post was adapted from the second edition of Radical Candor: Be A Kick-Ass Boss Without Losing Your Humanity.
Need help practicing Radical Candor? Then you need The Feedback Loop (think Groundhog Day meets The Office), a 5-episode workplace comedy series starring David Alan Grier that brings to life Radical Candor’s simple framework for navigating candid conversations.
You’ll get an hour of hilarious content about a team whose feedback fails are costing them business; improv-inspired exercises to teach everyone the skills they need to work better together, and after-episode action plans you can put into practice immediately to up your helpful feedback EQ.
We’re offering Radical Candor readers 10% off the self-paced e-course. Follow this link and enter the promo code FEEDBACK at checkout.
Watch the Radical Candor Videobook
We’re excited to announce that Radical Candor is now available as an hour-long videobook that you can now stream at LIT Videobooks. Get yours to stream now >>