Michele Hansen - Standing up for User Research in the Age of AI (with Michele Hansen, Founder @ Geocodio & Author “Deploy Empathy“)

There is such incredible insight that comes out of really understanding people's problems and how you might solve them and how your, your product is currently solving them. There is simply no replacement for that. AI is incredible. It can speed up our research. It can also speed up development. The problem is when you add AI into a company that does not have a clear product vision, a coherent product strategy, and a deep understanding of the customers, you're not going to accelerate your product development. You are simply going to accelerate running in place or running in circles. Hello and welcome to One Night In Product, the show where I chat to some of the brightest minds in product from across the globe to help you see product management in a whole new light. If that sounds up your street, don't forget to dive into the back catalog on your favorite podcast app or on YouTube. And of course, follow, share, subscribe, drop me a comment or review, it all helps keep the lights on. My returning guest tonight is Michelle Hansen. Michelle is the co-founder and CEO of Geocodeo, a geocoding and data matching platform that's apparently used by thousands of happy customers. Some of which Michelle has clearly spoken to as a passionate advocate for user research and the author of 2021's Deploy Empathy, a practical guide to interviewing customers. She's recently brought out a new edition of the book and she's here to talk all about whether we need to do any of that pesky stuff anymore. Now we've got AI to tell us the answer to all of our questions instead. Michelle, thanks for coming and welcome back to the show. Thank you so much for having me back on. I'm really excited. Well, it's been a minute, right? But, you know, hopefully we can sort of pick up from where we left off and see where we can go in this new age of future that we're in right now. But like we said, it has been a while. So what are you up to these days? What's going on? She's still running the company, got the book. But like what in general is keeping you busy these days? Honestly, the company is keeping me quite busy. So just for context on my background, I was a product manager and then in 2017 went full time on my own company that my husband and I had started as a side project a couple of years beforehand. So that was in 2017. I went full time and we ran it just the two of us for some years. He's a developer. You know, I'm in product. So it was a nice match. And but now, you know, we have a team of eight. We've got a ton of new features coming out. It's been it's it's been a lot of growth in the last couple of years. And honestly, I don't know how I found time to write a new edition of the book. But as you said, with AI and everything, there was a real need to update the book as people kept asking me about that. But there's also other additions to it as well. Yeah, well, that's a very well, first of all, is an interesting point there, because you could cynically say or I could cynically say, well, the easiest way to just, you know, update the book would be to feed the whole transcript into one of these large context window LLMs and just get that to do it for you. But I'm assuming that you may be used a little bit on the side, but that this is still kind of this is stuff that's come out of your fingers still. It's entirely stuff that's come out of my fingers. And it was sort of funny. I use LLMs a little bit when I was editing the book. You know, there was a there was, of course, a professional proofreader and indexer who went through it. But then I naturally, being the brilliant person I am, made edits to it after they had edited it. And so I had to make sure that I hadn't introduced typos or change the page references in the index by mistake. And I was like, oh, I can just have Claude help me with this. And it couldn't. It turns out Claude really, really struggles with PDFs. It kept telling me that there were typos where there weren't typos. It was it was not nearly as good. But all of the content came from me. But some of that content is around how you can use AI in tandem with your customer research to accelerate, you know, the cycle time of doing research. Well, let's talk about some of those differences. And again, like you've mentioned, AI a couple of times. I imagine that's one of the big differences. But, you know, the book itself. Well, actually, before we even talk about differences, you know, maybe some people missed it first time around or they missed this interview or the interview last time around. What's the actual kind of in a nutshell, a core proposition of the book as a whole, like the original version and the new one before we talk about differences? Yeah, so it's called Deploy Empathy, a Practical Guide to Interviewing Customers. And that's what it focuses on is doing customer interviews. This came out of my own experiences working with other founders who many of them were developers. They're running small companies, startups. They're getting them started. And they really didn't know how to talk to customers, some of them even people at all. Like they've told me that they really struggle with social interaction in general, but they really wanted to understand what should I build for people? You know, how do I understand what they need? How do I build something that they're going to pay for? And so I kind of was looking at all of the existing work on on customer interviews. And first, I was sending people a jumble of resources that probably didn't make any sense and didn't make it seem very actionable. Book chapters here, podcasts there, you know, sorts of stuff like that. But really what I what I found is I need I needed a book that taught people how to do customer interviews from the ground up, assumed no experience with product, no experience with UX, no experience with research. And with something that they could just, you know, skim through and then figure out, OK, I've got a churn problem. Here's the script I need. I can go run with this and, you know, talk to five people about it. The the core, the sort of the heart of the book, as many people have said, is this whole section called how to talk so people will talk. I find that when I talk to people and they say that they tried customer interviews, but they didn't get anything out of it. Often it's because they weren't taught how to behave and treat the other person in the interviews so that they feel comfortable with them. It's a very particular way that you interact that is very different than a normal conversation. Quite frankly, it's more like a therapist interacts with their patients. And and so people need to learn that. And so in that section, there's all of these ways that people can practice that on their spouse and friends before they try it out on customers. Don't make me talk to my wife. Come on. That was actually the really amazing thing is that people told me that it had not only improved their work life and the results are getting out of it, but actually had improved their relationships. And that was like such an incredible, incredible thing. Really moved me. But so what surprised me was I wrote this book basically for developer founders because I was like, there's already so many good books out there for product people, for UX people, for design people. Like there's enough books for them. I'm just going to write for this one really small audience. And then it turned out that product leaders and design leaders like they needed a book to teach their employees how to do interviews as well. And they've been some of the most avid adopters of it. There's people who've led workshops at their product meetups with the there's a whole customer interview workshop bit. Actually, that's in the new book. And one of the appendices is a workshop guide. And so I've been really, really surprised, quite frankly, because I just I thought there were enough books for us, but apparently not. There's always another book to read. But I also hope you're getting royalties for the workshops as well. You know, you've got to get your beak wet. But what are some of the new things? We mentioned AI, but is it just updated for the age of AI or the other kind of bits in there or maybe even changes? You know, things that you've adapted your thinking and over the intervening years? Yeah, there are some things I adapted my thinking on. For example, when I wrote the book, it was early 2021. And so, you know, I was in the middle of a COVID lockdown. At that point, a lot of customer interviews were still over the phone. I think the switch to Zoom and video had really started. Before COVID, I never did customer interviews over Zoom. It was either in person or it was on the phone. And so I had written something in there, you know, saying that I suggest doing it on the phone. But now that's the normal way to interact with people, which I think is a little bit more pressure on a new interviewer because you have to monitor your facial expressions. You have to monitor their facial expressions. It's a little bit easier when you only have voice to pay attention to on both sides, but it is helpful. It does really add a lot. I added to a bunch of chapters. So part of the book is there is a sample customer interview, both in transcript and audio, if you go to the book website. And then I do a sample analysis chapter. And so I added a bunch of new analysis pieces in there and also examples of them using the sample interview. So, for example, there's an example forces diagram. There's a switch timeline, and now there's actually an image of the timeline. One of my favorite additions to it, which I think this is just sort of my favorite chapter, is on the neuroscience of listening. And in the original edition, I talked about what happens when someone is listened to. So they've put people in fMRI machines and scanned their brains and said, what happens when we have them talk about a general topic or another person versus they talk about their own experiences? And then what happens when they talk about that just, you know, into the ether versus they're actually know that they're talking to another person, another person is hearing them. And parts of the brain related to pleasure and enjoyment light up when people talk about their own experiences to other people in a way that they don't when they're talking about someone else or not talking to someone. And I have found that this applies even in a business context. I think, quite frankly, a lot of the things that we dive into in a customer interview, nobody has ever cared to ask them about their laborious invoicing process that they have to do for two hours a week. Right. Nobody has ever bothered. And then here you have someone who is absolutely wrapped and waiting on your every word and is like, so, so then, OK, then you click that button. OK, what happens next? Tell me. Right. And of course, we don't use that tone, but that's basically how the other person feels. Right. That someone is excited to hear about this kind of boring thing. Right. But something I added to the new edition that I really loved is also what happens in the interviewer's brain. So not just the person who is being interviewed. And again, these are fMRI studies where. So they put people into the scanner and they showed them a bunch of fuzzy images. And what this is simulating is the experience of curiosity, and that is what we are diving into when we're in an interview. We're trying to understand what this person's process is. We probably have a sense for the general picture, given all of the other customers we've talked to. But every organization is weird and unique in its own way, in the same way that every person is weird and unique and delightful in their own way. And so what this study did was to put people in the fMRI machine, showed them fuzzy images. This was a very frustrating experience for them. Right. Their brain is like, what is going on in this picture? Don't know what it is. And then when they eventually showed people what the images were, those same parts of the brain that lit up when people were talking to other people about their experiences, the ones associated with pleasure and enjoyment, those light up when curiosity is satisfied. And so this not only means that the person you're interviewing is finding it to be an enjoyable experience, but you yourself are going to find it to be an enjoyable experience because your curiosity is going to be satisfied. And I think this is something that especially people who are new to interviewing are nervous about. They're like, is someone really going to want to tell me about their boring invoicing process? Like, and like, are they going to want to talk to me? Oh yeah, you can't stop them once they get started. Exactly. Right. Because it's, you know, the dopamine is firing in their brain. But then also for people who are like, well, am I going to, you know, am I going to know what questions to ask? Am I going to know how to drive this interview? And yes, because again, the dopamine is going to be firing in your brain. Your curiosity is going to be going. You're going to be trying to make those images crisper, take them from fuzzy to clear. And so that is, honestly, that's my favorite part of the new edition. And I'm so glad you asked because yeah, everybody wants to talk about AI. I want to talk about FMRIs. But just to be clear, it has AI in it as well, you know, you don't want to, you don't want to kind of diminish your potential reach to the brim. Yeah, there's a whole chapter about how to use AI in your customer research to speed up the process. And how I use it on a daily basis. This is everything from transcription, which when I started interviewing over 10 years ago, I mean, you had to like you do an interview and you get out of it and you'd be so excited to go back to your team and be like, oh, my God, they just said this amazing thing. And it was so good. It was like this perfect nugget. And they just like all this really interesting stuff about their process. But then you've got like maybe got some like post-it notes with like some scribbled notes on it, or maybe you printed out your script and you're kind of drawing and diagramming. But like you don't have that amazing money quote, right? Because you have to send off the recording to somebody and wait like two weeks for it to get transcribed. And then you need to sit there with your highlighter going through it, which I kind of love. But like, right. It was a long process to get back to that point of like, oh, my God, they just said this thing. And it was so, so great and so insightful. And so now you can just take the recording, throw it into Otter or a whole bunch of other transcription services. Get the transcript back right away. That's like 90 percent good. It's amazing. I mean, I can go from being in an interview and then getting it back to people in half an hour. I think it's also pretty decent at doing basic analysis. And I include some of the prompts that I use in the book. So, for example, constructing a jobs to be done timeline. I think it can do a decent job of that. I would say it's like 70, 80 percent good at that. I think it can pull out, you know, satisfaction, dissatisfaction quotes, you know, information on like what did they use before? What did they think of competitors? Where do they struggle right now? Pretty, pretty decent. I want to I'll give it like a 70. So where it really fails, though, is figuring out what's important from an interview, no matter how much context it has on my business. It cannot tell me what I found interesting and useful and insightful and that weird, spiky thing that was new that got me really curious, which is like the best part about an interview. And so in many ways, I like to think of AI like a research intern. So it's really good at clerical tasks. It's really good things with like sort of a very straightforward scope, right? Like think of it as a super smart, super competent, super fast intern who just started last week and has no context on your business and no practical life experience. And so like you would so like for your intern, you would have them transcribe something. You might say construct a job speed on timeline based on this interview. Pull out their key satisfaction, dissatisfaction quotes with our products versus other products. But you wouldn't say, how should I change my product strategy based on this interview? What should I prioritize in my roadmap based on this interview? You wouldn't ask your intern those questions. And if you do, you should hire them. All the time, right away. So I think that's sort of a framing that I like to use. And I include that in the book, that it's good at intern tasks. Things that are not intern tasks, things that really require strategic thinking that require, you know, the many years of industry understanding and customer understanding and all of that that you have. It can't replace that no matter how much you train it and how much additional context you give it. It's never going to be you and it's never going to have the level of strategic understanding that you have. But it can do so much to make those steps in the process that get you to that point. And communicating with your team can do that. It's also really helpful for digging into your customer research library. So there's a founder I know, for example, his name is Ben Aldred. He runs a company called ConsentKit, which is for user research teams to manage consent. And he uses it with all of his job stories. He has all of those in a database already. And then he can basically query that with AI and say, OK, I'm thinking about this. You know, which of my customers have a similar use case to that? And or who should I be talking to about this feature? But he already has that really strong base of quality research to pull from. And so to me, even in this age of AI, I think we have to keep humans at the core of it. We're doing human driven, but AI accelerated research. We're not talking to LLMs, we're not serving the LLMs, but we're talking to real people and then using the LLM to accelerate our process. That's my soapbox. I was going to say that sounds like just the sort of thing that a human would say if they were trying to fight back against the inevitable march of the machines. And I guess, yeah, obviously, I use AI for all of those things as well. And I'm sure you and I share very similar opinions about much of what you just said as well. I want to dig into every single one. But one interesting thing that I found, you know, I do a lot of interviews, both for the podcast, but for my day job as well, working with different companies and trying to dig into what's going on. And there was a point where I was like, hey, exactly as you said, yeah, I can just use AI for the transcriptions. I don't need to take any notes anymore. I can concentrate on the person, all good stuff. But then I started to be suspicious that actually, sometimes, you know, you talked about it being whatever, 70, 90, whatever percent good. Sometimes it would miss stuff that actually I felt was really important, like me, myself. And that's something that because I did the interview, I could just go back and check, you know, maybe check the transcript. But then I actually started to find myself just sort of taking notes alongside the notetaker as well. So that rather than just, you know, focusing on the person and just letting the machine do everything for me, I'd actually end up taking my own sort of stickies as I went to sort of almost remind me what to go back and check, because there were things that stuck out, which I wasn't 100% confident that the LLM would actually pick up, because sometimes it misses stuff, right? So I guess that's an interesting question. Like, is that now a solved problem? Is that something that still goes on? Are there kind of any strategies that you recommend to try and make sure that you don't just get the basic stuff, but you actually do get those little sort of nuggets that actually are the things that you really need to sort of drill into? Yeah, I have a similar process to yours. I have a pad of sticky notes here that I end up using during interviews and, you know, and sales calls as well, right? And we often kind of borrow a little bit from interviews. And, you know, it's things to come back to, you know, in the interview later, right? You don't want to forget that. And also things you want to make sure that you note later, because, you know, I find that I go into an interview wanting to hear about one thing, wanting to learn about one thing. And the best interviews are not necessarily the ones where I heard about that. It's where they heard about something, where I heard about something that I hadn't even realized was there, right? It's that weird spiky thing that's off to the side that, you know, if you were interviewing an LLM would never come up, right? It's that weird thing that makes my brain go, wait, hold on a second. Okay, there's something here, or there's more here than I realized, or this, okay, wait, this is vague. And it wasn't even necessarily what you came in there to learn. Those kinds of things, I find that the AI skips over because you're giving it a prompt. It kind of figures out, oh, what's the general topic of this? And then it stays on that general topic. And it's like, but actually the thing that I need to dive into deeper, that's the action item for me, is this thing that's way over there that the AI has completely skipped over. And so I think there is definitely still a value in taking your own notes. I still like to have a giant pad of artist's paper next to me when I am doing an interview, because I find myself, you know, diagramming their timeline, diagramming their process. And that just, and that's just for me, that it helps me think in the interview. And sort of, I tend to process as we go. I know some people really prefer to save all of the processing until later, but I don't know if it's because I'm ADHD or whatever it is, but like, I like to have that going on as well. And it really helps me dig into their process. And so I don't think the AI note taker, it's transcribing, but it's not notes. And I feel like notes are the important thing. Transcript is just what happened. It's not the why. It's not the so what. But there's a second part to that as well. It's not just about the note takers, although obviously that's a big part of it, and probably the kind of the headline use case for some of this stuff. But there is another trend that's coming up, and I'm sure you've seen it. You follow all the same people on LinkedIn. The kind of concept of synthetic users, and this idea that rather than going out and speaking to real people, and using AI to transcribe or to summarize and do all the things that you could do with that, that actually you can just go out and much more quickly and cheaply and reliably just go out and speak to a bunch of AI agents, and they'll come back and tell you what these types of people would want. How do you see that? I mean, that feels like it could be interesting as an experiment, but do you think that's actually a replacement or even an enhancement or just a distraction from going out and speaking to the real people that are going to actually be using your product? There's a quote on this that I saw a couple of months ago that I really liked. Quote, you're never going to stop talking to real people, and you shouldn't. There are some decisions that shouldn't even go to synthetic users. You should just go to humans. That's from the co-founder of synthetic users. Is that Hugo? Yeah, I think that's the kind of perspective that we should take, right? There are some things that you can use simulated users for and some things where you really need to go to humans. I want to actually shout out really quickly that Nielsen Norman Group, they did a really great blog post series this summer on AI and research, and research on AI and research, I should say. Really great series. I mean, their newsletter in general, I think everybody should be subscribed to it. So there's two different types of simulated users. There's digital twins, and then there's the simulated users. The simulated users, they're based on a generic LLM base, and they're intended to simulate the responses of a group. I think these are good for building general understanding. So for example, if you're going into, if you interview people in an industry that you know absolutely nothing about, and this industry happens to be full of people who are English speaking and talk about their industry a lot on the internet, then you can use it to say, okay, help me understand in general the problems that a medical office receptionist in the UK might face on a daily basis, right? Like you could probably just get a general sense if you have no understanding of that user. It's not going to replace talking to real people, but if you've never been in a medical office before, never talked to a medical office receptionist, it'll at least kind of get you started and help kick you off on your own research. But it requires that that group is English speaking and talks about themselves a lot on the internet. If you, by contrast, want to understand garment workers in rural Vietnam, you are not going to have a lot of content on the internet about them because the LLMs are mostly consuming this content in English. It's consuming content about people who are well-documented on the internet, right? And so there are really serious issues with bias and lack of understanding that come into this. I will say that there is a type of synthetic or simulated user that performs a lot better, which are digital twins. So digital twins are based on real research data, such as surveys and interviews, and they're intended to simulate a real user rather than a group of users. So for example, let's say you sent out a survey and you had a thousand customers fill it out, and then you had another 700 people who only completed 75% of it. They do much better when you feed them all of that survey data, and then you say, okay, if these 700 people had completed the survey, how might they have filled it out and can we get a fuller picture from this survey? The thing is that they require extensive real data to get that information. And while you can use it to backfill, what they found is that there is less difference in the responses that they get, which is to say the standard deviation in the responses is smaller. And kind of as we were already talking about, oftentimes it's those like spiky, weird responses that you get that end up being the most interesting and the most thing you really wanna dive into. And so when you're using digital twins, they're not going to give you that range. They're not going to give you those spiky answers, but it can help you sort of fill out the rest of that survey. And again, get a broader picture, but they have to be based on real actual users of yours. Yeah, it reminds me of some work that I did back in the past, an old job. This was pre-LLM, although we were certainly playing around with some stuff that was itself kind of a precursor to LLMs. And we're doing a lot of work around trying to basically predict the future by taking down social data and using natural language processing to predict the future of consumer packaged goods. So it was, yeah, people were talking a lot about this stuff online and they were talking about their sodas or they were talking about their potato chips or crisps or whatever you wanna call them, depending on which side of the pond you wanna put your chips or crisps. But we had a real problem when we started moving into markets or moving into categories that again, to your point, weren't really being spoken about. But it is interesting because like the very value proposition of that company and of a lot of the companies that are out there now with the LLMs that I wish we'd had back then, is that to some extent, human behavior is predictable. You know, you've got the spiky stuff around. is predictable. You've got the spiky stuff around the side that you've kind of mentioned already, and obviously agree that the only way to get that is to go and speak to actual people that have those kind of out there, kind of around the middle kind of needs or whatever. But at the same time, there are a lot of things that are probably pretty predictable to your point around medical receptionists or whatever. Again, with the caveat of the bias and underserved people and such, but there are many human attitudes or requirements that are fairly straightforward, and you could probably assume based on just general knowledge. So is it really a problem if you're going to get the same answer back anyway? Like, you could go out there and do bunches of interviews with a bunch of people and get the same data back basically that you would have got 90% as you would have got from the LLM because the problems that you're going out there are actually so generic that actually anyone would have said them and you just didn't know them yet. I guess it's kind of just reinforcing your point. But is it a problem? I would say it is. And there's another study that looked at process. And I would agree with the general point that a lot can be generalized. A lot of things are in common. I think this is what the strength of jobs to be done is, right? Outlining a buying process. There's similarities in an industry. There's going to be similarities. They're going to be sort of these, you know, sort of headline points in a process that are going to be shared. But every human and every human organization does things differently in their own way. And then they have their own idiosyncratic reasons that don't make sense, or they do things that are outside of that process. So there was one study with, I want to say this was with synthetic users. And I want to say it was actually Jan Vermouth of the Product Quest podcast who told me about this one. They're doing research on doctors and their process with patients. And it was talking about their process with meeting with a patient and then, you know, taking notes on the visit, you know, charting it and all of that. And what they found from the LLM was that it basically gave it that process, but it was an idealized version of that process where the doctor goes from, you know, sitting in their office, ready to receive the patient. The patient comes in. They take the patient's symptoms. They use their knowledge to figure out what the patient might be experiencing. They sit there. They order tests. They tell the patient what they should do. You know, the patient goes off on their merry way. The doctor then types up all of the notes. This is a visit, right? When they actually went to talk to people, they found, well, OK, well, the doctor was tired that day and they hadn't had their coffee yet. They actually, you know, they told the patient they were going to order the blood test, but actually they forgot. They didn't do it until six hours later. You know, the patient left, but then they came back five minutes later because there was something else they had to mention to the doctor that they had forgotten about. Like all of these other different things happen, even if that it's it can be true that there is a general process that applies across organizations or across context. But it's also true that things never go according to the exact plan as it looks in a textbook or in an LLM. And I think that's where the really interesting stuff is. Right. Because if I was building software for doctors and I found out, oh, they forgot to order the blood test until, you know, you know, six hours later. I know that there's a lot of like AI transcription tools for doctors now. Like that would be the thing I would make sure gets on the doctor's to do list. Right. Like that's like maybe there's some sort of product here. And I wouldn't have known that had I not gone to speak to this real person and understand their reality on a daily basis. Now, of course, you can't take one person's experience, one person's idiosyncrasies and make a product out of that. But often you find those idiosyncrasies across organizations and across people. And that's really where the magic is. And and I think that is is what you miss. And. But also, even if you already felt like you really understood that process, you're going to learn things that you didn't know. It always happens. I have been in the geocoding and location data analytics business for 12 years and talked to thousands of customers. And I learned something new every single time, whether it's about why they're using the product, how their organization works, how they make decisions, the types of tools they're using. All of that is constantly changing. And I'm always learning something new and and sharing that. And and I don't remember if it's Christina Walker or Erica Hall who said that the point of research is not to do it as fast as possible. The point of research is to change you. It is to change your organization. And I think this is Erica Hall that, you know, the point of the research process and bringing in. And I think her point is more around bringing other people into the process. Right. So instead of she's another point that I love, which is, you know, instead of being handed the insights from the smart people, it's so much more fun to be one of the smart people with the insights. Right. And so bringing the rest of your organization with you in that discovery process, in that process of curiosity and the images becoming from fuzzy to clear. And so I think that's what you miss if you're only talking to an LLM for your research. It's. Being changed by the experience, realizing that the idea you had at first is actually not the biggest opportunity. There's this whole other opportunity somewhere else that you didn't even realize was there. Well, this is an interesting point, because whatever we say about AI kind of touches on a more fundamental point about doing any of this stuff. And that's just the very concept of the value of talking to users in the first place. Right. Now, we chat about this a little bit last time we spoke, and it's something I know that's very much on your mind. And I know that because you did, you've done at least one talk on it, back at minor product, where you, let's say complained about a very specific quote, that people always trot out when they're talking about the value of doing or not doing user research, which is the allegedly famous Henry Ford quote of if I'd asked people what they'd wanted, they would have just said faster horses. Now you see that on LinkedIn posts all the time. And people, you know, keep trying out this arguable, in fact, probable that Henry Ford never even said it. But apart from the fact that he didn't say it, what's the problem? Or what are the main problems that you have with that quote and what it kind of implies or means for the value of user research? Yeah, I don't think there's a product manager or a UX researcher out there who hasn't had this quote lobbed at them to completely disparage and discredit the idea of talking to customers and the value of it. I think it kind of makes all of our blood boil to hear it. And as you said, so first of all, Ford never even said that quote. He never said, if I just asked people what they wanted, they would have said faster horses. He never said that. Okay. He said a lot of other stuff that we shouldn't talk about. No, but certainly look him up on Wikipedia for some of his interesting opinions. What is it with these kind of business entrepreneurs, these automotive entrepreneurs that say that, you know, get lauded for their technical and business credibility, but at the same time, just don't seem to be able to stop insulting one minority or another. Like, what's with these people? Well, maybe Ford started that trend. There you go. Yeah. So I think something with this quote, first of all, it implies that Ford invented the car, which he didn't. So first of all, Daimler invented the engine and Carl Benz invented the first car, which was a redesigned bicycle. So Ford didn't invent the car. The quote implies he wouldn't have invented the car if he had asked people what they wanted. Okay. He didn't invent the car. Second one, people would not have asked for faster horses. So think, you know, let's go back in history a little bit. This is around the turn of the 20th century, late 1800s. Cities were full of horses, how people are getting around, how a lot of goods are moving through cities. And the problem with horses is first they get tired. Second of all, they poop a lot. So now... Sounds like me at my age. So cities at the time were absolutely full of horse manure. Now, in the early 1800s, they were actually able to sell the manure back to farmers in the countryside, use it for fertilizer, the food then comes into the city. What a nice little circular process we have. Well, there were so many horses in cities and cities had grown so big by the end of the 1800s that there was no market in selling the manure out of the city. And so it just piled up on city streets. And so you had piles of manure and also piles of dead horses because people would just leave them in the street. And so there would be streets that were completely blocked because of the manure. When Victorians talk about mud, right, and how, you know, men would stand on the outside to protect women from the mud, right? What they actually mean is this sludge of, you know, water and horse manure that was covering all of the cities. This is also why you had a lot of issues with diseases like cholera. Huge, huge problem. So there were manure crises. So people would not have asked for faster horses. They would have asked for horses that poop less. And quite frankly, if you put it on that perspective, a car pollutes more than a horse does. So I think it failed on that regard as well. He also didn't design his car. He had, you know, a sort of a Waz basically, who was the one who did a lot of the design work. Of course, Ford contributed to it. You know, he created the assembly line after watching Chicago meat packing plants. I learned this recently that disassembly lines in meat packing plants to take apart, you know, a cow or whatever, the disassembly line was invented before the assembly line. So did you say taking apart a cow? Yes, butchering it. Wow. I just never heard of it referred to as taking it apart before it kind of. I was trying to be delicate. Maybe there's vegetarians listening or just people who like cows. But yeah, actually, the Chicago had a huge meat packing industry and those that they invented the disassembly line and then Ford saw that and invented the assembly line. But anyway, so he was not the sole designer of his car. There was a Waz there who did a lot of the design. And I think the big point here is Ford, he also completely ignored customer preferences. So the Model T, it only came out in one color. You couldn't customize it in any way. It didn't come with a windscreen. There are all of these things that it didn't come with. And there was actually a time when he went off to Europe, I believe. And so his second in command decided to redesign it based on all of this feedback that he had heard from the customers. And then when Ford got back, he was so mad that he smashed it with a baseball bat. The biggest customer preference that he ignored was how people paid for it. So Ford required that everybody pay for their car upfront and then wait for the car to be built. But that required people to have a lot of money. Cars in the early days, they were basically toys for rich people. And GM, they recognized that if they could finance it, if they could let people do installment payments, they would actually have a much higher likelihood of selling cars. They would have, they would sell a lot more cars if they did that. And Ford refused to. And so even though Ford had this early mover advantage, you might even say sort of a first mover advantage in the U.S., once GM came online with cars in different colors, with ridiculous luxuries like windscreens, with the ability to finance the car, Ford started losing market share. So Ford's market share peaked at 55% in about 1923. After GM introduced, you know, the customization and the financing, Ford's market share dropped to 10%. And it has never recovered. Ford has never recovered their market share compared to GM or in the industry in general. And so I think if this is a person we're saying, well, if Henry Ford had asked people what they wanted, they would have said faster horses. He didn't invent the car. They didn't want faster horses. And he wasn't even very good at business because he lost his first mover advantage and plummeted his market share. So is this really someone we want to look to as an example? And that's not even getting into all of his opinions on minorities and how he treated his workers. We're not even getting into any of that. But on the other hand, he didn't actually say it. Right. It was made up in a marketing week, a letter to the editor. This is the first anyone had ever been able to find this quote being attributed to him. And I think it's good to start with that because when somebody, you know, says this to you, when somebody just sort of lights this quote on fire and throws it at you to extinguish all of your dreams about having a customer-driven product strategy. Right. Like it's so convenient to be able to say, well, he never even said it. Right. Like as a like, well, actually. But I think we really like dive into, right. Like why would somebody bring up this quote? Right. Like it's really tempting to, well, actually it and say, well, actually Henry Ford never said it. Actually, GM ended up eating Ford's lunch and their product and their growth stagnated. You know, actually horse manure was a big problem. Nobody would have wanted faster horses or they would have wanted once it poops less. I think what that person is telling you that they're afraid. Right. Like if you're in an organization where, you know, there's the founder who maybe feels like they're spending more time in meetings than they are interacting with the customers and in sales as they used to, maybe they feel like they're out of touch and they're afraid. And that's the way they're keeping their control over what that vision is because they're afraid to give it up to, you know, a product manager talking to a customer. Like, who do you think you are? Right. Who do we think we are to do that? Right. I think it really comes from a place of fear. And I think in the same way that we use empathy to understand our customers, I think we should also use empathy with people who are saying that. Right. So like, even looking at sort of from a jobs to be done perspective, right, like from a functional perspective, they don't understand the value of talking to customers. From a social perspective, maybe in your organization, you know, ideas usually come from the top, hippo, rather than the highest paid person in the room, rather than from the bottom, which is where a lot of organizations put customers or an emotional, maybe they're threatened by a new way of doing things. And I think we want to well actually it, but maybe we keep that in our brain, all of the facts we know about that quote and about Ford and say, it sounds like you're not sure about the value of talking to customers and the value it would bring. And maybe it doesn't jive with how we've historically done things here. Can you tell me about that? Right. Like dig into it with them. And then I find that the easiest way to get people on board who are skeptical is to get them into usability testing, especially if it's like a high converting landing page. And if you can show them like, oh, wait a minute, hold on the person, like they actually can't find the buy button. Wait, no, no, no, the login button. No, it's it's right up over there on the right. No, no, wait, why are you going back to your email right now? Hold on a minute. What's going on here? Like, like, why, why? Like, why didn't you read this section? And you know, right, like they get like excited and like everybody's sitting in the room being like, find the login button, find the login button. Right. And it creates this sense of them being one of the smart people with the insights. Right. If you can bring them into the process. And I think actually there is a real quote from Ford that I think encapsulates this nicely, which is if there is any one secret to success, it lies in the ability to get into the other person's point of view and see things from that person's angle, as well as from your own. Now, like the quote we've talked about, it's unclear or doubtful, really, that he lived this. But yeah, I was going to say he probably could have done that a bit more in a number of scenarios. But he did see the value of it. And that's something you can play with. But is by by using that empathy on people, using your customer interview skills on your doubters, and then bringing them into the process and chuckling to yourself and your teammates about manure crises, and Ford losing his market share. I just pulled up another quote, actually, because you mentioned, you know, Steve Wozniak is the quote from Steve Jobs, which I've just pulled up my phone. Some people say, give the customers what they want. But that's not my approach. Our job is to figure out what they're going to want before they do. And then he says the Henry Ford quote as well. Thanks, Steve. Thanks, Steve. But this is another thing because of course, so many people want to be Steve Jobs, right? You know, probably fewer people want to be Henry Ford these days, especially given all of the stuff that you know, people now rightly judging for. But at the same time, you know, when you've got your Steve Jobs's of the world, a misquoting it, but be kind of doubling down on the idea. But to me, it almost kind of, I know, this is obviously kind of a core theme in your book as well, it almost shows this kind of idea that and the kind of the general objection to going out and speaking to customers or users is like, we can't go out and expect them to design a product for us. That's not what anyone's saying that you should do, right? Like, the point of going out to these people is to find out what problems they're having, for example, that their vehicles pooping all over the place and try and then come back and work with your team to work out the best solution to those problems, right? Rather than going out there and saying, Well, would you like a horse? Would you like a cow? Would you like a car with a motor in the front? Like, that's, I think that's the biggest thing for me is this kind of disconnect where people think that the reason that you're going out to speak to people is to ask them to tell you exactly what they want you to do. And it's not. No, it's not. It's not. I think that the customers, you know, job if they have one, right, is to have a problem that they're willing to have solved. And they will happily have problems because they have so many problems. People have so many problems. Organizations have so many problems. Your job as the product manager or as the company designing and selling a product or a service is to figure out, okay, given your capabilities, your experience, your connections, your network, your resources, your competitive positioning, what can you build for them or offer them that solves their problem in a way, you know, um, that is faster, better or cheaper, right? How, uh, you know, to quote Marty Kagan. So, um, that is your job. That is the organization's job. It is not the user's job to tell you what to build. And as you said, if you go out there and ask them what you should build, you're not going to get very good data back. It's not the point. The point is to understand, you know, how, how it should work for them or what their problems are. And, and, and that's where your own background, your own experience, your own, um, view of the industry and your own conversations with other customers come into play to figure out, okay, where is there actually an opportunity where people are willing to pay us specifically a lot of money for something where it's competitively, um, you know, it's different than competitors in a way that customers care about. And it's something that is possible for us to deliver, right? A customer is not going to know any of that and, and they shouldn't, and you should not ask them for product advice. That's, that's your job or your leadership's job. Unless you're building a product for product managers, but probably to me, that sounds almost like, you know, be careful what you wish for because you'll be getting all of the ideas then, and there'll be designing stuff for you on the fly. So, again, be careful what you wish for. But speaking of product managers, and I guess, obviously, as you say, from your kind of history, sort of startup founders, you know, people that are listening to this, that are building products, either early products, or maybe getting a bit further on and maybe they've got product market fit, but they want to work out what to do next. Like if there was one piece of actionable advice that you could give them or take away from this based on concepts that we've been talking about, the idea that, you know, speaking to customers and speaking to users is timeless, and even in the age of AI still needs to be done. What's a piece of advice you'd give them to kind of inspire them to do something either double down on something or to think about something differently? Honestly, it's talk to people, talk to your customers, talk to your prospects. There is such incredible insight that comes out of really understanding people's problems and how you might solve them and how your product is currently solving them. And there is simply no replacement for that. You know, I think AI is incredible. It can speed up our research. It can also speed up development, right? I mean, just speaking from my own company alone, all of our developers using cloud code for the last year, that has really sped up the pace of development in our product. It's been incredible. But the thing is, is it sped up the product and our trajectory because we already had product vision in place. We already knew what our product strategy was. We knew what we were good at. We knew our competitive advantages. We knew what our competitive advantages weren't. We already had that strategy and vision in place. And so once we got AI, we were able to accelerate our development, right? We already had that understanding of customers feeding into the product vision. That sped things up. But I think a lot of organizations, quite frankly, that don't understand their customers or their prospects, or they're not all on the same page about it, or they don't have a coherent product strategy, and it's just shiny ball syndrome, and the priority changes every week. And then those organizations are also adding in AI to the process, especially on the engineering side. And the problem is, when you add AI into an engineering organization, into a company that does not have a clear product vision, a coherent product strategy, and a deep understanding of the customers, you're not going to accelerate your product development. You are simply going to accelerate running in place or running in circles. It's not going to solve the underlying problem of a lack of product strategy and a lack of customer understanding. And so that comes out of understanding your customers. And while there are ways to backfill data that you've already done in sort of reasonably accurate ways, there is no replacement for going out and talking to real people, going out and surveying real people. That is what has to drive your product development through the entire organization. And you need to make sure that everybody, from leadership, to the engineers, to customer support, everybody understands the core problems that you're solving for users, why you're solving them, why you as an organization are the ones to solve them, why competitors don't solve them as well, why all of this continues your path as a company is a way to profitability or higher revenues, whatever your goals are, right? AI and AI development is not going to give you that clarity. And I think so it's advice, but it's sort of don't make the mistake that simply having your developers use cloud code is going to improve your product. Because if they don't understand customers, and you don't have a clear vision for them to execute on, you're just going to stand in place faster. Well, that's a sobering and chilling insight for all the people out there that are just hoping that if they just put as many vibe coded applications up that one of them will hit. And yeah, maybe one of them will, but it feels like a lot of them won't. And you probably get better odds at the casino, right? But if one of those people wants to come and find you after this, check out the latest edition of your book or chat to you about user research in general, or maybe come and start using your company's offerings, or maybe even start an argument about apocryphal thought leader quotes, where can they come and find you? Yeah. So actually, first, I want to say that I'm not against vibe coding, I have vibe coded things myself. And when I was a product manager, it would have been so amazing to vibe code a prototype rather than getting the green light for resources, even to get the okay to ask the user researcher to create a prototype for us, I had to get a green light for it. It would have been so amazing if I could vibe code things. So anyway, so people can find me online. So I am on LinkedIn and Blue Sky mostly. MJW Hansen is my handle. I'm not really on Twitter anymore. But I am also going to be at your meetup in January. In January in London. I'm so excited. I'm going to be there anyway, to meet with some founders. Oh, no, come on, you're supposed to make it sound like you're coming. I'm going there actually to do the meetup. And then I just arranged some meetings the day after just to give me an excuse. Yeah, exactly. It works really, really nicely. Yeah, I'm super excited for that. So I'll bring some new some copies of my new edition of my book Deploy Empathy, which you can also buy at deployempathy.com. But in addition to the book, I also run a geocoding company called Geocodeo. Currently, we're in the US and Canada. We're actually gonna be expanding to the UK next year. Yeah, so pretty exciting stuff. But yeah, and I also have a newsletter. I don't admittedly write it as much as I should. But I will have something coming out before the end of the year that is a product prioritization framework. Again, this is kind of my running track of, you know, product basically for non product people, especially developer founders, though, who knows, maybe it will surprise me and product people will find it useful. Stranger things have happened. Obviously, link all that into the show notes and try and get this episode out in time for people to come pick that up. But I'm very excited to be going to the pub. It's always fun to get people together. So I'll make sure to link that into the show notes as well. And hopefully we get a few old faces, old friendly faces and new people coming along to say hello and chat and chat to each other and sort of build that sense of community, which I think is so important for product people, wherever they are, especially in London, where I am. I've heard incredible things about the product community in London. And so I'm so excited to get to experience that for myself. It's January 15th, right? Is it Thursday? Off the top of my head, yes, I'd have to look it up. But yeah, I'm really excited then. I'm not very good at remembering things unless they're directly in front of me. So I'll make sure to put the right date into the show notes. But obviously, we've got a little bit of time before then. So all I can do is wish you the best of luck on your continuing journeys. Hope you get to talk to a lot more customers. And as for now, thanks for taking the time. Thank you so much for having me. It was really fun. Transcribed by https://otter.ai