As an engineering manager, I’ve been thinking how much sharing my opinion of someone’s performance and skills can influence others around me to think the same.
For example during the weekly meetings I have with my peers at work, if I praise or complain about a person in my area having some behavior, I will shift the perception my peers have of that person.
When you have to manage people, staying objective when assessing a person’s performance is always a challenge, no matter the experience or seniority. There are several traps to avoid, one of them is to rely on reputations too much because although reputations offer convenient mental shortcuts, they also bring their load of subjectivity.
So how exactly are reputations formed, how to verify if someone’s reputation is fair, and how to help bring someone’s reputation closer to what it is in reality?
Reputations are mental models
Everyone has a reputation. Your friend tells you about her new boyfriend being annoying, and even though you never met the guy, you start forming an opinion about him. And the more she tells you similar stories, the more you reinforce that opinion of him in your mind until it becomes his reputation.
A reputation is nothing more than a mental model for a person. It works as such: after a few times of observing or hearing of the same behavior in someone, we build a mental model of who we believe that person to be.
It’s a useful mechanism because if we had to carefully evaluate every single action or word from everyone we interact with, we would be overwhelmed, so we need those mental models to simplify our view of the world.
In the workplace, everybody has a reputation too, which can be good or bad based on people’s actions and how those actions are perceived by others within the common beliefs of the shared workplace culture.
When reputations are exchanged among a group of trusted peers, they can be very effective, because all members of the group can rely on their peers for information, which saves everyone the time-consuming effort of building a reputation themselves. And in turn, sharing information reinforces the trust and bond within the group.
Propagating reputations is also a crucial element of social order in any organization because by reinforcing or sanctioning specific behaviors, an organization can spread those behaviors and have them become the basis of its culture.
Judging people by proxy is hard
When you manage someone, you have to be the judge of that person’s professional performance, which for knowledge workers is difficult because it’s almost impossible to find strictly objective benchmarks. Managers must therefore walk a thin line where they need to be objective while not being afraid of using their judgment, even if judgment is subjective.
It was Korzybski who said that “the map is not the territory”. In other words, no matter how good our models are, they are all imperfect. Research suggests that we are not particularly good at judging other people’s opinions and skills, as the 2006 psychology paper “Making Decisions for Other People” states, “To make judgements about other people, we tend to retrieve knowledge about ourselves and use it to make assessments about others.”
And then when you manage not just individual contributors, but also managers, things become even more complicated. Those managers tell you about the people reporting to them, one skip level away from you, and you have to judge the performance of those skip-level people by proxy, through the prisms of other managers’ judgment and mental models. And same goes for you: you become the source of information about your direct reports to your peers and to your manager, who learn about your direct reports through the prism of your own judgment and mental models.
Of course, you want to trust that your colleagues and friends are honest to you and that the judgments they share of others, good or bad, are honest too. But because we are all humans, we must also trust that our models will always be imperfect, regardless of us being honest about them or not.
So how do you make sure that you are as objective and fair as possible when judging people’s performance by proxy? And how do you make sure that when you talk about people in your organization, you’re not creating an unfair reputation for them, by being either too positive or too negative?
To verify a reputation, gather data
Let’s be honest, nobody’s reputation can match reality perfectly. That’s because again, reputations are mental models, and all mental models are imperfect. So the question is not how to avoid propagating imperfect or unfair reputations, but rather, how to make sure that we constantly adjust the reputations of people around us to make them less imperfect, and get them as close as possible to fairly represent reality as people keep changing.
So how do you verify if someone’s reputation matches reality? Simple, use facts and data. For example, if you hear somebody say that a particular person tends to be cynical, or even if you make that observation yourself, your first step should be to look for data in your past and present interactions with that person. If you have either memories or present facts of behaviors where the person acts cynically, then there is your proof.
However, it’s easy to misinterpret such data, because we all tend to use ourselves as benchmarks to judge others, therefore the data you’re looking at may appear as a fact just because of the subjective standards you have for yourself. And like everything that involves a judgment or assessment, there are many other ways to fool ourselves due to biases of some sort.
Three biases to be aware of
The first bias that is often involved in misjudging a person’s reputation or performance is confirmation bias, the tendency to look for evidence that confirms our existing beliefs. Once we’ve formed a reputation for a person in our mind, we tend to look for behaviors that support that reputation, because it shows how right we were and that makes us feel good about ourselves.
That also means that we tend to filter out evidence that doesn’t match our beliefs, therefore we pay less attention to the absence of behavior, and we fail to recognize when a person has changed.
The second bias is the halo effect, also called the horn effect, when a positive or negative trait of a person influences the general perception of that person. For example, a person might be doing something great in one area, and by association, his work in another area is also judged as great, even if there is no data to support that judgment.
A third phenomenon that’s not exactly a bias, is belief perseverance, when a belief is maintained even in the face of new contradicting evidence. This often comes up during performance reviews when someone’s past performance is used as an input for their present performance.
Let’s take the example of a person who was rated by her manager as “exceeding expectations” during past performance reviews. That person may keep getting high reviews even when she no longer shows the behaviors that would justify the rating, because her manager remains consistent with the outdated reputation.
These are only three of the biases that may come into play, but there are many other ways to fall prey to our own judgment shortcuts. So what are some strategies to help minimize the effects of biases on us?
Talk to others to minimize the effects of biases
If you think your biases might influence your judgment of a person, the only way to be certain is to get outside of your own head and ask for someone’s opinion. If that someone is close to you, such as a peer or your manager, you can be very direct about it: “I’ve observed behaviors A and B in my interactions with that person, I’m interpreting that as X, what do you think?”
You can also gather feedback from people who work directly with the person, mixed across different types of work relationships: peers, reports, customers, and stakeholders. It’s also good to gather feedback in different formats: company feedback tool, emails, face-to-face, etc., because the format will change the qualitative aspect of the feedback you get.
And remember, those people have their own mental models and may have been tampered by the same mis-judgment as you. Therefore when you collect feedback, always look for factual behaviors, not “I’ve heard, felt, or sensed,” but “during that interaction this is what happened, then this, then that, etc.”
Once you have collected data about that person and her behaviors from the perspectives of various people, you want to look for overlapping evidence and opinions, which would confirm the assumption you made. And on the contrary, if the data shows poor overlapping evidence and opinions, or little mention of the behavior at all, then you will have to accept that most likely your judgment was wrong, and it’s time to update your mental model of that person.
A final step is to take the time to reflect and figure out what assumptions you tend to make and where those assumptions come from, along with what biases are frequently influencing you. By doing this, you will learn about yourself over time and will be able to implement safeguards to avoid falling for the same assumptions and biases in the future.
Someone has changed? Talk about it!
The problem with reputations is that sometimes they maintain themselves even when the behaviors or people that sparked them are long gone. Worse, sometimes a person unconsciously understands the false reputation he’s been given, and to be consistent with what is expected of him, lives up to that reputation which becomes a self-fulfilling prophecy. So let’s imagine you’ve just discovered that the reputation of a person doesn’t match reality well enough, what do you do next? It’s very simple: talk about it!
During a staff meeting with your peers and manager, or during a 1-on-1 with a colleague, just take two minutes to help clarify someone’s reputation, either on the positive or negative side. It’s as simple as saying a few words along the lines of:
- “All the feedback we gave her really paid off, she worked hard on improving her behavior and I haven’t seen her do it for a few months now. She has changed!”
- “I thought he was cynical, and I’ve been saying that to you a lot, but I was wrong.”
- “It is true that the impact of his work exceeded expectations last quarter. He’s done great things during this quarter, but I do not see the data that supports giving him a rating of exceeding expectations again for this quarter.”
And sometimes, the responsibility doesn’t lie in a single person, but in the collective, “we’ve been too hard with how we judged her behavior, the impact wasn’t that negative, we should reconsider how we view her.” Fixing someone’s reputation takes time, and you will likely have to repeat the message multiple times until everybody’s mental models are adjusted.
In the case where a person’s behavior is negative, another important step is to reach out to that person, and give her feedback on her improved behavior. Then, also tell the person that she should work on communicating directly to others around her that she has worked on improving her behavior, and that she hopes people are no longer seeing her as her old self.
So today, take five minutes to think who within your organization has an overly good or bad reputation, and reflect on whether or not that reputation is factual and matches reality. And if you any have doubts, gather data from yourself and others, adjust your mental models accordingly, and talk to people around you to help them adjust their mental models too.
Nicely written Manu, being a leader or a manager one has to be very careful about the unconscious biases. Even when you ask someone “I’ve observed behaviors A and B in my interactions with that person, I’m interpreting that as X, what do you think?” You may be injecting the reputation bias in that person. Maybe we should start by asking “Have you interacted in person with the person Y and what do you think of that person?” Then go with your observed behavior and ask your opinion again.
There is no right or wrong approach I think, it is also an agile process. Every person is different thus they should be managed differently, using any kind of predefined template usually creates problems.