ANNOUNCER: Welcome back to the Augmented Podcast. Augmented is a podcast for industrial leaders, process engineers, and shop floor operators, hosted by futurist Trond Arne Undheim and presented by Tulip. TROND: Kimberly, welcome to the podcast. How are you? KIMBERLY: I'm good. Thanks for having me, Trond. TROND: Yeah. So, I thought we would talk about your...well, one of your favorite subjects, I don't know, design? KIMBERLY: Gosh, yeah. There is not so much I like more than design. [laughs] TROND: So, here's the situation. I think you have worked on industrial design for a while. You certainly have a degree in industrial design from Virginia Tech. And you have worked on these issues in industry for a while at Rocket Software and now at industrial tech company Tulip. You say I think on your LinkedIn, that you're driven by curiosity, complexity, and infinite ideas. Is that a way to clue me in on your view of design? It should be curious, complex, and infinite. KIMBERLY: Absolutely. I think kind of the power of design is being able to ask a lot of questions about really what's going on in a situation, and why something is happening, and how you can really make that better, and brainstorming ad infinitum on ways to make that possible. TROND: But there is a tension there, though, because, in industrial production and stuff, there's a lot of talk about quality, efficiency, consistency. So, there's a system that you have to relate to as well when you're doing industrial design. KIMBERLY: Exactly. I think one of the core tenets of good design is really understanding users, and prototyping with them, and being able to test, is this the right solution? So, when we design features at Tulip, we always talk to our customers about it and ask them to complete a few tasks to see. Are we thinking about this the way that they're thinking about it, and did that line up right? TROND: So, that's what exactly I'm going to do before we get there because I wanted to ask you more about your background. Before we start talking about the topic in and of itself, I just want you to tell me, have I understood you correctly? You are a designer. And how did you become one? What was it that drew you to this world? I don't know, in terms of education, but perhaps even before you got a degree in it. KIMBERLY: Yeah, let's see. When I was seven, found a piece of paper that says I wanted to be an inventor when I grow up. [laughs] And when I was starting out college, I thought that the best way to do that was starting out with mechanical engineering. Turns out thermodynamics is not really my strong suit, [laughs] not things I find fun. And transferred into industrial design to really think about how I could build physical products in the world and make things that fit people and use that power of invention to build stuff. And kind of fell into digital product design after college and really enjoyed focusing on user research. I had done a lot of coding growing up and found that I could apply that in digital product design. And when I found out about Tulip, it was really kind of the best of both worlds in bringing together how much I absolutely love about physical products with being able to make tech tools. TROND: So, there's a lot of stuff to unpack here. But designing for manufacturing is, in and of itself, challenging, right? Because there are these constraints that we just barely started talking about. But then digital design, in theory, is, you know, lots of opportunity, but I'm sure it adds some other constraints that are important to talk about. Why don't you kick us off a little bit on sort of designing for manufacturing first? And then we can move into what you think has been some of the real opportunities of digital both now and, you know, as you see the field kind of evolving. But design for manufacturability or DFM, right? There is this notion. There's a paradigm in engineering now that design is important, which perhaps it wasn't always, right? KIMBERLY: It took a little time to catch on [laughs] for a lot of folks. TROND: Yeah. So, what's happened there? And, you know, what does all that really mean to someone who's trying to be a designer, or, you know, indeed, firms that are considering what this is? Is it a trend? Is this something they need to catch on to? What is design in manufacturing, you know, in your mind? KIMBERLY: Yeah, I think designing and, we'll say, designing tools for manufacturing rather than, like, designing physical products for manufacturing processes is a whole separate issue. But it really involves thinking about not just someone's immediate environment but thinking about what's going on around them as well. So, one of the things that we think about a lot in digital tools for manufacturing is what is the operator actually doing. How big do the buttons that we use on the Tulip Player need to be to make sure that someone can touch them with giant gloves on or see them if they're in the side of their peripheral vision while they're focusing on another task at hand? And really making sure that that's as optimal and clear and obvious as possible, without at the same time drawing them away from that critical task that they're working on. TROND: You know, this is so interesting to me because if I think about the early days of not industrial automation but of kind of, like, the computer age, some of these software tools, it's not that they weren't well designed. I'm sure they also were. But they were trying to draw us into the computer. You seem to be saying, "I'm trying to do the opposite." You're trying to make the tool almost kind of invisible, in a sense, that it certainly doesn't distract you from what you actually were trying to do. KIMBERLY: Exactly. And I think that's the whole notion of user experience. Design is understanding that total workflow of what someone's working on and really looking at how to match tools to the real process that happens. Because I think one of the huge issues that a lot of manufacturers face today is, like, how much training it takes to put people through. Someone was telling me about a situation where they had to train every new operator because they had machines that they had to check on every hour and look for certain attributes in those. And they had to tell each operator where to find each of 150 different attributes across, like, 17 different screens, and write them all down, and [laughs] keep them on a sheet of paper. And then, [laughs] somebody else had to record that in another system to make sure that it was getting from that piece of paper to their digital records. And that's a crazy use of training time. And it means that you've got all of this knowledge that's stuck in someone's head. That's the sort of thing where you could absolutely build a system that could make it really clear instructions on where to find all of them: Look here for that, look there for that, look there for that. And walk them through it and have them write it all down. But that's not really the end goal, right? The end goal is for the user to have available to their recording needs what those attributes are. They don't freaking care about finding [laughs] them. So, setting up a process where they could just hit one button and feed all of those attributes in so that they could pay attention to the ones they needed to know. They didn't have to pay attention to the ones they didn't need. And then, when they even got a new machine, being able to just map those new data points automatically, rather than have to come up with an entirely separate training process. TROND: So, I want to talk about some specifics of some of those examples. But right before that, you laughed when you speak about this, but we're not talking about the distant past here. KIMBERLY: We're not at all, yeah. TROND: We're talking about some existing factories, right? KIMBERLY: [laughs] TROND: That's why it's kind of bittersweet. It's not that funny, actually. It's pretty sad in a sense. But, you know, there are reasons why, which we probably shouldn't get into. It's complicated. But let's focus on the design side of things. It doesn't make a lot of sense. But can you still understand how somebody who has such a process in their—well, that is their operator reality—how they sort of can't seem to think how they could get out of it? But you're saying it could be quite simple to design something that's almost, like, ten times easier. Is that what you're trying to say? KIMBERLY: I think that particular customer found that it was much, much more than ten times, probably closer to 120 times. [laughs] But yeah, we get into these types of processes because we find the way that it works for the first people who handle it, and then we get used to that. And we create training materials, and we onboard one person after another, and ensure their training materials work. And it takes so much time that, oftentimes, people don't have the time to think about how could we do this better. TROND: Do you almost think, like, at some point because you were trained on something, you almost have an incentive to not ask for it to change because it took you that long to learn it? Once you've learned it, not only are you the expert on it, but it's sort of, like, it almost denies the effort you put in if someone suddenly says, "Actually, there's a whole other way here." KIMBERLY: It's the curse of mastery, absolutely. TROND: So, talk to me about some specifics. I know you had an example. You mentioned to me earlier about torque. So, torque is something that I'm not an expert on, but it certainly is the measure of force multiplied by distance, I guess, technically. But it is, you know, torque is important in quality control for various tools, at least that is, like, my simplistic understanding of it. KIMBERLY: Yeah. So, kind of another case where we really found that there were processes that weren't meeting those operator needs as much as possible and probably because people got used to doing it the way that they did. We found that one of our customers, every hour, was writing down four torque values and manually calculating the sum and average of those values and writing those down on paper, plotting those points on a physical graph on paper, drawing lines between them every hour. And then, a supervisor would have to come over, plot those points on their own paper, and bring it back to their computer to transfer it into Excel. TROND: So, the way you're explaining it sounds insane to me. But on the other hand, right? Over-torquing a bolt or something can have catastrophic failure. KIMBERLY: Exactly. TROND: Or if someone catches you, you know, you've produced 1,000 bolts with the wrong torque, or you've put on it a torque that it doesn't have. This is not nice to put into an airplane or even anything else. KIMBERLY: Exactly. And when we think about this as designers and thinking about what those user goals are, we know that that core need is data integrity. We know that that core need is I need to transfer that information to a digital system. And so, we think about what are the absolute minimum number of inputs that I need from the operator? What's the fewest number of people I need involved in this process? Because people have better things to do with their time. TROND: So, tell me then, how could this be simplified so much? KIMBERLY: They were able to do it as one step in a Tulip app with four input fields that automatically graphed in a chart. Every hour, one person has to fill out four fields. And then it's only being written down once, which means that you have half of the amount of error potential that you did before when two separate people were writing it down on their own pieces of paper. And it also means that you can do checks for, are those within reasonable bounds? Because if you are writing it on paper and typing it in Excel, you don't have the ability to look at, other than what's in your brain, like, is this a normal torque value or not? But when you're doing it exactly as that operator is writing it down, you're reducing that error, and you're making it much more possible to pay attention to quality earlier in the process and get ahead of that. TROND: Was that an enormously long and drawn-out process to get to that step? Do you know if the customer then went through 10 different iterations of how this could have been done or? KIMBERLY: My understanding is that they did it in about an hour and revisited how they were charting that diagram two or three times. TROND: They did it in an hour. And then how much time do you roughly save from this or certainly agony you save? Because what you're talking about is there's now an operator who actually has the additional information, additional control over what they're doing. And they can compare with the last hour or -- KIMBERLY: You're saving 10 minutes every hour. TROND: Right. So that would seem to be a good investment on time then to do that. KIMBERLY: Yep. [laughs] TROND: Well, end users, right? They're notoriously central to UX design, right? So, like, what does the user want? But something that's always been puzzling to me is, in your examples here, it's not immediately clear to me that the user is the expert in the sense that do they truly know what their goal is or what they want? Like, they, at some point, get so...and we would, each of us gets so embedded in whatever you think you're doing and think you were asked to do. So, there's a bit of digging involved in a customer interview, I'm guessing. KIMBERLY: Yeah. I think one of the things that is really difficult for a lot of people to internalize when they first start working on design processes, whether they're a professional designer or whether they're designing an app to meet their own needs as part of a different day job, is not paying attention to what the customer wants but what the customer needs, and being able to ask follow-up questions. So, looking at, okay, well, you asked for copy-paste. Tell me about why you want copy-paste. What are you trying to copy-paste? And you're copy-pasting that because you need to do this action 20 times in a row. That makes a lot of sense as to why copy-paste would meet that need. But are there other ways that we could handle that in addition to copy-pasting? So, for example, that actually came up in a lot of our user interviews around app triggers and logic. And we've worked on copy-paste. We're working on it more. But really, a lot of the time, we found that it was because people they needed to write information to tables in multiple...one time for every field. So, one thing that we did when we created automations was we made it possible to write that in a single action instead of having separate actions for each field. And so, we'll still work on copy-paste. But we also solved that core need of why'd you have to do that in the first place. And how can we make that simpler for you? TROND: Can you just explain what an app trigger is for people who are not inside of Tulip lingo? KIMBERLY: It's a way to include logic in a Tulip app. So, when this happens, I want Tulip to respond in this specific way. TROND: All right. So I can see how a bunch of sort of discrete individual operations that were unnecessarily complex could sort of simplify this way. But what about the more complicated larger picture things like supply chains and the effects of small, complicated actions on the entire chains of product steps? The reason I'm asking is that when you're interviewing one user, all right, so they already don't really fully...you need to unearth what they're actually...what the complexity is. When you add 2, 3, or 300 users, I mean, that's like a whole other level of complexity. KIMBERLY: Yep. And I'll also add there that we like to recommend always interviewing five to seven users because when you interview one to two, you're really getting people's individual opinions. But when you do five to seven, you're noticing overall patterns. And above that, many you're largely creating echo chambers. [laughs] But large processes really are not inherently different than small processes. Large processes are a series of small processes that may be sequential or maybe branched. And so, I always look at those types of large processes as breaking them down. One thing that we always start out with is task flow diagrams. What really happens next, and how can I create a diagram of decisions? And handoffs, where, who needs to know what happens next? And does that require a different system? Does that require a user input? What information is necessary at each step in that diagram? And then, you look for the handoff points between them or the different systems that need to be coordinated and can break it out into separate processes from there to focus on. So, whether it's developing things in small processes or large processes, breaking it down really helps even just get into what can our milestones be? And how can we improve this incrementally? TROND: Kimberly, I'm curious, how do you do all this if you were not trained in design? This thing I have in mind is a typical manufacturer may not have a full-time designer on board, or if they do, it could be that that person is either a software designer or literally has spent their life designing physical products. And then suddenly, they're faced with this hybrid situation where they're dealing with a digital transformation. Or they're just dealing with a lot of different design challenges; it would seem to me that are either not taught in school or, you know, these people were never designers in school. They just were told, "Here's the product, and you're in charge of the product." And there is no designer, so, obviously, you're also the designer. How does that work out? Are all these things necessarily you take courses in this, or what should you do? KIMBERLY: It's really true. It can be overwhelming for a lot of people to first look at. But I fundamentally believe that everyone is a designer because you're designing things constantly. It's really just a question of how intentional that design is, right? And so, the way to be a designer, even if you haven't been a designer before, is to start asking why. I think one of the really, really beautiful things about manufacturing is the Kaizen process and the 5 Whys from the Toyota Production System of looking at, okay, well, this is the problem that I heard. This step in our process takes a ton of time and causes a lot of training. Okay, why did that happen? Well, because someone doesn't know where to get this information. Okay, well, why don't they know that? Because it's hard to find here. Okay, so how can we make that simpler to find? Questions like that and being able to really dig into this are the core of design process and something that even five-year-olds can be great at. [laughs] TROND: Hmm. What are some of the biggest challenges, though, in a company where, you know, like the one where you work in terms of implementing all this? I mean, you are one designer or head of design, and then you're saying everyone's a designer. And I guess that means that everyone who works in your company and in each of the client companies they also have to be these designers. Then, how do you sort of set a direction of design? And is that kind of what you're talking about right now, and as long as everyone has roughly the same mindset, then this thing should work pretty smoothly? Or are there frameworks that each company...because we were talking in the prep call, you and I, about how customers are starting to build kind of design systems. What does that mean when you're creating kind of a company-specific way of doing things? What does that mean? Is that a handbook of 100 pages? What does it entail? KIMBERLY: So, ours is a website. But, at Tulip, designers really play the role of researchers and advocates for users, and stewards of best practices, and people who are accountable for making sure that the right thing to meet the user's needs makes it into the product. And one of the ways that we do this is by having team agreements on this is what color our primary buttons are. So, at any place within our product, there should be one clear, obvious next action and a maximum of one. That's our primary action for that. TROND: You know, that sounds so simple. KIMBERLY: [laughs] TROND: But I can tell you the number of software systems, at least that I have come across, which probably at this point is hundreds, they did not have such a button. And when you sign up to somebody's website on an everyday basis, even today, I don't see that green button or blue button. It just doesn't always exist. KIMBERLY: And that's where I think some learning about best practices and really thinking about what is that most important next step in task flows can help with, is looking at, you know, my goal is for the user to know what they can do. And so, my goal as a designer is to create hierarchy in what buttons are most prominent, what text is most prominent, what images are guiding or distracting from when the person first glances at the page. So, we've seen both our Tulip library team and even a few customers start creating templates where they say, "This is what our most important buttons look like. If it's not the most important button, this is what a button looks like. This is how we do a layout of a particular step in our process. This is how we format images. This is where our logo goes," and setting those up as things that people can copy-paste from. TROND: Is it simply branding and color, or are there other aspects of the process steps that can be templated that way? KIMBERLY: Branding and color are fundamental ways to help. TROND: Yeah, I'm not minimizing that because I think that is the core. So many steps could have been actually fairly understandable, even if that was done. KIMBERLY: Exactly. TROND: But I'm saying apart from that, just help me reflect on what other items could go into this kind of template or design system. KIMBERLY: Other things are looking at how do I show the heading text versus how do I show paragraph text? And one of the tricks I like to tell people who didn't go to design school is, if you close your eyes for 30 seconds and then open them up and look at the interface, what's the very first thing you notice? And is that the most important thing for your user to notice? If it's not, then take a look at what is supposed to be the most important and make it more clear, whether that's bigger, bolder, a different color. And keep adjusting like that until it works. And you can do that kind of as a gut check of your own. And then, when you get to something that you're most comfortable with, test it out with an operator and try out, is that the first thing that they saw? Is that the thing they thought that they were supposed to do next? TROND: So, all of this brings me to a question about automation. Because if you think about what's next in most people's notion of industry and stuff, it's like, you know, industry 4.0, and automation, and AI-generated X, Y, and Z, AI-generated user interfaces. What's the promise there? There certainly is hype, right? KIMBERLY: Absolutely TROND: We're going to automate everything. These machines can tailor it all. Like, why are we even asking these questions? Because they have been asked once by someone, and now we can put it into a system. In a certain sense, everything you've said does imply that there is, you know, an answer to many of these things. And you don't have to reinvent the wheel every time. So, there certainly must be potential here. KIMBERLY: Exactly. There are definitely best practices and guidelines to follow. TROND: And the machines can, to some extent, understand and implement those. KIMBERLY: It's something that's really started to come out. Within the last six months, our design team is paying a lot of attention to some of the brand-new tools that are coming out in this. So, like, Figma, which is our design tool of choice for building digital products, they just bought an AI design company, and some other new ones are popping up. We've played with Uizard a little bit. But I think one of the things that's tricky is that oftentimes, the answer to design problems can be, it depends. Because, like I've said a few times, it depends on what's really most important, and we haven't yet seen a lot of AI tools have a great understanding of that. But they are great at noticing what common industry patterns should be because there's a UX law that people spend 95% of their time in other products. And they bring those expectations to other interfaces. So, you'll see almost the entire internet has logos in the upper left-hand corner. Why is that? Because that's where people look for logos now. [laughs] So, what AI tools are potentially really great at is doing some of that grunt work for you of where should I put my logo? Or, if I need an icon to illustrate recent, what are the most common icons that people associate with recent? So that you don't have to do that thinking of, you know, what's right? What are other things using? What concepts makes sense to me? But you can just type in recent icon, and it'll give you one that probably matches your need. TROND: So, if shoe manufacturing and shoes are kind of the extreme high point of customization these days, let's just say, where is enterprise software UX on this scale in terms of maturity? If shoe manufacturing is at a nine right now where, you know, you can get any shoe you want. It can fit your foot. It can fit your neighbor's foot. It can fit your foot today, and your foot tomorrow, and your foot when you're wet, or whatnot. And, you know, all the design choices, you can personalize it, and all of that has been researched. It's not a user group of one, you know, a user group of, you know, oh, you're a middle-aged lady, this is the shoe for you, or, you know, you're a youth, young person, that's the shoe for you. Where are we with enterprise software, and where can we get in terms of customization of the interface, and the overall attention to design, and the feeling that we're actually in a customized environment that matches? Or, I don't know, would it ever exceed consumer products? KIMBERLY: I think enterprise software today is freaking fascinating. I love enterprise software for its complexity. The tricky thing about enterprise is that every business is made up of people. And those people are different, and those people have different needs. And those business processes, as a result, are different. So, the core nature of enterprise user experience is being able to balance, letting each of our customers customize things differently. And that's a tricky problem to solve because you have to look at what some overall patterns are and how to make that customization available to the user in a way that's not overwhelming. At the same time, I think, right now, the employee expectations of the software they use at work have really, really changed in that 15 years ago, ten years ago, you went to work, and IT said, "You got to use this." And you said, "Okay, fine. I hate that, but sure." TROND: You always said that you hated it, right? Because we kind of always knew that it wasn't optimal. KIMBERLY: Exactly. TROND: I don't think anyone was ever duped. KIMBERLY: Exactly. [laughs] TROND: We were not duped. We knew this is not what it should be. You just gave it up. Like, you didn't even think there was a choice. KIMBERLY: Yeah. My wife has explained my work in enterprise user experience to her co-workers as, you know, that timesheet app that we all hate? She makes software for work like that, not be things we hate. [laughs] But the expectations that people come into work with have changed. And they've said, "Now that I have all of these really great apps that I get to use at home, why can't the tools I use at work be that simple too?" So, we're having a lot more software companies look at how can I invest in user experience? If those employees are now actually driving the purchasing decisions or at least the decisions of what types of tools to look into, how can we meet that need? How can we [chuckles] better have our users accept what IT says? Or how can we win over IT? Becomes the question there. And they start building a lot of customization tools from which we can learn. Well, if teams are implementing enterprise-level features across a wide variety of platforms today, what sort of industry patterns are coming up that we can then share and make that experience familiar? TROND: Yeah, it just strikes me that, you know, so much of what we think of when the phrase digital transformation is used, it's kind of shocking that the word design is not part of that. Because if you're transforming it digitally, you're not doing much, right? I mean, what does that even mean? KIMBERLY: Right. And what we were talking about before, if we simply made a digital tool that told people where to find those 150 things and write...that's not helpful. TROND: No, right. KIMBERLY: That takes exactly the same amount of time. TROND: Exactly. So, you know, digitizing something, in and of itself, is not where the efficiency is because if that were the case, then these industrial tools we've had for 30 years would have brought us to a different place, right? Flying cars and whatnot 20 years earlier. But they just didn't, right? It took a long time. It seems to be there's a little confluence of maturity of design thinking and maturity of digital interfaces, and the two they need to work together, I guess, to get these improvements. So, does that mean you're optimistic now? Or are we in a good...the digital opportunities with the design sensibilities it's starting to gel? KIMBERLY: Yeah, I think we're at a really cool inflection point. And we've got customers and people all over the world looking at how can I make this system best for my factory and my people? And being able to answer, what does it mean to be best? How can I meet all of those goals and designing what that process can really look like? TROND: Well, thrilling. You seem enthusiastic about where this is going. I hope that inspires some people out there who are designers who are working on enterprise products for industrial uses. So, thanks for elucidating some of the challenges, but also, I guess, kind of infusing some optimism here that we're on the path towards products and work realities, operator, you know, experiences, if we're working in factories and other places, that are starting to make more sense. Working in a way that feels natural, where the interface is not disturbing you and irritating you so that when you go back to your hobby interfaces, there's this discrepancy because that's what you've been talking about, right? The discrepancy has been enormously large. KIMBERLY: Exactly. TROND: But it's no surprise, I guess, because if design was never thought about, then why would they automatically turn into some brilliant user interface? Is that always the case, or are there inventors who can skip this entire process? I mean, Steve Jobs famously quipped that...I don't know exactly what the quote is. But, you know, his intention was always to build something for the user. So, you didn't really have a product unless you had thought about the user. I think that's sort of the thrust of what his whole design philosophy is. But not everybody is that way. KIMBERLY: I think one of the Job's quotes that I personally love best is, "Design is how it works, not how it looks." I think that's something that people can get confused about a lot because they think design is about making it pretty. Pretty comes because goals are met, right? Pretty comes because it's clear what the balance between elements is, what the purpose of something is, and what you can do with it. I think that comes naturally for some people, and some people learn by asking why and asking how things can be better. TROND: Great. Thank you so much for explaining all that. It's been a pleasure having you. And good luck explaining how these things work for new generations of designers and engineers. KIMBERLY: Thank you, Trond. Great chatting with you. TROND: Thanks for listening. If you liked the show, subscribe at augmentedpodcast.co or in your preferred podcast player, and rate us with five stars. The Augmented Podcast is created in association with Tulip, the Frontline Operation Platform, connecting people, machines, devices, and systems in a physical location. Tulip is democratizing technology but also, importantly, empowering those closest to operations to solve problems. Tulip is also hiring, and you can find Tulip at tulip.co. Please share this show with colleagues who care about where Industrial Tech is heading. You can find us on social media. We are Augmented Pod on LinkedIn and Twitter and Augmented Podcast on Facebook and YouTube. Augmented: Industrial Conversations that Matter. See you next time.