Chicago Booth Review Podcast How Should You Deliver Negative Feedback?

George Wu: I think sometimes when you say, "I enjoyed your essay, I enjoyed reading your essay and there were lots of really good ideas, but," they're really expecting the but.

Hal Weitzman: How would you tell one of your direct reports that their work just isn’t good enough? Especially if that person seems blissfully unaware that they are underperforming?

Welcome to the Chicago Booth Review Podcast, where we bring you groundbreaking academic research in a clear and straightforward way. I’m Hal Weitzman.

Today, we’re launching a new podcast miniseries, the audio version of our popular Business Practice column, where we asked people to script what they would say in a challenging workplace scenario, submit their responses, and rate each other’s contributions. Then we bring in Chicago Booth behavioral science professor George Wu to analyze the results.

Today, George is talking about the delicate task of delivering negative feedback, or constructive criticism. First, let’s hear the scenario.

Reader: You manage a team of analysts at Trendline, a market research firm. Their role is to collect data on the restaurant industry and compile it in written reports. The reports are recognized as among the best in the business, in part because they're easy to read. Because of this, you value your analysts as writers, as much as researchers. You hired Stephanie as an analyst about six months ago and several of her assignments have come due.

She has mentioned that it was a joy to write the reports and that she's excited to get feedback from you, but there's a problem. Her prose is awful, almost unreadable. It has problems that go way beyond the straight typo. This morning she came by your office and asked how you like the reports. You hardly knew what to say, so you asked her to come back in the afternoon to talk. She smiled, said yes, and bounced out the door. What do you say to her when you meet?

Hal Weitzman: All right, George. This is a situation that many of us will have faced. What makes it so difficult?

George Wu: Well, it was set up in a way where poor Stephanie or Steven, we'll talk about that in a bit, is a bit clueless. Maybe very clueless. And so I think because she's clueless or he's clueless, that telling that person anything that resembles the truth is going to be very surprising to that person. So that's going to create a situation where one, it's just going to set up a situation for conflict and it may be something where you don't expect them to be exactly very receptive.

I guess the other thing which behavioral science has suggested is that we oftentimes think of these situations as likely to be unpleasant. In many situations, we overestimate how actually unpleasant they are. They're probably going to be unpleasant, but not as much as you think. But if you believe that something is going to be really unpleasant, you don't actually engage. And so because you never engage, there's no opportunity to actually calibrate yourself and learn that it's actually not as bad as you think it is.

Hal Weitzman: That certainly plays out. In the corporate world, you would say that there's very little feedback and what there is is pretty insubstantial. Good job. And so negative feedback is really a much harder thing, isn't it?

George Wu: It's much harder. And in some ways, let's just say people don't experiment on making feedback a little bit harsher or maybe a little bit more truthful or whatever because they have this idea, sometimes appropriate, sometimes not, that feedback won't be taken well.

Hal Weitzman: Like all business practice columns, we had people vote on what they thought was the most effective and least effective responses. What were the kind of general principles that came out of that?

George Wu: Well, it's interesting that the responses had certain elements. Obviously these elements were used differentially. A lot of people use what's oftentimes called a sandwich approach. You kind of start positively, you get the real good stuff, which is the stuff that's negative in the middle, and then you end with something that is positive. Others basically had some kind of overture or offer to mentor.

Others had specific help on writing. Some things were just about trying to express positivity and enthusiasm, maybe sugarcoat some of these kind of things. And then some things were actually pretty direct, like negative or condescending tones, or maybe even more directly like suggestions that if you don't shape up pretty quickly, you will be eliminated from this organization. I think the threat.

Hal Weitzman: A threat can be quite motivating but perhaps isn't optimal. So what did we find that the people who voted in this poll found was optimal?

George Wu: Well, I'll start with the negative stuff. You probably figured out that threatening was not considered exactly the thing that anybody should be doing. So negative tones and threats and things like that. It's interesting that a lot of people used, probably the second most common thing was to use a sandwich approach. No effect. So whether you use sandwich or you didn't use sandwich had no effect. But I think the most positive was to express some desire to mentor and help this person to do a better job.

Hal Weitzman: So just to go back to the sandwich thing, because that is so common. I tell someone something that they're doing really well, then I deliver the negative feedback, and then I try to end on an up note. But that makes no difference?

George Wu: Well, we do that with our students when we give feedback. So it makes no difference, which was a little surprising. I would've thought that it would be positive just because it is so recognizable. But I think it's because, this is a suspicion, but I think it's because it's just so transparently instrumental in some way that I think that people see through it.

Hal Weitzman: Are they waiting for what's in, I don't want to say what's normally in the sandwich, but are they waiting for what's in there? Is that the point? If I call you into my office and say, "George, you know I think you're terrific," then you ignore all of that and you are wondering what the hell this is about. And then I come to the real thing and by that point, when I turn to the positive part at the end, you are already devastated.

George Wu: I think sometimes when you say, "I enjoyed your essay, I enjoyed reading your essay and there were lots of really good ideas, but," they're really expecting the but.

Hal Weitzman: So am I right in saying that the optimal response according to our people who weighed in here, was for you to offer to be a mentor regardless of the structure?

George Wu: Yeah, I think so. And certainly the structure, I don't know how general that is. There are things that surely the structure matters, but probably in more subtle ways than just the sandwich. And I think the sort of overwhelming at least conclusion from this particular scenario, is that it's important to not just criticize, but to provide constructive ways of getting out of that, which is feedback and mentorship.

Hal Weitzman: Now we also collected some interesting data here about gender. Depending on whether the employee in question was male or female. What did we learn there?

George Wu: Yeah, so this was a tricky one. We actually, somebody programmed this so that people randomly either got a person called Stephanie or a person called Steven, and only later on did they find out when they read the column that there were actually two different people. And so I think there are a couple of things that are different and interesting, is that the first one is a little disturbing, which is that the overtures to be a mentor were differentially applied to the man and the woman. So 66% of those who got Steven, the man, offered to mentor. And 47% who got Stephanie, the woman, offered to mentor in that situation. So there was a difference of 20% in the willingness to mentor. And that was obviously in favor of being much more open to mentor men than women. There were two other gender asymmetries so to speak, which is that unfortunately the call for termination, it's a small effect, but only 3% threatened to terminate Steven. 11% threatened to terminate poor Stephanie.

And that in terms of being praised, there was a sense that there was the opposite kind of gender effect, which is Stephanie was more likely to be praised, about 47%, than Steven, 22%. But although that might be interpreted as being positive, I think that oftentimes praise might be seen as a mistaken belief that women are unable to accept real feedback. And so what we need to do is kind of sugarcoat stuff because women can't take it. And obviously to the extent that that's not true, that there's going to be really differences in terms of the feedback that men and women get.

Hal Weitzman: So let's turn to the actual responses. What were the one that was highly rated?

George Wu: So this one was rated 5.4 out of seven, which is by business practice standards very good. So they got Steven. So, "Hi Steven, thanks for dropping in. I have been going through some of your reports and I am so glad to have you on the team. You have generated quite some valuable insights from the restaurant industry data and I'm sure it will be useful for anticipating future trends. It's great that you are proactively seeking feedback and I did want to confess that the quality of pros in those reports does not do any justice to the valuable insights you have come up with.

Engaging in lucid pros is one of the major reasons our reports are considered valuable apart from the industry insights. I believe we can easily magnify your impact by working the language a little as the underlying analysis is great. I found a few online learning modules that may be helpful as refresher courses as you begin editing. You can also reach out to our editing teams directly for advice. Please let me know if you can think of anything else that we can initiate to help resolve this on priority. I love your energy and passion to help our customers and I'm excited to support your progress on this journey."

Hal Weitzman: This is a sandwich, but the part in the middle of the sandwich is delivered always with sort of the sense that it's constructive criticism but also saying you can do better. Your insights are so great, let's just bring the language up to match-

George Wu: Yeah, absolutely.

Hal Weitzman: ...the excellence of the insights. What were one of the less highly rated responses?

George Wu: So this one was near the bottom, but, "As an analyst you should understand that your essential job is to report on complicated business with concise and clear language. I don't doubt your ability in analyzing situations and I like your passion for the project. However, you have to ensure your writing is well organized in order to make the progress more efficient. Next time, come back with better formatting."

Hal Weitzman: Which sounds like a bit of a threat.

George Wu: And pretty direct.

Hal Weitzman: All right, let's turn to ChatGPT, because you do this fun thing where you put this scenario into ChatGPT to see how an AI respondent would advise that we go. Or what did you get there?

George Wu: One thing that's interesting is that it came up with categories of things to focus on. So it says "Start positively," so it believes in a sandwich. "Express appreciation. For example, highlight the positive aspects of her reports first. Be specific, use I statements, offer guidance, encourage questions, emphasize growth, set clear expectations, and positively." So I think most of those things, I think are good. How to actually put them all together is not necessarily very easy.

I think the one other thing which I did for this one is I asked it to anticipate what they thought were the big risks. So here's what it gave and we can talk about this. "Impact on confidence." So it may undercut Stephanie's confidence. "Responsiveness to feedback. Individuals vary in their openness to feedback. Communication breakdown, time management, overwhelming, fear of judgment, resistance to change, cultural sensitivity, lack of follow-through, team dynamics."

I think some of those things are pretty interesting. I think one of the last things I asked it was what should I do if Stephanie gets defensive? And it said, "Stay calm and be empathetic, use I statements, clarify intentions, focus on specifics, encourage open dialogue, ask questions, reassure support, take a break if necessary." So I think most of this is pretty good advice. It's certainly, I think at the very least, it focuses people's attention on things that may be obvious after the fact but are easy to miss.

Hal Weitzman: So I noticed that ChatGPT pointed out the importance of I statements. What would be an example of an effective I statement in this scenario?

George Wu: Well, I'll give you what ChatGPT thought was an effective I statement. "I found it a bit difficult to follow the flow in some places. It's not a critique of you as a person, but rather a suggestion for improvement in the way we present the information."

Hal Weitzman: It's kind of the active listening type. I hear you saying this. Okay. And the point there would be to put any misunderstanding on me rather than on you, and to make it not about blaming you.

George Wu: Though of course the inference is that I am a pretty common reader of the kinds of content that you're putting forth.

Hal Weitzman: Right. As an expert, I didn't follow what the hell you were talking about.

George Wu: Absolutely.

Hal Weitzman: So George, what do you think are the big lessons that we can draw from this scenario from behavioral science research?

George Wu: So I think one of the most interesting things are the gender effects, which are very subtle. I mean, this was not telegraphed as something that was about gender. We just assigned some people a man and some people a woman. And we saw that there were differences in terms of how people behave. And I think that those kinds of differences, although this particular scenario has not been studied, are in some sense indicative of the kinds of gender differences that we oftentimes see. And I think part of it is just to be aware that you may be acting very differently for a man and a woman in ways that I think are not necessarily the way that you would like to act and also not necessarily productive.

Hal Weitzman: Okay George. Well I'm going to give you some feedback. Good job.

George Wu: Hooray.

Hal Weitzman: And we would love to have you back again on the Chicago Booth Review Podcast.

George Wu: Love to be back.

Hal Weitzman: That’s it for this episode of the Chicago Booth Review Podcast. You can find the rest of the Business Practice series on our website, chicagobooth.edu/review/business-practice. When you’re there, sign up for our weekly newsletter so you never miss the latest in business-focused academic research. This episode was produced by Josh Stunkel, and the scenario was read by Madeline Bunke. If you enjoyed this podcast, please subscribe and please do leave us a 5-star review.

I’m Hal Weitzman. Thanks for listening. We’ll be back next week.