Name: Is the Performance Review Broken?
Uploaded: 2026-05-20
Duration: 18 min 33 s

Performance reviews have a reputation, and not a good one. For many employees and managers, they feel like a bureaucratic exercise that creates more anxiety than insight. What if the problem isn’t feedback itself, but the way organizations structure it? This conversation takes a hard look at how companies are rethinking a dreaded once-a-year obligation to focus on real conversations, clearer expectations, and continuous development. Learn what great performance management actually looks like— and what's standing in the way of getting there. 	 

David Hassell, Chief Executive Officer, 15Five
Brooke Weddle, Senior Partner, McKinsey & Company  
Moderator:  Adam Banicki, Head of Video, Fortune

Is the Performance Review Broken?

Conferences

OK, I wanna start with a, a show of fans. Uh, who here thinks performance reviews are transparent, fair, and inspire better performances? Uh oh, not a lot of hands. Well, according to a recent Gallup poll, only 1 in 5 employees believe they are. So let's dissect where the disconnect is. Brooke, David, thank you for being here. You both work at companies trying to improve, improve performance. What is the most common thing you see not working with performance reviews today? Well, I, I'm happy to start. I think as the room, uh, clearly senses there are many things that are not working, and I think the key point today is that most performance systems were designed for a world that is moving more slowly than the world in which we are working in. And um I can think of a couple of failure modes. uh I would say annual performance reviews, uh, definitely a failure mode, not delivering feedback uh in a continuous way looking for measures of productivity proxies that don't actually get at outcomes and so when you shift to what we think the future is there are a couple of, um, big ideas. One is first of all. Of understanding how the outcomes are changing and the skills and capabilities required of today's workforce look differently, so McKinsey found that AI literacy has jumped in reported reported importance by employers by 70%. Um, that is just one capability. Certainly others like learning agility, uh, are also spiking. The other key. Uh, part of the future, if you will, is the performance review process itself, um, and how we are using AI and act to actually transform some of those processes, uh, and leading organizations across sectors are, um, I would say beyond experimentation at this point and actually hardwiring those processes different, uh, differently through tech, uh, technology enablement. David, what about you? You know, when I look at the, the statistics around this, um, you know, the vast majority of managers and employees, I think it was a, a, a Deloitte study that showed that they just don't trust the process. Um, I think it was Gallup that said only 2% of CHROs actually believe that their performance management practice is actually effective. I mean, that's stunning. So maybe there's 2 out of 100 in this room who think, think it's working. Um, and I think building on what Brook said, uh, you know, the fact that these, this process was designed for a different era is, is part of the issue. Um, annual reviews made sense when the world was pretty static, and we're at an accelerating pace of, of technology and change, and, uh, the, the practice just hasn't kept up. You mentioned changing annual reviews. What should change that process? Are we talking like routine check-ins with managers? Like, where do you see the best option? Yeah, look, 10 years ago there was a big movement toward, hey, we should throw out the annual review and make everything continuous, and 15.5 was at the forefront of that. We build technology for, uh, weekly check-ins and one on one. Ones and things like that and we found and our customers found that we kind of threw the baby out with the bathwater. The continuous piece is absolutely essential, but you do need a period of time where you reflect and you look at you assess performance when you can take a step back. So I think it really is the combination of those two things, but the annual cadence is part of the problem, um, you know, the, uh, the reason it's typically is done annually is because it is so onerous to, to do one of these practices. And um what needs to happen is it needs to be simplified and it can be more frequent. We have some customers doing it as many as 4 times a year. I don't recommend that for everybody, um, but at least twice a year with a lighter weight process, uh, combined with the kind of the weekly feedback is very important. And I know you're both, uh, using AI to help this process. Brooke, how is McKenzie implementing this and how is it being received? Yeah, well, I, I can talk about what McKinsey is doing, but I, I can also say that it is representative of what other organizations are doing as well. So one of the reasons why the annual performance review feels so heavy is because there is a lot of transactional work that goes into a performance review. Everyone in this room knows that. And so what McKinsey is specifically doing is um using AI to intake data from 360 evaluations um one of the things that McKinsey places a lot of importance on this won't surprise anyone is of course it's human capital. Without our human capital and expertise. We have no business model. So in a performance evaluation you were talking about, you know, upwards of 20 interviews, right, that you are intaking data, uh, to evaluate a candidate. It's even greater when you think about kind of milestones in someone's career like partner election. Uh, so you, you intake, uh, that, that data AI actually is used to generate, um, the, the content for the review, an initial version of course the human is, is most clearly in the loop here and that content generation is actually synthesized in the context of our leadership development model and increasingly we are using. Using um our tools to drive more predictive, uh, analysis as well so that I as an evaluator, um, will have, um, some steers right on how to interpret, uh, the content that, uh, is being teed up for me and of course, um, that helps with things like biases and my own blind spots in terms of my pattern recognition. So I'm, I'm curious, like for many employees when they're going through the, uh, performance review process, they feel like managers become robotic as they're, as we're implementing AI into this process. How is that communicated to employees in a way that makes them feel comfortable and that this is still a human conversation? Uh, you know, one of the things that we're really focused on is, is, you know, I, there's a, there is a big conversation about there, but what is the role of AI in all of our work, right? And, uh, we're, we're firm believers that AI should be, uh, used to elevate humans and to elevate our judgment and to, to, uh, building on what Brooke said, keeping humans in the loop. And so, you know, one of the things that we are, uh, focused on at 15.5 is figuring out how to, uh, use AI to streamline the process to, to, to have managers have what they need to be able to make better decisions, um, and, uh, and so, you know, right now actually, uh, we're in June we'll be releasing a product. That now integrates into about 18 different systems of work, uh, so that managers can be in real time given insights into, hey, this is an opportunity for feedback potentially over here or hey, here's, here's an area that you may wanna look at. There may be some blind spots, but really keeping the, the full onus on the manager as the, as, as the person with the judgment. Yeah. So I think we've talked about trust and, and employees want clarity and accountability. Where do you both stand on ratings at reviews time? I've always appreciated that that grade at the end of the year that sort of told me where I stand with my manager. But, uh, is, is that a good thing? Is it a bad thing? I'm curious. Yeah, well, we, we've spent some time discussing this, uh, and, um, the McKinsey data on this is, is pretty clear that, um, ratings are actually. not the centerpiece of an evaluation for an employee, uh, and in fact there is no negative stigma around, uh, ratings for most of the, um, global, uh, employees that we, uh, surveyed a couple of years ago. The negative stigma comes from perceptions of unfairness, from opaqueness, uh, in, in the process. I don't understand what is the difference between a strongly performing versus a distinctive, uh, performance, and, and. So ratings actually give people the opportunity to understand where they stand uh and I think that's quite important and and the data would back that up and in fact in my personal experience uh working with many organizations that moved away from ratings they in fact continue to use shadow ratings in their internal discussions which I think uh defeats the purpose entirely um we need to share where people. Are, uh, in terms of their evaluation but then couple that with strong focus on development I would say actually that the competitive advantage today of most leading organizations is derived by how quickly they can help develop human potential, uh, and that actually going back to your question on communication should be I think centerpiece of any kind of communication on performance based AI um that you know companies are using. Just wanna build on that, um, you know, uh. We went through a movement also probably about 10 years ago where we saw a lot of organizations say hey we should get rid of the ratings and to, you know, to what Brick shared about the, the shadow ratings we saw that's true. I mean companies need to assess their people and make talent decisions. So whether that's a rating or some other methodology as an organization as an HR leader, you have to do that. And so to not share that with the person I think does create a lot of, uh, um, challenge in, in the employee-employer relationship. Because there is this lack of transparency we found, um, one of the challenges we've seen with ratings is just uh it is a communication problem because sometimes people say, you know, on a 5 point scale a 3 is actually pretty good that's where you should be and then there's, you know, kind of deviations aside from that we all grow up thinking like we should get A's and you know we should all be a 5 and so I think the, the challenge is in the perception of the conversation, not necessarily the rating itself. I, I once had a manager that said you should get maybe 1 or 25 ratings and you're nearer and, and a 3 it means you're performing really great and you're, you're showing up every day so I understand that. Uh, I'm curious if companies remove ratings, how should they handle compensation, performance, promotion, those sort of things? Well, even if a company chooses to remove ratings, I still think the centerpiece needs to be around this mismanagement. dialogue and um I've seen increased importance on this notion of dialogue or interlocution right um whether you're interacting with a robot robot uh an agent or a human this is a new capability that's emerging. And I don't think managers are um equipped in all cases to be able to engage in a thoughtful dialogue. So certainly things like pushing pushing transactional activities first and foremost to things like AI will help. Second, um, there is typically some managerial upskilling that needs to happen and I, what I've seen work really well is positioning that in the. The context of broader cultural transformation, so you don't get managers to emphasize and role model continuous development if that's not a cultural priority led and role modeled by the CEO and then finally there needs to be accountability. So if we're gonna ask managers to do continuous feedback better, we better hold them accountable in their own performance reviews and talk about what great looks like. I, I'm curious, like when we think about asking managers to do more, uh, are we asking too much of them if, if it's going to be a continuous feedback loop? Is that more work being put on, on management to make sure that they're delivering these results? David, what do you see, you know, I, I read, I read some statistics that showed that, uh, in, in many organizations managers spend upwards of 210 hours a year doing performance management like the formal practice, and it's mostly backward looking, and they spend only about 13% coaching. Uh, in, in real time, which I think those numbers really need to get flipped, and so I do think that. Uh, we put too much on managers when we have these big onerous practices that don't actually deliver the value which is more forward looking than backward looking. So how do we create systems and, and, and, and a culture where managers have the appropriate training because every manager that comes into your organization has, you know, baggage of training from whatever other organizations they have. So you have to get them aligned around your ethos, uh, and then, and then how do we, um, you know. Really empower them with, with the, the data and the intelligence and what they need to make, uh, to, to have those coaching conversations ongoing in the flow of work and not as some additional thing that they need to do. Yeah, at a place like McKinsey, tens of thousands of employees, how important is it to create a model that's scalable to all of those employees and how does AI help with that? Yeah, I mean, it, it, you know, we, we do do continuous feedback. We have, um. You know biannual, uh, reviews as well that serve as kind of anchor points, uh, but, but like I shared we we are using AI to again, you know, kind of shift some of that transactional work and holding people accountable, um, to, to really spending time, uh, with their people and their teams, uh, to try to. Drive development. I also think there's a mindset shift here which is when you think about the manager irrespective of the company, it, I, I think the mindset shift is viewing development as performance, right? So we're development is not separate from managing someone's performance. If you're managing their development, you are. Actually managing their performance and I think that kind of shift is important if you're really going to embrace the notion of adaptability, uh, continuous learning, learning quotient. There's all sorts of buzzwords out there to get at this notion, um, but that, that would be a core kind of cultural or mindset underpinning that I think is quite important. What separates a useful performance conversation from a bureaucratic one when you're going through the motions and doing this continuously? You know, I think the first thing is, you know, and, and, and this actually I think really highlights the difference between the, the performance review process which is, you know, this big conversation, it happens once a year, typically very, you know, very much subject to recency bias and, and, and whatnot. Uh, versus, uh, a conversation that's happening in the moment about something that's happening in real time. Most of that growth, I mean, if you're gonna have a conversation once a year about something that happened 3 or 4 months ago, um, you're not really going to be able to help that person develop around that one thing. But if you're noticing it and calling it out and have a conversation in the moment, you have this opportunity to, to, to shift the behavior. Uh, what it takes, however, is the managers, right, having a good practice of behavior around that, and you know, we tried to solve that problem with our. 15.5 check-ins and one on ones 10 years ago, but, uh, it was really hard for the, the, the, the software was great, but the, the change of management was the hard piece and so, um, so now I think AI now makes it even uh more possible than ever to provide the right kind of insights and nudges and things to managers to be for them to be able to have that conversation. That's the problem we're solving now, um, and I think what again it comes back to making sure they're, uh, addressing something that's happening in real time, yeah. So I mean you could go in so many different directions with that question. I think we've all had a review where you're kind of like wow, that's, uh, that was not what I had anticipated. I think, um, where when I think of a bad review I think of kind of something that you know is awash in platitudes and niceties where you walk away not actually knowing what the thing is that you need to be working on and um. Interestingly enough, we're we're also seeing AI be used to try to help managers and leaders be more specific, be more intentional about the feedback that they are giving, um, so AI giving feedback to humans on how to deliver feedback that's essentially what I'm saying, and that is happening. Um, I think the other thing where, um, you, you see kind of the failure mode is having the feedback feel disconnected from, from the outcome or from the, the flow of work, um, and going back to our research again, one of the key things that we found is that over 70% of employees wanna have clear goals as a part of their continuous dialogue and process so that would be the other kind of, you know, failure I would say that. Often happens and again connecting that back to specific outcomes is going to help drive specificity. Yeah, I just want to build on that. I think the, I think having, like you said, the clear goals and clear expectations are the foundation and, and what needs to be in place in order to have effective feedback, right, because otherwise, um, if someone doesn't know what's expected of them and where they stand, it's challenging to deliver, uh, something that's going to be useful for them to understand how can they grow and, and, and meet those expectations better over time. OK, I wanna end with um one last question, uh. What is one thing leaders should stop doing and start doing as they think about their review process this year, David? Uh, let's see, stop and start. I would say, um, uh, well, let's start with start. I think that, uh, companies should really think about, um, lightening up the processes as much as possible, so the annual process, uh, making it as lightweight as possible, and then really focusing on that continuous, uh, feedback, uh, throughout the year. There should never be a surprise at review time, right? So, uh, really at that point, you should, you should be documenting the past, but organ organizing the entire process around the future. How can we grow? How can we be better in the future? And, uh, you know, yes, you have to do assessments, but, but, but, but limiting the, the time investment on that. I would say uh stop creating kind of endless steps in the process fundamentally take a clean sheet to your performance evaluation process starting back back looking back from the outcomes that you want to achieve, um, you know, taking a page from our our work on reimagining work flows that's what I. I would, I would stop creating the process. I would start reimagining and working back from those specific outcomes you want to achieve. And I would bet that some of the outcomes you want to achieve are actually new skills and capabilities that you want your workforce to embrace and then designing systems to help actually those employees live into those capabilities and skills. Brooke, David, thank you so much. Thank you. Thank you.