WEBVTT 00:00:00.000 --> 00:00:02.080 Welcome back to the Deep Dive. I want to start 00:00:02.080 --> 00:00:04.240 today with a scenario that I think is going to 00:00:04.240 --> 00:00:08.800 trigger a legitimate physiological response in 00:00:08.800 --> 00:00:10.779 about 90 % of our listeners. You're starting 00:00:10.779 --> 00:00:13.679 with a threat. I like it. Bold strategy. It's 00:00:13.679 --> 00:00:16.519 a Tuesday afternoon. You're in the zone. You 00:00:16.519 --> 00:00:18.359 know, you're in that flow state, just getting 00:00:18.359 --> 00:00:21.820 work done. And then suddenly your phone buzzes. 00:00:21.820 --> 00:00:24.239 A calendar invite pops up. It's from your boss. 00:00:24.399 --> 00:00:26.679 And the subject line is just two words. Oh, I 00:00:26.679 --> 00:00:31.129 know they are. Performance review. Yeah, that's 00:00:31.129 --> 00:00:33.450 it. That is the corporate equivalent of seeing 00:00:33.450 --> 00:00:35.750 a police car flip its lights on behind you. It 00:00:35.750 --> 00:00:38.329 really is. Even if you haven't been speeding, 00:00:38.530 --> 00:00:41.450 your heart rate just spikes, your palms get a 00:00:41.450 --> 00:00:44.409 little sweaty. It's a visceral biological reaction. 00:00:44.429 --> 00:00:46.990 It's the principal's office for adults, 100%. 00:00:46.990 --> 00:00:49.289 And what is so fascinating to me is that this 00:00:49.289 --> 00:00:51.710 isn't just a you problem or a me problem. No, 00:00:51.710 --> 00:00:55.850 no. It is universal. And here's the kicker. The 00:00:55.850 --> 00:00:58.920 managers. They hate it just as much as the employees 00:00:58.920 --> 00:01:01.420 do. Oh, absolutely. Their research backs that 00:01:01.420 --> 00:01:06.480 up entirely. It is this ritual that consumes 00:01:06.480 --> 00:01:10.739 millions of hours and billions of dollars. Yet 00:01:10.739 --> 00:01:13.120 it's widely regarded by the very people doing 00:01:13.120 --> 00:01:17.200 it as uncomfortable, bureaucratic, and a lot 00:01:17.200 --> 00:01:20.180 of the time totally useless. So that is the mystery 00:01:20.180 --> 00:01:23.019 we are solving today. Why do we do this? If the 00:01:23.019 --> 00:01:25.700 performance appraisal is so universally loathed, 00:01:25.700 --> 00:01:29.099 why is it like the bedrock of modern employment? 00:01:29.260 --> 00:01:31.079 Right. We aren't just going to talk about how 00:01:31.079 --> 00:01:33.680 to survive your next review. We are going to 00:01:33.680 --> 00:01:36.010 dismantle the entire machine. We're going to 00:01:36.010 --> 00:01:39.469 look at the economic theories that justify it, 00:01:39.530 --> 00:01:42.290 the absolute mess of cognitive biases that make 00:01:42.290 --> 00:01:45.109 it so unfair, and all the legal landmines that 00:01:45.109 --> 00:01:46.989 companies are just dancing around every single 00:01:46.989 --> 00:01:48.510 time they fill out one of those forms. And we're 00:01:48.510 --> 00:01:50.170 going to find out if we are actually measuring 00:01:50.170 --> 00:01:52.909 performance or if we're just measuring how much 00:01:52.909 --> 00:01:55.170 our boss likes our personality. Well, I think 00:01:55.170 --> 00:01:57.150 the data suggests it might be a little more of 00:01:57.150 --> 00:01:58.769 the latter than you'd probably like to admit. 00:01:58.930 --> 00:02:02.510 I knew it. I knew it. Okay, before we get into 00:02:02.510 --> 00:02:05.200 the... The depression of it all. Let's define 00:02:05.200 --> 00:02:07.459 our terms. Yeah. Because companies love to dress 00:02:07.459 --> 00:02:09.379 this up, don't they? Oh, they do. They call it 00:02:09.379 --> 00:02:11.280 talent connect, career development discussion, 00:02:11.460 --> 00:02:14.099 or my favorite, the quarterly sync. Yes. But 00:02:14.099 --> 00:02:17.319 stripped of all the HR branding, what is a performance 00:02:17.319 --> 00:02:20.539 appraisal fundamentally? At its core, a performance 00:02:20.539 --> 00:02:22.520 appraisal, and it really doesn't matter what. 00:02:22.840 --> 00:02:26.020 fancy name they give it, is a systematic periodic 00:02:26.020 --> 00:02:29.659 process where an employee's job performance is 00:02:29.659 --> 00:02:32.259 documented and evaluated. And it's evaluated 00:02:32.259 --> 00:02:35.620 against a set of pre -established criteria. The 00:02:35.620 --> 00:02:37.960 word documented there feels really heavy. It 00:02:37.960 --> 00:02:39.699 is the most important word. A conversation is 00:02:39.699 --> 00:02:42.560 just a chat. Right. An appraisal is a record. 00:02:42.719 --> 00:02:45.919 It goes into a permanent file. It is the data 00:02:45.919 --> 00:02:48.439 point that will determine your pay, your promotability. 00:02:48.840 --> 00:02:52.060 And this is the crucial part. It creates the 00:02:52.060 --> 00:02:54.939 paper trail that can be used to fire you. So 00:02:54.939 --> 00:02:57.539 despite all the soft language about growth and 00:02:57.539 --> 00:03:00.139 development, it's effectively a legal instrument. 00:03:00.300 --> 00:03:02.800 In many, many ways, yes. But if we step back 00:03:02.800 --> 00:03:04.919 and look at the mission, you know, the management 00:03:04.919 --> 00:03:07.259 theory behind it, it's not supposed to be a weapon. 00:03:07.460 --> 00:03:10.550 OK. What is it supposed to be? Well, the academic 00:03:10.550 --> 00:03:13.129 literature, it generally agrees that the primary 00:03:13.129 --> 00:03:15.849 objective is increasing organizational efficiency. 00:03:16.389 --> 00:03:18.669 Which, I mean, that makes sense on paper. If 00:03:18.669 --> 00:03:20.710 I'm the CEO of a car company, I need to make 00:03:20.710 --> 00:03:22.750 sure the person on the assembly line is actually 00:03:22.750 --> 00:03:25.490 building cars and not... you know, toaster ovens. 00:03:25.490 --> 00:03:27.810 Exactly. It's all about alignment. You have these 00:03:27.810 --> 00:03:30.030 big corporate goals at the top and you need to 00:03:30.030 --> 00:03:32.330 cascade those down to the individual actions 00:03:32.330 --> 00:03:34.930 of thousands and thousands of people. Right. 00:03:35.069 --> 00:03:37.469 If you are working incredibly hard, but you're 00:03:37.469 --> 00:03:39.530 working on the wrong things, you are technically 00:03:39.530 --> 00:03:41.789 productive. But from the organization's point 00:03:41.789 --> 00:03:44.629 of view, you're useless. The appraisal is that 00:03:44.629 --> 00:03:46.949 course correction mechanism. So it's like the 00:03:46.949 --> 00:03:49.740 GPS for the workforce. Right. The recalculating 00:03:49.740 --> 00:03:52.479 voice. Ideally, yes. It's also meant to facilitate 00:03:52.479 --> 00:03:55.379 collaboration, discuss pay, and clarify expectations. 00:03:56.020 --> 00:03:59.199 But there is a deeper, more, I don't know, cynical 00:03:59.199 --> 00:04:02.020 or perhaps just more realistic economic theory 00:04:02.020 --> 00:04:04.280 that explains why this whole structure exists. 00:04:04.699 --> 00:04:07.259 Okay. It's called the principal -agent framework. 00:04:07.659 --> 00:04:10.199 I saw this in our notes. It sounds like something 00:04:10.199 --> 00:04:13.219 out of a spy thriller. The principal and the 00:04:13.219 --> 00:04:16.519 agent. It kind of does, but it's really the foundational 00:04:16.519 --> 00:04:19.980 concept of labor economics. So in this scenario, 00:04:20.220 --> 00:04:22.879 the employer is the principal. They're the one 00:04:22.879 --> 00:04:24.860 with the capital. Got it. The boss. The boss. 00:04:25.000 --> 00:04:28.699 The employee is the agent, the one hired to perform 00:04:28.699 --> 00:04:31.720 a task. And the central problem in this relationship 00:04:31.720 --> 00:04:34.519 is something called information asymmetry. Meaning, 00:04:34.600 --> 00:04:37.800 I know what I'm doing all day, but my boss doesn't, 00:04:37.800 --> 00:04:41.040 really. Correct. The principal hires the agent 00:04:41.040 --> 00:04:44.279 to do a job. But the principal cannot stand over 00:04:44.279 --> 00:04:46.540 the agent's shoulder 24 hours a day. It's way 00:04:46.540 --> 00:04:49.459 too expensive and inefficient. Right. So the 00:04:49.459 --> 00:04:51.579 principal has a blind spot. They don't know if 00:04:51.579 --> 00:04:53.720 the agent is working at 100 percent capacity 00:04:53.720 --> 00:04:55.740 or if they're scrolling through social media 00:04:55.740 --> 00:04:58.399 or cutting corners. So the principal is basically 00:04:58.399 --> 00:05:02.040 paranoid. They're paying for a service that they 00:05:02.040 --> 00:05:05.079 can't fully verify in real time. In economics, 00:05:05.259 --> 00:05:08.060 yes. This creates what's called moral hazard. 00:05:08.839 --> 00:05:11.920 Because the principal can't see everything, the 00:05:11.920 --> 00:05:14.779 agent has an incentive to do the absolute minimum 00:05:14.779 --> 00:05:17.660 amount of work necessary to get paid. Ah, okay. 00:05:17.860 --> 00:05:20.839 The performance review was invented as the solution 00:05:20.839 --> 00:05:24.420 to this problem. It is the mechanism to manage 00:05:24.420 --> 00:05:27.319 the direct effect and response, as they call 00:05:27.319 --> 00:05:29.680 it. It's a monitoring device? Okay. It's a way 00:05:29.680 --> 00:05:32.500 to audit the agent? It connects the agent's behavior 00:05:32.500 --> 00:05:35.899 to the principal's goals. It bridges that information 00:05:35.899 --> 00:05:38.230 gap. And when you understand that, you start 00:05:38.230 --> 00:05:41.329 to realize why these reviews feel so adversarial. 00:05:41.550 --> 00:05:43.870 Because they are. They are structurally designed 00:05:43.870 --> 00:05:46.029 to be an investigation from the very beginning. 00:05:46.250 --> 00:05:49.230 That is incredibly clinical. Yeah. But here's 00:05:49.230 --> 00:05:51.110 a twist from the research that I just did not 00:05:51.110 --> 00:05:53.459 expect at all. We usually hear about how these 00:05:53.459 --> 00:05:55.279 systems are biased, and we are definitely going 00:05:55.279 --> 00:05:57.600 to get to that. But there's an argument that 00:05:57.600 --> 00:05:59.720 these rigid appraisals can actually increase 00:05:59.720 --> 00:06:02.500 diversity. Yeah, this is a really important nuance. 00:06:02.639 --> 00:06:04.819 It sounds completely counterintuitive, right? 00:06:04.920 --> 00:06:06.800 You'd think a rigid system would hurt marginalized 00:06:06.800 --> 00:06:08.959 groups. I would, yeah. But you have to consider 00:06:08.959 --> 00:06:11.439 the alternative. And the alternative is what? 00:06:12.540 --> 00:06:14.860 No system at all. Right. If you don't have a 00:06:14.860 --> 00:06:18.060 formal documented appraisal system, how do promotions 00:06:18.060 --> 00:06:21.339 happen? It's who you know. It's golf. It's drinks 00:06:21.339 --> 00:06:23.800 after work. It's, you know, I like the cut of 00:06:23.800 --> 00:06:27.399 his jeb. Exactly. Informal systems are absolute 00:06:27.399 --> 00:06:30.139 breeding grounds for implicit bias. If there's 00:06:30.139 --> 00:06:33.279 no form, no criteria, no documentation, managers 00:06:33.279 --> 00:06:35.600 just promote the people they feel most comfortable 00:06:35.600 --> 00:06:38.060 with. Which usually means people who look, talk, 00:06:38.220 --> 00:06:40.319 and act just like them. That's how you get the 00:06:40.319 --> 00:06:42.800 old boys club. So the argument is that even a 00:06:42.800 --> 00:06:45.759 flawed form is better than total chaos, because 00:06:45.759 --> 00:06:48.040 at least it forces the manager to justify their 00:06:48.040 --> 00:06:50.860 decision on paper. Precisely. If it's done in 00:06:50.860 --> 00:06:53.500 a culturally sensitive manner, a structured appraisal 00:06:53.500 --> 00:06:55.800 can ensure fairness in hiring and promotion for 00:06:55.800 --> 00:06:58.360 socially excluded backgrounds. It provides a 00:06:58.360 --> 00:07:00.980 clear, objective ladder to climb instead of a 00:07:00.980 --> 00:07:03.480 secret handshake. That makes sense. However, 00:07:03.620 --> 00:07:05.839 not everyone agrees that individual performance 00:07:05.839 --> 00:07:07.899 is even the right thing to be measuring in the 00:07:07.899 --> 00:07:09.990 first place. I'm looking at the notes on labor 00:07:09.990 --> 00:07:12.790 unions here. Ah, the stakeholder view. Yeah, 00:07:12.889 --> 00:07:16.970 unions often have a deep philosophical disagreement 00:07:16.970 --> 00:07:19.490 with the very concept of the individual review. 00:07:19.790 --> 00:07:22.569 And their argument is what? They argue that in 00:07:22.569 --> 00:07:24.970 a modern workplace, individual performance is 00:07:24.970 --> 00:07:28.069 largely a myth. A myth. That's a pretty strong 00:07:28.069 --> 00:07:30.589 word. Well, think about your own job. I mean, 00:07:30.589 --> 00:07:32.269 think about it. How much of what you achieve 00:07:32.269 --> 00:07:36.720 is solely 100 % because of your own effort? Not 00:07:36.720 --> 00:07:39.639 much when you put it like that. If IT doesn't 00:07:39.639 --> 00:07:42.579 fix my laptop, I can't work. Right. If marketing 00:07:42.579 --> 00:07:44.680 doesn't bring in the leads, sales can't sell. 00:07:45.040 --> 00:07:47.240 If the receptionist is rude to a client, the 00:07:47.240 --> 00:07:49.420 deal falls through before you even walk in the 00:07:49.420 --> 00:07:51.819 room. We are all entangled. My output depends 00:07:51.819 --> 00:07:54.540 entirely on your input. Exactly. We're a collective. 00:07:54.980 --> 00:07:57.360 Unions will often argue that trying to isolate 00:07:57.360 --> 00:07:59.740 one person's contribution and then grade it on 00:07:59.740 --> 00:08:02.779 a curve isn't just inaccurate, it actively destroys 00:08:02.779 --> 00:08:04.779 the solidarity needed for the work to actually 00:08:04.779 --> 00:08:07.279 get done. Because it encourages people to say, 00:08:07.399 --> 00:08:10.660 that's not my job, instead of helping the team 00:08:10.660 --> 00:08:13.040 out. It pits people against each other. That's 00:08:13.040 --> 00:08:15.699 a fascinating tension. The company wants to measure 00:08:15.699 --> 00:08:18.079 the agent, but the agent is actually just part 00:08:18.079 --> 00:08:20.740 of a hive mind. And that leads us to the practical 00:08:20.740 --> 00:08:24.439 question. If we're going to do this, if we're 00:08:24.439 --> 00:08:27.569 going to try and measure the agent. How do we 00:08:27.569 --> 00:08:30.250 even do it? How do we collect the data? Right, 00:08:30.310 --> 00:08:33.190 the how. The sources break this down into three 00:08:33.190 --> 00:08:36.570 basic methods. And as we go through these, I 00:08:36.570 --> 00:08:38.970 really want you, the listener, to think about 00:08:38.970 --> 00:08:41.889 which one your company uses. Let's start with 00:08:41.889 --> 00:08:45.129 the one that seems the most fair on the surface. 00:08:46.009 --> 00:08:48.149 Objective production. This is what I call the 00:08:48.149 --> 00:08:50.419 widget count. It's very straightforward. You 00:08:50.419 --> 00:08:52.860 measure sales figures, units produced, lines 00:08:52.860 --> 00:08:55.360 of code written. The hard numbers. Exactly. It 00:08:55.360 --> 00:08:58.000 deals with unambiguous criteria. You sold $100 00:08:58.000 --> 00:09:00.620 ,000 worth of paper or you didn't. Numbers don't 00:09:00.620 --> 00:09:02.700 lie. Numbers don't lie, but they also don't tell 00:09:02.700 --> 00:09:05.120 the whole truth. The source material points out 00:09:05.120 --> 00:09:08.899 a massive flaw here, opportunity bias, or as 00:09:08.899 --> 00:09:10.860 it's sometimes called, criterion contamination. 00:09:11.480 --> 00:09:13.200 Okay, can you give me an example of that? Sure. 00:09:13.519 --> 00:09:15.720 Let's say we are both selling the exact same 00:09:15.720 --> 00:09:19.340 car. You are assigned a territory in a booming 00:09:19.340 --> 00:09:22.080 city center with tons of foot traffic. Okay. 00:09:22.159 --> 00:09:25.440 I am assigned a territory in a rural town where 00:09:25.440 --> 00:09:28.220 the main factory just shut down and pretty much 00:09:28.220 --> 00:09:30.519 everyone is unemployed. Right. I see where this 00:09:30.519 --> 00:09:33.200 is going. You sell 20 cars. I sell two. Are you 00:09:33.200 --> 00:09:36.299 10 times a better salesperson than me? No, not 00:09:36.299 --> 00:09:38.159 at all. I just got lucky with the map. Exactly. 00:09:38.620 --> 00:09:41.740 The variability in our performance was due almost 00:09:41.740 --> 00:09:44.779 entirely to factors outside of our control. But 00:09:44.779 --> 00:09:47.080 the objective production method punishes me, 00:09:47.220 --> 00:09:50.659 and it rewards you. It creates this false sense 00:09:50.659 --> 00:09:53.580 of meritocracy. Plus, it completely ignores quality. 00:09:53.879 --> 00:09:56.340 If you measure a customer service rep purely 00:09:56.340 --> 00:09:58.580 on the number of calls taken, they're just going 00:09:58.580 --> 00:10:00.399 to start rushing people off the phone to hit 00:10:00.399 --> 00:10:02.480 their number. You get what you measure. It's 00:10:02.480 --> 00:10:04.480 a classic management problem. If you measure 00:10:04.480 --> 00:10:06.820 speed, you get speed, but you might lose customer 00:10:06.820 --> 00:10:10.019 loyalty in the process. So to try and fix that, 00:10:10.139 --> 00:10:12.460 some companies use the second method, the personnel 00:10:12.460 --> 00:10:14.659 method. This one sounds like the naughty list. 00:10:14.820 --> 00:10:16.779 It essentially is. This is recording what they 00:10:16.779 --> 00:10:20.279 call withdrawal behaviors. So absenteeism, being 00:10:20.279 --> 00:10:23.289 late, having accidents. Yeah. It's basically 00:10:23.289 --> 00:10:25.210 tracking the negatives. So you showed up on time 00:10:25.210 --> 00:10:26.909 every day. Congratulations, you're a good employee. 00:10:27.149 --> 00:10:29.009 Well, it sets a floor, but it doesn't really 00:10:29.009 --> 00:10:31.769 set a ceiling. You know, it's very useful for 00:10:31.769 --> 00:10:34.730 identifying problem employees, people who are 00:10:34.730 --> 00:10:38.029 literally not showing up to do the job. But it 00:10:38.029 --> 00:10:40.330 is completely useless for distinguishing between 00:10:40.330 --> 00:10:42.649 an average worker and a superstar. You can be 00:10:42.649 --> 00:10:45.470 punctual and still be totally incompetent. Exactly. 00:10:46.009 --> 00:10:48.870 Which brings us to the third method, the one 00:10:48.870 --> 00:10:51.049 that almost everyone in a knowledge economy is 00:10:51.049 --> 00:10:54.139 subjected to. judgmental evaluation. This is 00:10:54.139 --> 00:10:56.259 the subjective one. This is the subjective rating. 00:10:56.419 --> 00:10:58.700 This is where your manager sits down, thinks 00:10:58.700 --> 00:11:01.259 about your year and decides if you are a four 00:11:01.259 --> 00:11:04.860 or five on leadership. And this, this is where 00:11:04.860 --> 00:11:06.700 the wheels come off the whole thing. Yeah. The 00:11:06.700 --> 00:11:10.399 section on cognitive biases here is frankly terrifying 00:11:10.399 --> 00:11:12.980 because we like to think our bosses are these 00:11:12.980 --> 00:11:15.720 rational, objective judges. We do. But the science 00:11:15.720 --> 00:11:17.799 says they are bringing a lifetime of baggage 00:11:17.799 --> 00:11:20.330 into that meeting with them. It's not just baggage. 00:11:20.389 --> 00:11:23.389 It's like a funhouse mirror. Let's talk about 00:11:23.389 --> 00:11:26.649 the big one first. The idiosyncratic rater effect. 00:11:27.250 --> 00:11:30.049 This comes from some incredible research by Marcus 00:11:30.049 --> 00:11:32.629 Buckingham and Ashley Goodall. I love this concept. 00:11:32.789 --> 00:11:35.309 Break it down for us. Okay. So they argue that 00:11:35.309 --> 00:11:38.350 when a manager rates you on something, the rating 00:11:38.350 --> 00:11:41.169 reveals more about the rater than it does about 00:11:41.169 --> 00:11:43.090 the person being rated. Wait, hold on. So if 00:11:43.090 --> 00:11:46.190 my boss says I'm bad at strategic thinking, that's 00:11:46.190 --> 00:11:48.330 actually about their idea of strategic thinking. 00:11:48.669 --> 00:11:50.649 It's about their definition of the term. Yes. 00:11:50.809 --> 00:11:53.929 And it's about their mood that day, their general 00:11:53.929 --> 00:11:56.529 tendency to be a hard grader or an easy grader. 00:11:56.649 --> 00:11:59.690 The data suggests that over half, sometimes as 00:11:59.690 --> 00:12:02.090 much as 60 percent of the variance in a rating 00:12:02.090 --> 00:12:03.809 has nothing to do with the employee's actual 00:12:03.809 --> 00:12:06.929 performance. Over half. Over half. It's all about 00:12:06.929 --> 00:12:09.610 the manager's unique rating tendencies, their 00:12:09.610 --> 00:12:12.230 own personal biases. I'm being measured with 00:12:12.230 --> 00:12:14.149 a ruler that changes length depending on who 00:12:14.149 --> 00:12:16.659 is holding it. That is a perfect analogy. That's 00:12:16.659 --> 00:12:18.580 exactly what's happening. If I'm a pessimist, 00:12:18.659 --> 00:12:21.039 I might never give a five out of five because, 00:12:21.100 --> 00:12:23.940 you know, nobody is perfect. If you're a people 00:12:23.940 --> 00:12:26.220 pleaser, you might give everyone fives just to 00:12:26.220 --> 00:12:29.000 avoid having a difficult conversation. We aren't 00:12:29.000 --> 00:12:31.240 measuring the work. We are measuring the manager's 00:12:31.240 --> 00:12:33.460 personality. And then you layer the halo effect 00:12:33.460 --> 00:12:36.100 on top of that. Oh, the halo effect is a classic. 00:12:36.139 --> 00:12:39.059 It's a cognitive shortcut. The human brain is, 00:12:39.139 --> 00:12:41.740 you know, it's lazy. It doesn't want to evaluate 00:12:41.740 --> 00:12:44.850 20 different traits separately. So if I have 00:12:44.850 --> 00:12:47.269 a generally positive impression of you, maybe 00:12:47.269 --> 00:12:50.830 you're tall, confident, you dress well, my brain 00:12:50.830 --> 00:12:53.169 creates a halo. I just assume that because you 00:12:53.169 --> 00:12:55.370 are good at one thing, like looking professional, 00:12:55.649 --> 00:12:57.929 you must be good at everything. The financial 00:12:57.929 --> 00:13:01.029 analysis, team management, coding, all of it. 00:13:01.049 --> 00:13:02.929 All of it gets a little bump. It's the pretty 00:13:02.929 --> 00:13:05.389 privilege of the corporate world in a way. And 00:13:05.389 --> 00:13:07.990 the reverse is the horns effect, I assume. Exactly. 00:13:07.990 --> 00:13:11.090 If you have a messy desk or you made one awkward 00:13:11.090 --> 00:13:13.950 comment in a meeting six months ago, I might 00:13:13.950 --> 00:13:16.409 subconsciously downgrade your entire year's output. 00:13:16.809 --> 00:13:19.289 He has a messy desk, so his thinking must be 00:13:19.289 --> 00:13:22.549 disorganized. It's a leap, but our brains make 00:13:22.549 --> 00:13:24.129 it all the time. And there's the anchoring effect. 00:13:24.269 --> 00:13:26.009 This is about first impressions. Or just the 00:13:26.009 --> 00:13:28.610 first piece of data available. Let's say I look 00:13:28.610 --> 00:13:30.730 at your file before our review and I see that 00:13:30.730 --> 00:13:33.779 your previous boss rated you a three. Okay. My 00:13:33.779 --> 00:13:36.500 brain is now anchored to that number. Even if 00:13:36.500 --> 00:13:39.340 you performed like a five this year, it is psychologically 00:13:39.340 --> 00:13:42.379 very, very difficult for me to deviate far from 00:13:42.379 --> 00:13:44.460 that anchor. I will now look for evidence to 00:13:44.460 --> 00:13:47.440 confirm the three, not to challenge it. So we 00:13:47.440 --> 00:13:49.879 are fighting our boss's personality, their general 00:13:49.879 --> 00:13:52.580 vibe check of us, and whatever number they happen 00:13:52.580 --> 00:13:55.100 to see first. And we can't forget leniency and 00:13:55.100 --> 00:13:58.000 centrality bias. These are huge. Leniency is 00:13:58.000 --> 00:14:00.679 simply rating everyone high to avoid the awkward 00:14:00.679 --> 00:14:03.720 conversation. No manager wants to have that fight. 00:14:03.919 --> 00:14:07.740 And centrality is rating everyone average or 00:14:07.740 --> 00:14:10.820 in the middle to play it safe. Everyone meets 00:14:10.820 --> 00:14:14.159 expectations. Both of them make the data completely 00:14:14.159 --> 00:14:16.700 useless for actually differentiating performance. 00:14:17.179 --> 00:14:19.340 There was one more concept in this section that 00:14:19.340 --> 00:14:21.740 I found really interesting. What? Organizational 00:14:21.740 --> 00:14:27.220 citizenship behavior, or OCB. Ah, yes. The good 00:14:27.220 --> 00:14:29.559 soldier syndrome. This is all the stuff that 00:14:29.559 --> 00:14:31.220 isn't actually in your job description, right? 00:14:31.360 --> 00:14:33.399 Correct. It's helping a colleague when their 00:14:33.399 --> 00:14:36.000 printer jams. It's planning the office holiday 00:14:36.000 --> 00:14:38.179 party. It's staying late to wait for a delivery 00:14:38.179 --> 00:14:40.000 that isn't even for your department. The little 00:14:40.000 --> 00:14:42.340 things. The little things. The research shows 00:14:42.340 --> 00:14:44.419 that managers are heavily, heavily influenced 00:14:44.419 --> 00:14:47.460 by these dimensions of altruism and civic virtue. 00:14:47.700 --> 00:14:49.740 But isn't that a good thing? I mean, we want 00:14:49.740 --> 00:14:52.019 people to be nice and helpful. It's good for 00:14:52.019 --> 00:14:55.549 the culture, definitely. but it biases the performance 00:14:55.549 --> 00:14:57.529 review against people who just want to do their 00:14:57.529 --> 00:14:59.909 job and go home. Right. If you're an introvert 00:14:59.909 --> 00:15:02.210 who hits every single sales target, but you don't 00:15:02.210 --> 00:15:04.389 want to organize the potluck, your rating might 00:15:04.389 --> 00:15:06.250 actually be lower than the extrovert who sells 00:15:06.250 --> 00:15:08.789 less but always brings donuts on Friday. So we 00:15:08.789 --> 00:15:11.950 are rewarding personality and compliance, not 00:15:11.950 --> 00:15:14.590 just pure output. We're rewarding the good citizen, 00:15:14.769 --> 00:15:17.710 which is not always the same as the good worker. 00:15:18.320 --> 00:15:20.320 Okay, so we know that data is just a complete 00:15:20.320 --> 00:15:23.179 mess. But let's look at the tools themselves, 00:15:23.500 --> 00:15:26.620 the actual forms and scales. Because the sources 00:15:26.620 --> 00:15:29.139 list a whole menu of methods, and some of them 00:15:29.139 --> 00:15:31.759 sound, well, frankly, they sound like a reality 00:15:31.759 --> 00:15:34.059 TV show elimination. They range from the simple 00:15:34.059 --> 00:15:36.740 to the downright brutal. The most common, by 00:15:36.740 --> 00:15:39.360 far, is the graphic rating scale. This is the 00:15:39.360 --> 00:15:41.500 standard rate from one to five checklist, right? 00:15:41.580 --> 00:15:44.200 Yes. You list a bunch of traits, dependability, 00:15:44.519 --> 00:15:47.039 quality of work, initiative. And you rate them 00:15:47.039 --> 00:15:49.639 on a scale. It could be 5 points, 7 points. I've 00:15:49.639 --> 00:15:52.000 even seen 20 -point scales. On a scale of 1 to 00:15:52.000 --> 00:15:54.299 20, how much initiative do you have? That seems 00:15:54.299 --> 00:15:56.320 impossibly granular. Who knows the difference 00:15:56.320 --> 00:15:59.460 between a 14 and a 15? Nobody. It creates this 00:15:59.460 --> 00:16:02.399 false sense of precision. A 14 and a 15 are basically 00:16:02.399 --> 00:16:04.580 the same thing, but the number makes it look 00:16:04.580 --> 00:16:07.139 all scientific. But the bigger controversy comes 00:16:07.139 --> 00:16:09.759 with the employee comparison methods. Okay. This 00:16:09.759 --> 00:16:11.480 is where we stop comparing you to a standard 00:16:11.480 --> 00:16:13.379 and we start comparing you to each other. Ah, 00:16:13.580 --> 00:16:16.580 the Hunger Games approach. In a way, yeah. There's 00:16:16.580 --> 00:16:19.139 the rank order method. Yeah. That is literally 00:16:19.139 --> 00:16:22.419 listing your subordinates from best to worst. 00:16:22.600 --> 00:16:25.259 Number one, number two, all the way down. But 00:16:25.259 --> 00:16:28.820 what if I have a team of 10 superstars? Someone 00:16:28.820 --> 00:16:31.860 still has to be ranked worst, even if they're 00:16:31.860 --> 00:16:35.700 amazing. That is the fundamental flaw. You could 00:16:35.700 --> 00:16:37.820 have the best team in the entire company, but 00:16:37.820 --> 00:16:40.159 on paper, half of them look below average just 00:16:40.159 --> 00:16:42.220 by definition of the ranking. That's terrible. 00:16:42.500 --> 00:16:44.440 Then there's paired comparison. You take two 00:16:44.440 --> 00:16:46.440 employees, you pick the best, then you take another 00:16:46.440 --> 00:16:48.299 two, pick the best. You repeat this over and 00:16:48.299 --> 00:16:51.220 over until a full hierarchy forms. It sounds 00:16:51.220 --> 00:16:52.980 like a tournament bracket for people's careers. 00:16:53.259 --> 00:16:55.820 It is. But the most controversial version of 00:16:55.820 --> 00:16:58.419 this whole idea is forced distribution, which 00:16:58.419 --> 00:17:00.860 is often linked to a practice called top greeting. 00:17:01.100 --> 00:17:03.059 Is this the bell curve thing I've heard about? 00:17:03.200 --> 00:17:06.289 Yes. The force fit method. Managers are required 00:17:06.289 --> 00:17:08.710 to place their employees into a normal distribution. 00:17:09.150 --> 00:17:12.609 You must have, say, a top 20 percent, a middle 00:17:12.609 --> 00:17:15.190 70 percent and a bottom 10 percent. You must. 00:17:15.369 --> 00:17:17.970 Even if it's not true. Even if it's not true. 00:17:18.210 --> 00:17:20.430 And top grading is where they take that bottom 00:17:20.430 --> 00:17:23.650 percentage, the bottom 10 percent, and they fire 00:17:23.650 --> 00:17:27.859 them. fire them. Often, yes. It was made famous 00:17:27.859 --> 00:17:31.480 by Jack Welch at General Electric. The idea is 00:17:31.480 --> 00:17:33.940 to constantly prune the workforce to raise the 00:17:33.940 --> 00:17:36.099 average level of talent. But doesn't that just 00:17:36.099 --> 00:17:39.140 destroy teamwork? I mean, if I know that for 00:17:39.140 --> 00:17:41.019 me to stay safe, you have to be in the bottom 00:17:41.019 --> 00:17:43.440 10 percent. I am not going to help you with your 00:17:43.440 --> 00:17:45.920 project. That is the primary criticism. It creates 00:17:45.920 --> 00:17:48.559 a cutthroat, toxic environment. It turns teammates 00:17:48.559 --> 00:17:51.539 into rivals. I can't imagine working like that. 00:17:51.720 --> 00:17:54.359 Microsoft used a version of this for years and 00:17:54.359 --> 00:17:56.619 famously abandoned it because employees said 00:17:56.619 --> 00:17:59.279 it created a culture of cannibalism. People were 00:17:59.279 --> 00:18:02.119 actively sabotaging each other. It also assumes 00:18:02.119 --> 00:18:04.339 that performance follows a bell curve, which 00:18:04.339 --> 00:18:06.960 in many high performance jobs, it just doesn't. 00:18:06.980 --> 00:18:09.140 You might genuinely have a team of all stars. 00:18:09.380 --> 00:18:11.640 So are there any methods that try to be less... 00:18:12.009 --> 00:18:13.809 I don't know. Arbitrary. Well, there's a movement 00:18:13.809 --> 00:18:16.450 toward behavioral methods, things like bars. 00:18:16.650 --> 00:18:20.210 That's behaviorally anchored rating scales. Bars. 00:18:20.390 --> 00:18:23.990 I love these HR acronyms. What is bars? So instead 00:18:23.990 --> 00:18:26.430 of just a generic one to five for a word like 00:18:26.430 --> 00:18:29.329 communication, the points on the scale are described 00:18:29.329 --> 00:18:32.869 by specific narrative examples of behavior. OK, 00:18:32.930 --> 00:18:36.009 give me an example. So a five might say. consistently 00:18:36.009 --> 00:18:38.690 anticipates communication breakdowns and proactively 00:18:38.690 --> 00:18:41.589 addresses them with the team while a one might 00:18:41.589 --> 00:18:44.670 say frequently fails to return important emails 00:18:44.670 --> 00:18:47.750 within 48 hours oh that seems much more helpful 00:18:47.750 --> 00:18:49.670 it's concrete it gives you a specific target 00:18:49.670 --> 00:18:52.349 to aim for it is better much better but it's 00:18:52.349 --> 00:18:55.230 incredibly expensive and time consuming to develop 00:18:55.230 --> 00:18:57.829 because you need to create these specific examples 00:18:57.829 --> 00:19:01.130 for every single job role in the company of course 00:19:01.130 --> 00:19:04.309 there's also bos behavioral observation scale 00:19:04.809 --> 00:19:07.049 which is similar, but it rates the frequency 00:19:07.049 --> 00:19:09.609 of these critical incidents. How often does the 00:19:09.609 --> 00:19:12.430 employee do X? And we can't ignore the role of 00:19:12.430 --> 00:19:14.369 technology here. The source mentions electronic 00:19:14.369 --> 00:19:16.690 performance monitoring. This is the new frontier, 00:19:16.849 --> 00:19:19.589 isn't it? It really is. This is the modern evolution 00:19:19.589 --> 00:19:22.170 of all of this. Computers recording huge amounts 00:19:22.170 --> 00:19:24.210 of data on multiple dimensions of your work, 00:19:24.349 --> 00:19:27.990 keystrokes, email sent, time on calls. The benefit 00:19:27.990 --> 00:19:30.799 being... that it potentially removes human bias 00:19:30.799 --> 00:19:32.500 because the computer doesn't care if you dress 00:19:32.500 --> 00:19:34.740 well or bring donuts. It doesn't care at all. 00:19:34.759 --> 00:19:37.160 It just cares about the data. But it introduces 00:19:37.160 --> 00:19:40.759 this massive new problem of surveillance. Right. 00:19:40.799 --> 00:19:43.500 The big brother aspect. It's a huge concern and 00:19:43.500 --> 00:19:46.299 it can measure activity. Are you typing? But 00:19:46.299 --> 00:19:48.920 it really struggles to measure thought. If my 00:19:48.920 --> 00:19:51.900 job is to solve complex problems, I might be 00:19:51.900 --> 00:19:54.420 most productive when I'm staring out the window 00:19:54.420 --> 00:19:57.279 thinking. But the computer just sees that as. 00:19:57.880 --> 00:20:00.559 idle time. So you're being judged by an algorithm 00:20:00.559 --> 00:20:02.819 that can't understand the most valuable part 00:20:02.819 --> 00:20:04.799 of your job. That's a whole other level of anxiety. 00:20:05.059 --> 00:20:07.980 It is a totally new one. So somewhere along the 00:20:07.980 --> 00:20:10.940 line, companies realize that maybe one boss's 00:20:10.940 --> 00:20:13.140 opinion isn't enough, especially with all these 00:20:13.140 --> 00:20:16.960 biases. So enter the 360 degree feedback. Right. 00:20:17.390 --> 00:20:19.250 The concept here is triangulation. Let's get 00:20:19.250 --> 00:20:21.609 more data points. You get evaluations not just 00:20:21.609 --> 00:20:23.589 from your supervisor, but also from your peers, 00:20:23.690 --> 00:20:25.670 your subordinates, and sometimes even your customers. 00:20:25.970 --> 00:20:28.849 The idea being that if everyone, your boss, your 00:20:28.849 --> 00:20:31.230 team, the people who work for you, if they all 00:20:31.230 --> 00:20:33.509 say you're difficult to work with, it's probably 00:20:33.509 --> 00:20:37.730 true. Theoretically, yes. It should reduce bias 00:20:37.730 --> 00:20:39.869 by providing a more comprehensive view of the 00:20:39.869 --> 00:20:42.890 person. But here is where the research throws 00:20:42.890 --> 00:20:46.140 a massive curveball. Okay. Studies show that 00:20:46.140 --> 00:20:48.819 the source of the feedback doesn't actually change 00:20:48.819 --> 00:20:51.480 how the employee behaves afterward. Wait, really? 00:20:51.599 --> 00:20:54.339 So if my boss tells me I need to improve my communication 00:20:54.339 --> 00:20:58.039 or my peer tells me the exact same thing, I react 00:20:58.039 --> 00:20:59.740 the same way. It doesn't matter who says it. 00:20:59.859 --> 00:21:02.119 In terms of influencing your subsequent behavior, 00:21:02.440 --> 00:21:04.480 the research says, no, it doesn't really matter. 00:21:04.579 --> 00:21:07.859 And there are huge risks with peer ratings. Like 00:21:07.859 --> 00:21:10.740 what? Well, peers might be biased by friendships. 00:21:10.819 --> 00:21:13.440 They'll rate their friends higher. Or they might 00:21:13.440 --> 00:21:16.609 fear retaliation. If I rate you low, you're going 00:21:16.609 --> 00:21:18.789 to rate me low next time. It's mutually assured 00:21:18.789 --> 00:21:22.089 destruction. Or mutually assured inflation. Everyone 00:21:22.089 --> 00:21:24.990 just rates everyone high to keep the peace and 00:21:24.990 --> 00:21:27.789 the data becomes meaningless again. There's also 00:21:27.789 --> 00:21:30.630 the anxiety of impression management. You end 00:21:30.630 --> 00:21:32.329 up spending more time trying to look good to 00:21:32.329 --> 00:21:34.349 your peers. You're basically campaigning for 00:21:34.349 --> 00:21:37.190 votes than actually doing your work. And what 00:21:37.190 --> 00:21:39.809 about asking the employee to rate themselves? 00:21:40.549 --> 00:21:43.380 The self -assessment. This is a really tricky 00:21:43.380 --> 00:21:45.740 one. The sources warn about something they call 00:21:45.740 --> 00:21:49.359 flawed self -assessment. Which is a very polite 00:21:49.359 --> 00:21:51.160 way of saying we were all delusional about our 00:21:51.160 --> 00:21:53.650 own ability. Oh, yeah, basically. Psychological 00:21:53.650 --> 00:21:56.549 research, like the Dunning -Kruger effect, shows 00:21:56.549 --> 00:21:59.250 that self -perception often fails to align with 00:21:59.250 --> 00:22:02.490 reality. Low performers often dramatically overestimate 00:22:02.490 --> 00:22:04.869 their abilities because they lack the very competence 00:22:04.869 --> 00:22:07.450 needed to realize they're incompetent, while 00:22:07.450 --> 00:22:09.029 high performers might actually underestimate 00:22:09.029 --> 00:22:11.369 themselves. So the people who need the most help 00:22:11.369 --> 00:22:13.009 are the ones who think they're doing great. Often, 00:22:13.049 --> 00:22:16.210 yes. It's a real paradox. However, despite that 00:22:16.210 --> 00:22:19.069 inaccuracy, the source notes a very, very important 00:22:19.069 --> 00:22:22.500 benefit of self -assessment. Participation. When 00:22:22.500 --> 00:22:25.099 employees get to participate in the process by 00:22:25.099 --> 00:22:27.460 rating themselves, their satisfaction with the 00:22:27.460 --> 00:22:29.880 whole system increases dramatically. They feel 00:22:29.880 --> 00:22:32.599 like they've been heard. So even if I'm completely 00:22:32.599 --> 00:22:35.299 wrong about how great I am, just the act of letting 00:22:35.299 --> 00:22:38.420 me say it makes me hate the process less. It's 00:22:38.420 --> 00:22:40.880 about procedural justice, feeling like the process 00:22:40.880 --> 00:22:43.740 was fair. Exactly. It gives you a voice, and 00:22:43.740 --> 00:22:46.190 that's powerful. So we've got all these different 00:22:46.190 --> 00:22:48.509 methods, all these deep -seated biases. This 00:22:48.509 --> 00:22:51.049 leads to this massive conflict in the world of 00:22:51.049 --> 00:22:53.730 work that the sources call the scientist practitioner 00:22:53.730 --> 00:22:56.990 gap. This is a major theme, yeah. You have this 00:22:56.990 --> 00:22:59.970 huge disconnect. On one side, you have academia, 00:23:00.269 --> 00:23:03.049 the scientists, saying appraisals work, they 00:23:03.049 --> 00:23:05.910 improve effectiveness, they identify competencies. 00:23:06.069 --> 00:23:07.829 And on the other side, you have literally everyone 00:23:07.829 --> 00:23:10.410 else saying, please, for the love of God, make 00:23:10.410 --> 00:23:13.480 it stop. Right. Practitioners and employees perceive 00:23:13.480 --> 00:23:16.420 them as uncomfortable, stressful, and a source 00:23:16.420 --> 00:23:19.279 of deep distrust. And if the person being appraised 00:23:19.279 --> 00:23:21.819 doesn't trust the employer or the process, the 00:23:21.819 --> 00:23:24.480 entire exercise is completely pointless. And 00:23:24.480 --> 00:23:26.420 there's some pretty famous management theorists 00:23:26.420 --> 00:23:28.339 who actually agree with the employees on this, 00:23:28.420 --> 00:23:31.039 right? I see a mention of total quality management 00:23:31.039 --> 00:23:35.539 or TQM. Yes. W. Edwards Deming, who is basically 00:23:35.539 --> 00:23:38.740 the father of the entire quality movement, was 00:23:38.740 --> 00:23:41.180 a huge critic of individual performance reviews. 00:23:41.460 --> 00:23:43.819 He hated them. And his argument was what? The 00:23:43.819 --> 00:23:46.779 argument in TQM is that the system dictates performance 00:23:46.779 --> 00:23:50.059 far more than the individual does. Can you explain 00:23:50.059 --> 00:23:52.500 that a bit more? Sure. Imagine you work on an 00:23:52.500 --> 00:23:55.200 assembly line. If the machine you use is slow 00:23:55.200 --> 00:23:57.740 or the parts you're given are defective, you 00:23:57.740 --> 00:24:00.200 simply cannot be a high performer no matter how 00:24:00.200 --> 00:24:02.680 hard you try. My performance is capped by the 00:24:02.680 --> 00:24:05.859 system I'm in. Exactly. Deming argued that blaming 00:24:05.859 --> 00:24:08.240 the individual for what are actually seldom flaws 00:24:08.240 --> 00:24:11.339 is not only unfair, it's counterproductive. You 00:24:11.339 --> 00:24:13.700 should be fixing the system, not yelling at the 00:24:13.700 --> 00:24:15.799 worker. That makes so much sense. If the process 00:24:15.799 --> 00:24:18.480 is broken, reviewing the person is just cruel. 00:24:19.440 --> 00:24:22.259 Precisely. And this kind of skepticism is what's 00:24:22.259 --> 00:24:25.220 driving this big new trend. We are seeing a real 00:24:25.220 --> 00:24:27.579 shift away from the traditional dreaded annual 00:24:27.579 --> 00:24:29.559 review. Thank goodness for that. The biggest 00:24:29.559 --> 00:24:31.940 criticism is that annual feedback is just too 00:24:31.940 --> 00:24:34.579 infrequent to be useful. If you screw something 00:24:34.579 --> 00:24:36.740 up in January and you don't hear about it until 00:24:36.740 --> 00:24:39.319 December, how are you supposed to fix it? You 00:24:39.319 --> 00:24:41.140 can't. You've probably been making that same 00:24:41.140 --> 00:24:43.480 mistake for 11 months without knowing. Exactly. 00:24:43.799 --> 00:24:46.119 So companies are moving toward what they call 00:24:46.119 --> 00:24:48.640 continuous performance management. Which means 00:24:48.640 --> 00:24:51.259 what exactly? It means more frequent informal 00:24:51.259 --> 00:24:55.420 check -ins, weekly or monthly. It's less about 00:24:55.420 --> 00:24:58.079 giving someone a grade and more about coaching 00:24:58.079 --> 00:25:01.160 and removing roadblocks. The source says about 00:25:01.160 --> 00:25:03.619 a third of U .S. private companies are switching 00:25:03.619 --> 00:25:06.799 to this kind of model. Yes, the trend is significant. 00:25:06.900 --> 00:25:09.859 The focus is shifting to talent development and 00:25:09.859 --> 00:25:13.079 business agility. It's really hard to be agile 00:25:13.079 --> 00:25:16.019 as a company if you only set goals and give feedback 00:25:16.019 --> 00:25:18.869 once a year. Now, we absolutely have to talk 00:25:18.869 --> 00:25:20.970 about the legal side of this because we said 00:25:20.970 --> 00:25:22.930 it at the start. At the end of the day, these 00:25:22.930 --> 00:25:25.750 are legal documents. They absolutely are. And 00:25:25.750 --> 00:25:27.750 this is where companies can get into a lot of 00:25:27.750 --> 00:25:30.250 trouble. If a company fires someone and that 00:25:30.250 --> 00:25:33.190 person sues for wrongful termination or discrimination, 00:25:33.589 --> 00:25:36.069 the first thing their lawyer will ask for is 00:25:36.069 --> 00:25:38.170 their performance appraisals. Their exhibit A. 00:25:38.230 --> 00:25:43.240 So if I fire Bob for... Poor performance. But 00:25:43.240 --> 00:25:45.420 I gave him five stars on his review three months 00:25:45.420 --> 00:25:47.400 ago because I wanted to avoid an awkward conversation. 00:25:47.740 --> 00:25:50.599 You have a massive legal problem. The inconsistency 00:25:50.599 --> 00:25:53.319 is fatal in court. The jury will see that and 00:25:53.319 --> 00:25:55.759 immediately distrust the company. And the sources 00:25:55.759 --> 00:25:58.779 mention specific laws this relates to, like Title 00:25:58.779 --> 00:26:01.259 VII of the Civil Rights Act, the ADEA, which 00:26:01.259 --> 00:26:04.000 is age discrimination, and the ADA for disabilities. 00:26:04.519 --> 00:26:06.400 So how does a company protect itself legally? 00:26:06.839 --> 00:26:09.740 The defense is objectivity, or at least the appearance 00:26:09.740 --> 00:26:12.519 of it. To be legally sound, appraisals have to 00:26:12.519 --> 00:26:14.319 be job related. They must be behavior based. 00:26:14.480 --> 00:26:16.720 And the criteria have to be within the control 00:26:16.720 --> 00:26:19.140 of the employee. So no vague statements? No. 00:26:19.220 --> 00:26:21.599 They need to rely on specific functions, not 00:26:21.599 --> 00:26:24.720 global assessments like Bob is a bad fit or he 00:26:24.720 --> 00:26:27.359 has a bad attitude. That will not hold up. It 00:26:27.359 --> 00:26:29.559 sounds like those behaviorally anchored scales, 00:26:29.859 --> 00:26:32.319 the bars we talked about earlier, would be much 00:26:32.319 --> 00:26:34.579 legally safer. Much, much safer. They provide 00:26:34.579 --> 00:26:38.299 specific documented evidence of behavior, not 00:26:38.299 --> 00:26:41.259 just a manager's opinion. Now, here's where it 00:26:41.259 --> 00:26:43.240 gets really, really interesting for anyone listening 00:26:43.240 --> 00:26:45.920 who works in a global company. The culture clash. 00:26:46.359 --> 00:26:49.319 The source says you can't just copy paste a U 00:26:49.319 --> 00:26:52.160 .S. style review system into another country 00:26:52.160 --> 00:26:54.900 and expect it to work. This is a fascinating 00:26:54.900 --> 00:26:58.000 area of cross -cultural management. Performance 00:26:58.000 --> 00:27:00.759 appraisals as we know them are deeply, deeply 00:27:00.759 --> 00:27:04.140 rooted in Western and particularly American cultural 00:27:04.140 --> 00:27:07.579 norms. Norms like assertiveness and individualism. 00:27:07.599 --> 00:27:10.490 I'm number one. Correct. In what we call assertive 00:27:10.490 --> 00:27:13.450 cultures like the U .S. or Germany, we value 00:27:13.450 --> 00:27:15.630 personal accomplishment. We see the appraisal 00:27:15.630 --> 00:27:17.750 as a way to ensure equity. If I work harder, 00:27:17.809 --> 00:27:20.230 I should get paid more. Feedback is seen as a 00:27:20.230 --> 00:27:22.490 gift, a tool for self -improvement. Yeah. Tell 00:27:22.490 --> 00:27:24.349 me what I did wrong so I can fix it and get that 00:27:24.349 --> 00:27:27.029 promotion. That's the mindset. Exactly. But now 00:27:27.029 --> 00:27:30.029 let's look at cooperative or low assertiveness 00:27:30.029 --> 00:27:33.250 cultures. These are cultures that value interpersonal 00:27:33.250 --> 00:27:36.349 connection and group harmony far more than individual 00:27:36.349 --> 00:27:39.109 standing. So many parts of Asia or Latin America, 00:27:39.250 --> 00:27:42.990 for example. Yes, often. In these contexts, the 00:27:42.990 --> 00:27:46.609 very act of differentiating employees, of ranking 00:27:46.609 --> 00:27:49.529 one as best and another as worst, is seen as 00:27:49.529 --> 00:27:51.970 disrupting group harmony. It's considered offensive. 00:27:52.369 --> 00:27:54.670 It breaks the team cohesion. It's actively damaging. 00:27:54.930 --> 00:27:57.150 And direct critical feedback in these cultures 00:27:57.150 --> 00:27:59.529 is often viewed as threatening and obtrusive. 00:28:00.039 --> 00:28:02.319 It's not seen as coaching to help you. It's seen 00:28:02.319 --> 00:28:04.940 as a personal attack or a way to make you lose 00:28:04.940 --> 00:28:07.980 face in front of your peers. Wow. So if an American 00:28:07.980 --> 00:28:10.940 manager goes into a subsidiary in, say, Japan 00:28:10.940 --> 00:28:13.160 and tries to implement a forced distribution 00:28:13.160 --> 00:28:16.359 system. It will almost certainly fail spectacularly. 00:28:16.380 --> 00:28:18.720 It creates social separation in a culture that 00:28:18.720 --> 00:28:21.259 values social integration above all else. The 00:28:21.259 --> 00:28:23.740 implication is crystal clear. Management practices 00:28:23.740 --> 00:28:26.680 are not universal truths. They are cultural artifacts. 00:28:27.019 --> 00:28:29.549 That is a huge takeaway. We assume business is 00:28:29.549 --> 00:28:31.750 business everywhere, but business is run by people, 00:28:31.849 --> 00:28:34.549 and people are products of their culture. Precisely. 00:28:34.549 --> 00:28:38.390 You can't separate the two. So, wow. We've been 00:28:38.390 --> 00:28:41.130 on quite a journey here. We started with a universal 00:28:41.130 --> 00:28:43.849 dread of the annual review. That knot in your 00:28:43.849 --> 00:28:46.920 stomach. Exactly. We looked at the noble objectives, 00:28:47.240 --> 00:28:48.940 trying to solve that principal agent problem, 00:28:49.200 --> 00:28:52.079 trying to align the workforce. And then we uncovered 00:28:52.079 --> 00:28:54.779 the messy, messy reality of data collection, 00:28:55.039 --> 00:28:58.079 how counting widgets misses the point. But judging 00:28:58.079 --> 00:29:00.880 people is just riddled with biases. Things like 00:29:00.880 --> 00:29:03.900 the halo effect and my favorite, the idiosyncratic 00:29:03.900 --> 00:29:06.299 rater effect, which means it's not even about 00:29:06.299 --> 00:29:09.259 you. Right. And we looked at the tools from the 00:29:09.259 --> 00:29:11.980 simple one to five scales to the brutal force 00:29:11.980 --> 00:29:14.839 rankings. We saw how technology is changing. 00:29:14.990 --> 00:29:17.170 the game with electronic monitoring, and we touched 00:29:17.170 --> 00:29:19.710 on the massive legal risks and the cultural blind 00:29:19.710 --> 00:29:21.970 spots. It just highlights so perfectly that the 00:29:21.970 --> 00:29:23.869 performance appraisal is this textbook example 00:29:23.869 --> 00:29:26.329 of a system designed for pure logic that often 00:29:26.329 --> 00:29:28.349 fails completely because of human psychology. 00:29:28.690 --> 00:29:31.190 That's the core tension, isn't it? We try to 00:29:31.190 --> 00:29:33.710 turn messy human behavior into clean data points, 00:29:33.890 --> 00:29:36.250 but the instrument we use to measure it, the 00:29:36.250 --> 00:29:39.369 human manager, is inherently flawed and biased. 00:29:39.670 --> 00:29:42.329 It really is. As Buckingham and Goodall said, 00:29:42.799 --> 00:29:44.859 The rating tells us more about the raider than 00:29:44.859 --> 00:29:47.480 the ratee. Which makes you wonder about the future 00:29:47.480 --> 00:29:50.019 of all this. Where is it all going? It does. 00:29:50.200 --> 00:29:52.779 And that raises a pretty provocative thought 00:29:52.779 --> 00:29:55.480 to leave our listeners with. Let's hear it. We 00:29:55.480 --> 00:29:58.519 are seeing two completely diverging trends right 00:29:58.519 --> 00:30:01.500 now. On the one hand, you have this big push 00:30:01.500 --> 00:30:04.519 toward teamwork, total quality management, and 00:30:04.519 --> 00:30:07.670 psychological safety. A world which suggests 00:30:07.670 --> 00:30:09.710 the individual review should just go extinct. 00:30:09.890 --> 00:30:12.069 It's an outdated concept. Okay, that's one path. 00:30:12.210 --> 00:30:14.029 But on the other hand, we have the rapid rise 00:30:14.029 --> 00:30:16.890 of electronic monitoring and big data, which 00:30:16.890 --> 00:30:19.130 could make individual measurement more precise, 00:30:19.390 --> 00:30:22.170 more constant, and more invasive than ever before 00:30:22.170 --> 00:30:25.230 in human history. So the question is, will the 00:30:25.230 --> 00:30:28.269 performance review die out, a relic of a bygone 00:30:28.269 --> 00:30:31.809 era? Or will it just mutate into something new, 00:30:31.930 --> 00:30:33.829 something where the computer is the one giving 00:30:33.829 --> 00:30:36.869 us our score? Every single day. That is the question 00:30:36.869 --> 00:30:38.890 we're facing. Something to think about the next 00:30:38.890 --> 00:30:40.890 time that meeting invite pops up on your calendar. 00:30:41.210 --> 00:30:43.349 Thanks for listening to Deep Dive. We'll catch 00:30:43.349 --> 00:30:45.269 you on the next one. Goodbye, everyone.