What does Six Sigma have to do with your inbox, your slide decks, or your endless meetings? In this episode, Chris and Toby break down the meaning behind “Sigmas” - from factory floors to knowledge work - and ask a provocative question: how many errors per million does your team tolerate before calling it “good enough”? From polished PowerPoints to unpredictable processes, they explore why consistency in knowledge work is so rare, why novelty often hides defects, and why sometimes a little imperfection still moves the mission forward.
Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.
Transcript
From the show. Tap a time to play from there.
Hey, what's up, Toby? We've made another week. Nice. Hey, man. I know. It's processed out. We're going to jump right in. Let's do it. Today's topic comes from operations and manufacturing. And it's around the Six Sigma, or the Sigmas, or the number of failures within omnipotent records, which Six Sigma is what? Three parts per million? Is that a defect? Yeah. Well, it feels like all
we ever talk about, if we talk about sigmas in the workplace, we tolerate nothing less than six. That's all we talk about. And I think part of it is we refer to that kind of work around air reduction as Six Sigma work or lean work. Yeah, so. It's three parts per million. I think about manufacturing. That makes sense, right? If we're rolling tires or ball bearings off the line, we don't want to be digging through those.
I get that. Yep. But gosh, three errors per million transactions. So all right. So what one sigma. So one sigma is. 30 .9 % All right, so that's 690 ,000 defects per million So that's terrible That is a 30 % success rate, which is on par with probably most of the PowerPoints we've seen
Yeah, so 31 % defect rate which means only 61 % of them are defect free Yeah Did I say that right? Yeah, I know it's backwards. So like among a million, there's only going to be 310 ,000 that are defect, that are free of defect. That's one sigma. Right. Okay. So that's, you know,
so the, your good output, another way to say that your good output on one sigma is 30 .9%. Okay. It's brutal. Yeah. But I mean, again, for knowledge work, PowerPoints that I have seen. It's a good example though, right? I mean, yeah, like how many defects does a PowerPoint have? And that's, and that's a really good rule. What are the consequences of a spelling error in a PowerPoint or what are the consequences? It could
be an issue if it's an extra zero and you're talking about loss reduction or sales targets. Folks get antsy. If it's an indent, if it's a font error, things are the same, right? I would say that many defects are hidden by polished PowerPoints. Like what would McKinsey be without a polished PowerPoint? That's a frightening thought. Yeah. All right. Okay. So one of the things about
manufacturing is that to move up a sigma takes about the same effort, right? So going from one sigma to two sigma is basically going from 30 .9 % good to 69 .1 % good. So just moving it up, just one standard deviation basically. Which means that, you know, still 70%. This is probably
most I would say that this is probably most knowledge work. Yeah. If we think about that in the broadest sense, everything from communication to assorted manual tasks. Yeah. Yeah. Anything where a system's not guiding our action. Yep. Yeah. That feels pretty reasonable. Yeah. So if you think about one sigma, it's basically the process fails most of the time or it's highly inconsistent. Yeah. Right your PowerPoint presentation Template changes
multiple times a quarter, right? Like seconds, you know to two Sigma or you know, it's 69 % good output. Yeah still there or areas occur frequently Okay, now let's move that just up one level right and so the three Sigma It jumps from close to 70 % good to 93 .3 % good. Heart change. Yeah. It literally like, up to three
sigma, it's kind of almost like a hockey stick. Right? So again, it's, you know, from one sigma to two sigma, it's 39 % better. From two sigma to three sigma, it is 23 % better. And then the next jump from four sigma to five sigma, it's 93 % to 99 .38%. It's literally 7%. And then to five sigma, it jumps basically percentages
like 0 .4%. And these are literally the same amount of effort. Right. And yet I don't think when we think of business processes and knowledge work, we don't actually think about it as changing the process to improve the output. But it's about the consistency too. Yes. Right. And so when we do highly variable tasks, it's hard to get
a sense of our actual consistency. Correct. So the, did I deliver it on time is, is not really a Sigma ish kind of outcome because it depends on the task. So I think that's, you know, as we're talking about this, I'm thinking about having done, we've talked a lot about this, having done a lot of process improvement work years. You know, sometimes it's process improvement
at one desk or sometimes it's an entire team. But it feels like so often everybody is the big, is the organization gets larger, the work niches down and people have increasingly specialized. And so to be able to have, you know, to your point, the difference of shifting from high level of defects at one Sigma to a marked reduction at two Sigma, you know, in a large organization, you have to touch a lot of people. Yeah. Even
as we're talking about this, it makes me think, ah, this is why we start to see formal payroll systems being implemented or financial systems being upgraded and driving a lot of consistent behaviors just to reduce variability. Oh, no, totally true. And what's interesting is like, I think oftentimes in knowledge work, we don't think of it as a process of delivery. We think of it as a task that we need to go. more of like
doing the dishes or something. I don't even know how, I don't know, that was a terrible example. But we think of it as like, oh, my boss wants me to do the thing. Right. And we don't think of it, you know, we've talked about this quite a bit where if the thing is repeatable, turn it into a process. Totally. Yeah. But it's funny. I've been watching the World Series. Yeah. I'm going to deviate here from our topic here for a little bit, but Google Cloud has this commercial
on baseball players that hit the base with their bat and how that leads to hard hits or something like that. And in my mind, I'm like, that's novel. I appreciate that. And it is absolutely, there's no action that you can take. And I find it fascinating that like a lot of knowledge work comes down to novelty, which means that you can get, you
can basically hide all your defects in the corner because as long as it's novel and it follows some sort of, you know, the cognitive bias of the person with the highest salary in the room, then it's okay. Yeah. Processes don't have personalities to soften. message or build relationships the same way that people do. Yeah. Yeah. Yeah. But I just, I just think sometimes we don't, we just
don't think about it as something that even can be fixed. Yeah. And so, you know, and sometimes it's the person that needs to fix it as the manager. And if you're an IC, then it's not yours to fix. Well, Chris, it kind of makes me. Ask the question like is it even reasonable to talk about professional Interactions in this same concept of like air
free transactions You know manufacturing is a is a highly repeatable super linear process generally Yeah, not super linear. Then we invest a lot of time understanding all the permutations that it could go down Yeah, I think this is where it falls into like do your system support or do they complicate and Like it does your sis like Because I would say that there are some times in which a one sigma rate is okay Yeah,
and I would say that there are other times where you know, it needs to be in the in the highly accurate Highly consistent outcome, right? And I think sometimes those two get flopped Right, and I think I think some of it is about the novelty like and I I Trash on data a lot not because I don't like data but because data seems to be whatever you want it to be Yeah, and so I do think that when it comes to Consistent things
that we do we should aim to make them better You know and not aim to make them more novel It's interesting. So novelty is a is a source of defects and so that also makes me think of the failure rate that we have around innovation and that's Unfortunately come kind of this catch -all buzzword For I have a new approach that I'm testing to I'm just sloppy and I'm wildly inconsistent. So they're innovating, you know
well it is interesting because like your upstream and your downstream has to all match in a manufacturing, because you've got work in progress, you've got stuff moving through the line. Totally. Knowledge work oftentimes ends up being one person doing one thing. And the transferability of that knowledge isn't necessarily taken into consideration. I
saw a funny meme on someone doing math of 16 minus 9 as a McKinsey consultant and they're like, oh, you can obviously see that this is a sum of squares. And so if you turn that into three squared and four squared and then you multiply it all out and they go, they do four or five steps and they get to 16 minus nine and you're like, still seven. And there we are. And we made it to seven. No, it's true. Well, the transfer
of moving pieces and parts versus moving information, it's easier to know if all the pieces have been moved from point A to point B. Whereas it's hard to know if all the information has been translated from person A to person B. We are terrible at that as human beings. It's hard to know, isn't it? Because you don't check. Because we assume that everybody seems everybody thinks the exact same way that we do or doesn't action the same
way like if you ever demo anything Like in a in a scrum you will be like what in the world is that person doing? Don't they see the button to the left and then you're like, oh That was a semi clever way of doing it that I would have never thought of or figured it out Chris, I think that's so true. Shadowing somebody to watch how they navigate a familiar application is fascinating and painful. Yeah. Because you're right, you
get the same way. Yeah, I built a UI. I was really happy to build some stuff in there that stuff just never got used because it just was unusable. Yeah. Like for me, not doing the day -to -day work was what I thought would work. But that ended up being one of these knowledge deviations because I didn't know the work that was being done and so I didn't built it in the right way.
The reality is it's like, what's funny, the other funny thing is oftentimes it's not in the outcome in which it's. Like we become fascinated with being distraction. I kind of stopped there a little bit, but we're fascinated with the distraction and not the actual process or outcome. And so it's like, hey, if we, you know, did we sell the deal? Yeah, we did. Let's go celebrate. Did we sell the right deal? Did we sell it to the
right client? Does it align with what we're trying, where we're going? Like those are all really important. And I think... That's one of the deviation of saying yes to everything is a bad one. Well, Chris, I think as we're talking this through, the piece that I take away is that, A, being nervous when somebody comes to Six Sigma, your professional services approach, your college work, and the piece you just touched on really
resonates with me and it's that we have to step back and ask the bigger picture question. Where? Because sometimes a highly defective exchange can still move the mission forward. That can still achieve our outcomes. Even if the PowerPoint is error ridden and has, you know, all kinds of font mismatches. Yeah. So the other is you probably hear snoring in the back. That's my bulldog. Just just crashed out. Not helpful for a podcast, but having a grand old time just snoring
away. I thought that was your stomach. Better that it's a bulldog, sir. That's awesome. Well, we made it again, man. All right. Good stuff. All right. Catch you next time. All right. See you.
The Process Debt Podcast: Improving knowledge work processes to help you reach One Peace Flow. Stop curing cognitive burnout with physical inactivity. Discover how building a balanced ecosystem of off-screen hobbies can...
The Process Debt Podcast: Improving knowledge work processes to help you reach One Peace Flow. Chasing every new AI tool creates friction and burnout. Learn how to pinpoint your spot on the adoption curve and eliminate...
The Process Debt Podcast: Improving knowledge work processes to help you reach One Peace Flow. Overcome the costly last mile problem in knowledge work, streamline messy workflows, and unlock single-task focus with...
The Process Debt Podcast: Improving knowledge work processes to help you reach One Peace Flow. Stop losing productive hours to invisible friction, endless rework, and fuzzy handoffs. Here is how to audit hidden waste...
The Process Debt Podcast: Improving knowledge work processes to help you reach One Peace Flow. Discover how obsessing over vanity metrics creates hidden process debt, destroys deep focus, and how a four-step framework...