Amy Jiménez Márquez is Vice President of Experience Design at Zillow. Previously, she led design teams at Amazon and Compass. In this conversation, we focus on the role of information architecture in managing digital experiences at scale, with a particular focus on research and modeling.

Show notes

Show notes include Amazon affiliate links. We get a small commission for purchases made through these links.

If you're enjoying the show, please rate or review us in Apple's podcast directory:

This episode's transcript was produced by an AI. If you notice any errors, please get in touch.


Jorge: Amy, welcome to the show.

Amy: Hi Jorge. Nice to see you.

Jorge: It’s very nice to see you. We’ve been friends for a long time, and it’s a pleasure always talking with you. Folks who are listening in might not know about you and your background. How would you introduce yourself to them?

About Amy

Amy: Well, I would say, “Hi!” I’m Amy Jimenez Marquez. I am currently the Vice President of Experience Design at Zillow. I’ve been in the design industry for over 20 years. I’ve gone through everything from UX design to product design… like I started as a webmaster! You know, UX design, product design, voice design, personality design, information architecture, kind of spanning many areas of design. And I am a perpetual student and don’t ever want to stop learning. So one of the things I love about my current role is that it’s in real estate, which is the most complex domain of any domain I’ve ever encountered. And we can talk more about that if you’d like, but it’s mind-blowingly complex and rife with information architecture issues.

Jorge: I definitely want to get into that, but you mentioned a couple of design “sub-disciplines,” if I might call them that, that might have raised a few eyebrows. You said voice design and personality design. Can you pinch and zoom on those?

Amy: Sure, sure. For several years I led the Alexa Personality Design team. And that was the team that shaped who that voice agent is as a personified agent, and also any of the celebrity experiences that came along with that. So Samuel L. Jackson — working on: who his virtual self was versus who his real self was and differentiating between what people believe him to be and what he really is. And so, delving into the psychology of personality was something very rich, very fascinating.

How do you build a fleshed-out personality from something that’s an “it”? And how do you build a separate, distinct personality from someone who already exists? Those were fascinating challenges to tackle. As well as building the tooling that enabled the conversational designers to put those experiences in without additional engineering. So we did a lot of backend tooling design as well, and that was fascinating — learning in voice design that when you strip the visuals away from design and you’re left with voice, what you have is context, content, and IA, and it’s really paired down to that, and it’s fascinating.

Jorge: What do you mean by IA in that triad [of] content, context, and IA?

Amy: Let me describe all three of them. Who is the person? Where are they? What are they doing when they say, “tell me a fart joke!” Usually a kid, but oftentimes… you know, you’d be surprised! There are a lot of adults asking for that. What do they expect in return? And what potential conversations might they want to follow up with? So by IA, I mean, where could this go beyond that?

Jorge: That’s fascinating. So it sounds like you’re dealing with two personalities; the psychology of two personalities: both the personality of the smart cylinder — to give it a name — and the personality of the person who is interacting with a thing, right?

Amy: Right. And we kind of had to look at it as cohorts because there are very different kinds of people who approach the cylinder, as you say. It’s probably a better word for me to use since I haven’t muted mine over here.

Real estate IA challenges

Jorge: You mentioned that real estate is mind-blowingly complex and that it has these rich information architecture challenges. And I want to circle back to that and ask what you mean by that and what kind of challenges you are encountering.

Amy: Have you ever bought a home, Jorge?

Jorge: I have bought a home, yeah.

Amy: Was it a pleasant experience?

Jorge: There are a lot of forms to fill out, a lot of processes, a lot of actors involved, a lot of agencies, institutions.

Amy: Right. So imagine you have a lot of different tools that agents are using. You’ve got agents, listing coordinators, transaction coordinators, and advisors. You’ve got title and escrow coordinators. You’ve got all of these people on the back end that oftentimes a home buyer doesn’t really know about and really shouldn’t. Their main interface should be with their agents.

But there’s all of this going on in the background, and there are all of these disparate tools spread across this ecosystem that these poor agents and their team members have to dive in and out of, in and out of, in and out of, and it’s broken.

Jorge: I am assuming that your scope is within the US. Is that correct, or is it international?

Amy: It’s within the US. And my scope specifically is on the agent’s side of things. So all of the agent-facing tooling is what my team works on.

Jorge: Well, that was my next question, which is, my experience with Zillow has been primarily as somebody who’s used it to do things like look up the value of home prices — look up what is available in certain areas, that sort of thing — as somebody who would be the client end of the thing. The person who would be buying or renting or whatever, right? And my perception is that systems like Zillow allow people like myself to either buy or rent or sell — list for selling — homes. It enables this marketplace, right? But it sounds to me like there’s more to it than that because you’re talking about escrow; you’re talking about all these processes. What does the system cover?

Managing IA at Zillow

Amy: So right now, Zillow has a lot of different services that the agents can use to enable their clients to have a smoother experience buying and selling the home. I can’t go into detail about all of it. I’m like two months into this new job, and there’s a lot of stuff in development. But we try to make their jobs easier through easy-to-use tooling and through kind of an end-to-end system that helps them work with their clients more smoothly through this really, really complex process, regardless of what region they’re in.

Jorge: Yeah, and I don’t expect that we’re going to get into details. I guess the question was just to correct my own understanding, which appears to be somewhat superficial, right? I thought it was like, “Well, it’s mostly like a marketplace where real estate homes, apartments, properties, whatever are listed for people who are in the market.” But it sounds like it might be deeper than that, right?

Amy: Yeah. There’s a program called Zillow Premier Agent that’s out there in public where agents kind of opt-in to working with Zillow and Zillow tooling to give their clients a smoother experience, a better experience. And one of the things on the client side that they want to do is not just to help them search or post things but to help them dream and imagine the possibilities.

Jorge: Ooh, that’s really interesting. How would the system enable them to do that?

Amy: Well, if you don’t know what you can afford, you don’t know what you can dream of. So helping them understand affordability, helping them understand savings, helping people prepare to buy a home, and be better at understanding how to buy a home.

Jorge: Got it. And that might be part of the context triad that you mentioned, right? Content, context, and then the information architecture.

Principles and roles

Jorge: You’re a VP, which in my mind means that you have teams of folks working in your team. And what you’re describing sounds like a pretty complex domain. And I’m wondering about the dynamic between your perspective as a leader who has a systemic understanding of what needs doing and then the individual contributor designers in your team who need to understand these things. How does that work?

Amy: So one, and I don’t want to make this preachy about Zillow. But one of their principles is “turn on the lights.” And that’s about transparency. And it’s one of my favorite principles because there are so many companies that aren’t transparent. So I try to make it clear through every level of my organization, here’s why we’re doing what we’re doing. Because when people know why they’re doing what they’re doing, it’s much more important. They know they’re making a difference in people’s lives. And bringing that rationale to them is really, really important from the get-go.

Jorge: Okay, so transparency is one. But some of the things that you’re describing… like, you mentioned the word taxonomy, which might be a term that, for those of us who know information architecture and have been involved in that for a long time, it seems familiar. But it might be harder for other folks to grok.

Amy: So, I’ve been working with the content design team and talking with them as well about, like, “Hey, do you guys need a taxonomist? It kinda seems like you do!” And those are conversations where their eyes light up, and they get excited, and you know, they’re looking for somebody to champion bringing in a taxonomist. And I totally understand that. And on the agent side, it’s really something because there are so many different words used for so many different things in that specific domain. It’s something that’s really important.

Jorge: You’re working with content designers. What are other design sub-disciplines involved?

Amy: It’s exciting. So we’ve got four different kinds of research. We’ve got UX research, population science research, and behavioral research. And we have insights teams. And they all report through the same org. Let’s triangulate, but with four. Quadrangulate? You know, the insights really help make rich discoveries about where we should be going, what we should be doing and focusing on.

Jorge: And that’s a research organization. What is the output of the work that they do, and how does it inform these structures that you were talking about earlier?

Amy: So they’re within the design organization. So they’re very accessible, and the output is reports and recommendations. Oftentimes you’ll get like a really great deck that has verbatims pulled out from interviews with clients or with agents and just get some really rich insights. And you get trends, too, like population science really helps you look at trends over time. And behavioral science helps us understand, like, how we shift behaviors.

Lessons from Alexa

Jorge: This seems to connect — at least it does in my mind — with what you were talking about earlier with regard to the design of personalities and having to understand the psychology of these things. And I’m wondering how, if any, the work that you did in designing personality for the cylinders might inform the work that you’re talking about now in working with your teams.

Amy: It has helped a lot, not just in what I do professionally as a design person, but as a leader of people. I have basically inflicted on every team since the cylinder… you know, let’s do our MBTI, let’s do any personality test we can; Strength Finders. And look at — across the team — how we relate with each other and how we might conflict with each other and why we might conflict with each other and understand each other better. And that’s something Zillow actually does from the get-go. You take an insights kind of course where you learn more about yourself and the way that you are at work. So I find that to be like this… the culture here really fits in with the way that I like to lead anyway.

Jorge: The acronym MBTI. That’s…

Amy: Sorry! Myers Briggs Test Inventory. Yeah.

Jorge: And the idea there… and I’ve worked in teams where we’ve done this, where the idea is that you try to gain an understanding of your own personality: how you process information, how you share, how you work best, what kind of other personalities you’re compatible with, and then also get an inventory of the personalities of the people in your team so that you can develop better working dynamics.

Amy: Yeah, and it’s something where it’s fully transferable. Like we get a big Mural board, and we put everybody’s information on there. And then I subscribe to an online service called where I can pull up… Like, I’m an ENFJ, and I can pull up ENFJ and INTP and see “here’s how we get along best, here’s how we may conflict, here’s how we, you know, might react to each other at work.” It’s really very informative. And so… and you know, I let people know, this is directional; this is not the blueprint of who you are; this is high-level directional information.

But it can help you understand when somebody reacts in a way that you don’t expect to a situation, you can go back and say, “Oh, it’s because I approached them with a really urgent situation without a lot of information and they’re extremely conscientious, and that throws them off.”

Jorge: It’s funny because it strikes me as an information architecture challenge of sorts, right? Like you have these attributes. From what you just said, the thought that it brought to my mind is that we all require different levels of information to develop enough context to know why we are being asked to do something, right? Like you talked about transparency. Some other people just go on… I don’t know; they go on gut or just take for granted aspects of the context. And they don’t need as much, right?

Amy: Well, some people are just like, “point me in the direction, and I’ll go that way.”

Making IA more tangible

Jorge: Right. That’s fascinating. Much of this — by “this,” I mean the work of doing information architecture — the structural aspects of these systems — can be very abstract, right? Like speaking of personalities, some people deal better with abstractions than others. But you’re saying yes, and I can see you smiling, which I think… I feel like I might have touched a nerve! So I’m just going to stop and see if you have anything to say to that.

Amy: Yeah. So here’s something that I was kind of on the fence about OOUX — object-oriented UX — until I started using it to make IA more tangible to the people that I work with who don’t really get ambiguous things. And I think what Sophia Prater has done with OOUX is actually quite marvelous in that she’s made it accessible to people to whom IA is ambiguous.

Jorge: In the spirit of… it’s not transparency; it’s just pointing people to the fact that Sophia has been on the show and talked about object-oriented UX, and folks who want to learn more might want to circle back to that episode. Sophia has been on the show and talked about object-oriented UX, and folks who want to learn more might want to circle back to that episode. But I’m curious as to how, in your experience, it does that. Like, what about object-oriented UX helps do this?

Amy: Sure. So the most difficult thing for me to teach… I teach a course in IA. The most difficult thing for me to teach is concept modeling. People just don’t get it. But when you break it down into object identification, like where you separate nouns and verbs, there’s an exercise that Sophia has on her website — breaking things down into nouns and verbs and how the nouns relate to each other. It’s so much more tangible, and it’s so much easier to get people to start understanding what a concept model is.

Jorge: I have not gone into object-oriented UX as deeply as you have, but the affinities between that and conceptual modeling stood out to me. I recently read Mike Atherton and Carrie Hanes’ book Designing Connected Content, which came out years ago, and I should have read it a long time ago, but I only circled back…

Amy: Has it been years, really?

Jorge: Yeah!

Amy: It doesn’t feel like it’s been years.

Jorge: I think it came out at the end of 2018, so it’s been a few years. And that book mostly deals with domain modeling, right? This notion of nouns and verbs and making diagrams to make these things tangible and thinking through the attributes of the nouns. And before that… I mean, there are other resources. One that I always bring up is Jeff Johnson and Austin Henderson’s Concept Modeling Core To Good Design — which I think is the name of the book.

But these ideas, that you can take a domain and think through the nouns that people have to deal with and how those nouns might relate to each other, like the “verb” connections between them, does seem to be core to the structural work that needs to go into these systems. And what I hear you say there, and I just want to reflect it back to you to see if I’m getting it right, is that the value in object-oriented UX is that leveraging our understanding of objects suddenly makes it more manipulable, more tangible. Is that fair?

Amy: Yes! So she’s making something very amorphous, like conceptual modeling, into something tangible, like nouns and verbs and attributes.

The modeling process

Jorge: I’m wondering if you could describe for our listeners how you and your team are using that. Like, what are the mechanics of doing that? Obviously, without disclosing anything confidential, but a lot of us are working remotely, right? So I’ve done this type of exercise in a room with people using a whiteboard. Not exactly Sophia’s approach, but, you know, you use sticky notes to represent concepts, and you draw lines between them. How do you do it when you have to work remotely?

Amy: So there are workshops, of course, and some of them are in person, but if there’s one person remote, we all work remotely. So if one person has to be remote, we’ll all work in Mural. And it’s really not that difficult. You have one moderator. You have one person leading everything, and you’re led through exercises.

And the way that we’re approaching this is there are certain aspects of certain software that we’re just kind of completely pulling apart and trying to put back together because maybe there are flaws in the IA or things we didn’t understand or opportunities we didn’t understand. And another thing about the whole exercise is understanding conceptual models, understanding connections that you missed before.

Jorge: I’m curious about the role that research plays in informing those models. You mentioned that there are four kinds of research teams in the organization, and they’re doing all sorts of things. How do you take insights that come from research and have them inform an abstract domain model of the sort you’re describing?

Amy: Well, what we do is we make sure that we have someone from research in the room at all times when we’re having… If you’re having a design workshop, you don’t do it without research. At least we don’t do it without research. And preferably at least one product partner, if not more so that you can vet your own biases and understand what you’re missing.

Because our researchers will help us understand what we’re missing in the picture, like, perhaps there’s a word we’re using that they’ll let us know, “Hey, clients don’t use that word at all. That’s not something they relate to at. So that noun you’re using, there is not a noun a client would use.”

Jorge: So they’re there kind of as a gravity check as they say, right? So like, “let’s keep it real.”

Amy: And they have so much context in their brains of all the research that they’ve conducted and all the research they’ve reviewed that not having them there is a huge miss.

Jorge: What about the emergence of key distinctions? And I’m saying this kind of cautiously because even as I formulated the thought, it sounded abstract. But I see one of the advantages of working at this level of abstraction — the modeling stage — is identifying the key distinctions that users will encounter, right?

The example I always use is when working in financial services — a banking website or something like that — the user might have something that they call an account, which is where they expect to store their money. But then the system might have something called an account, which is the profile you use to log into the system. And if you use the same word — account — for both, that might be ambiguous, right? So there’s a distinction that needs to be drawn there, and I’m wondering how those distinctions emerge during these modeling exercises.

Amy: That’s a bit deep. I’m trying to think of an example. That’s a great question! I know that we’ve encountered things like this before in our workshops. I just can’t think of a specific example. But in that case, we look at, oftentimes, especially in real estate, it’s a difference between what a client is calling something and what an agent is calling something. And in that case, we just have to make sure that we’re surfacing the right word to the right audience.

Jorge: Part of what’s implied in what you’re saying here is that it’s up to the design team. I’m assuming that it’s the design team that’s facilitating this exercise, and it’s up to the design team to make these ambiguities visible and then help adjudicate the direction that you’re going to go into. You have to choose whether you’re going to use the client term or the agent term. Someone has to choose, right?

Amy: Well, It has to depend on who the audience is. Like, I mean, my argument is you should have a content management system that can say, “Oh, hey! A client’s looking at this; we’re going to surface the ‘client’ word.” Or, “Oh! Hey, an agent’s looking at this; we’re going to surface the ‘agent’ word.”

Jorge: All right, so that’s yet another contextual layer here, right?

Amy: Yeah! In a perfect world, that’s what happens. We don’t live in a perfect world!

Grokking the technical underpinnings

Jorge: Right. Well, and you mentioned earlier… I think you said it kind of in passing when you were introducing yourself that there was something about the understanding of the technical aspects that underpinned the experience, right? And you said that early on in your career, you were a webmaster, which, at least in my book, entails some degree of rolling up your sleeves and getting into the implementation of stuff.

Amy: Yeah. Self-taught HTML. Self-taught CSS and JavaScript in the late nineties. I mean, I got into all of this through the code. I didn’t get into it through the visual side.

Jorge: What I’m hearing here with this example of the agent versus client term and leveraging the CMS is to offer up the right language to the right audience implies that the CMS has some degree of personalization; the capability to do that. To recognize who I’m showing this to and to switch terms if needed.

Amy: Right. And this is in a fictional use case.

Jorge: Yeah. It’s hypothetical. But the reason I wanted to call it out is that it implies that this is a process that cannot happen properly. Or at least the implications cannot be played out properly in the absence of some understanding of the technologies that are going to be used to implement the model. Right?

Amy: Absolutely! As any person in any level of design, if you don’t understand the systems you’re working on, their limitations, and their capabilities, how do you know what you can design? And how do you know where to push?

Jorge: Right. So it’s the capabilities and constraints of the system; you have to grok those.

Amy: Absolutely. And you have to also be informed enough or have somebody in your design team informed enough to call BS on, “Oh, we can’t do that. That’ll be a million dollars.” To be able to say, “Actually, you can just do that on the front end with JavaScript. Let’s do this workaround.”

Jorge: I’m wondering about the other end of the process. Modeling, to me, is one of the most important phases in design, and I think it’s one of the least acknowledged. When you go on social networks and see discussions about UX design, most of what people are talking about seems to be about the user interface level.

Amy: Yeah.

Jorge: They’re talking screens. And what you’re talking about here is the underlying structural constructs that are going to support those experiences.

Models to screens

Jorge: We’ve talked about translating research into the model, and I’m wondering about the other end of it. So, what do you do with the model? Because eventually, you have to get to the screens. How does that translation happen for you?

Amy: In using the object-oriented UX exercise, when you get the high-level fundamental nouns, basically — your objects and their attributes — you start to understand how you could start to build out an information architecture for this environment you’re building. So you start building the bones. You start with the framing of the end-to-end experience that you’re building.

And I think too often we’re not building end-to-end experiences. We’re building pages, or we’re building micro-interactions, and then you completely lose the sense of the flow when you do that. So, I’m really happy with the teams that we’re working with that are kind of starting from scratch and trying to build this end-to-end system, like to see how the entire forest is laid out before they start looking at the trees.

Jorge: So you point them to the Mural board? Is that the gist of it?

Amy: That’s the starting point. There’s a lot of workshopping that has to come after that. There’s a lot of collective work that has to happen together. And you refine and rework, and eventually, you get into IA flows, and then you get into your traditional modeling out of what this interaction is going to look like.

Jorge: So it sounds like there’s an intermediate step between modeling and screens, right? Like where there’s playing out of the implications of the model.

Amy: Yeah. It’s almost like process modeling, process mapping. It’s the stuff that I find exciting that other people find boring. And honestly, when I was new at design, I was terrified of this side of IA. I thought it was just so complex. And now I just can’t do without it. I am flabbergasted when people aren’t doing it like that.

Jorge: Yeah. And my concern is that many people aren’t doing it like that, which is why I was so keen to talk with you about these issues; I know that you are a design leader who understands the importance of doing this, and you’ve been working like this for a long time.

The impact on personal work

Jorge: So what impact, if any, has working in this way had on your own work; how you manage your own personal information?

Amy: Oh! That’s interesting.

Jorge: Because what we’re talking about here is designing a large system or something like that. But I think that thinking in this way must influence how you frame problems in general.

Amy: Yeah, it helps me a lot, especially with design organization. Like rather than looking at the silos of my team, I look at every individual on the team and say, “Okay, is everybody in the right place? Are they in the place where they’re going to be the most effective? Are they in the place that actually speaks to what they enjoy doing, what they get fulfillment out of doing?”

And if not, let’s talk about this, and let’s look at the team as a whole and find different places for them to be, if this is not where they should be. So, rather than looking at this as a defined system that can’t move, look at it as something where they have the opportunity to have a say in their destiny in this work that they’re doing.

Jorge: I’m going to reflect it back to you because I think it’s an important insight. It sounds like what both have in common — both the modeling exercise and what you’re talking about here — is honing this ability to work at granular levels — you know, the screen level, the process level, the level of an individual contributor, and their role and what function they’re providing to the team as a whole — and then also being able to zoom out to take in the big picture, to understand the system as a whole, to see the forest and not just the trees.

Amy: That’s something I enjoy doing. Yeah, I enjoy doing that.

Jorge: I think it’s really interesting to think about how modeling might be a way of seeing the forest and understanding how the trees relate to each other. And it sounds like it’s applicable to more than just the design work; it’s also applicable to designing the thing that will design the thing.

Amy: Yeah, absolutely. Absolutely. And it makes people happier. I mean, you can expand it out to the entire organization where because of the way organizations are designed, they have applications named after that side of the organization, and they get married to that. And when you come in and say, “Hey, I hate to tell you this, but this is the wrong application,” it blows people’s minds because you are looking at everything end-to-end with new eyes, and you have to tell them, “Hey, this thing you named your team after shouldn’t be. It should be connected to all of these other things, and they should be one cohesive thing.”

As somebody who comes in at a VP level, It’s my job to influence that kind of thinking. And it’s a little scary and sometimes intimidating because people fall in love with their own ideas and what they own and the ownership of that thing. And when you tell them that thing really isn’t quite right and should be stepping back and looking at it, it should really be this end-to-end cohesive thing that’s much bigger than this one, and then they’re going to have to share ownership of that. It’s a very delicate balance to bring people around to that.

Jorge: One of the things that’s so exciting about hearing you say that is that it points out that one of the advantages of thinking with the mindset or the framing of information architecture is that we become much more conscious about labels. And the fact that labels are just a label, they’re not the thing. And what I’m hearing you say there is that people can become fixated with labels and distinctions. Like you’re saying, it’s like, that’s the name of the team! Of course, we do it…

Amy: Well, yeah. They make t-shirts! They put little logos about what they name the team. And it’s named after this app that actually isn’t the right thing.

Jorge: So, the big takeaway there is, think about what the thing is and what role it plays and maybe suspend your attachment to labels and work from the overall fit in the system and what value it provides.

Amy: Absolutely. At the company I worked for prior to Zillow, they changed the label of my team so many times that our Slack channel was just called Team Amy. And I said, “This is not about ego; this is about: I’m done changing the name of this channel. It will change when you have a new leader.”

Jorge: I was going to say they’re in trouble now because it can no longer be called team Amy, right?

Amy: Right.

Jorge: Well, Amy, this has been fantastic. Thank you so much for sharing your experience and know-how with us. I expect that people are going to want to find out more. Where can they get in touch with you?

Amy: Oh, I’m on Twitter @amymarquez, which tells you my age; basically it’s my name. And I’m on LinkedIn as well. Amy Jimenez Marquez. Easy to find.

Jorge: Great. And we didn’t get into this, but you’re also the publisher of Boxes and Arrows.

Amy: Right! right. As you know, our mutual friend, Christina Woodkte, actually founded that and handed it over to me several years ago. And me, along with Alesha Arp, we’re running… Alesha’s running the social media arm of it for me. We’ve been publishing it for the past couple of years. And right now, we’re on hiatus until January. We both are a bit burnt out. So we aren’t publishing anything new until January, but we are very interested if anybody wants to reach out to if you have any ideas for posts or articles; happy to talk to you about it and get them ready for the new year.

Jorge: And I would like to encourage everyone to do that because Boxes and Arrows is a very valuable resource, and it has been for a long time. And I’m very excited to know that it is in good hands. And it has stewards who care for it and want to keep it growing. So I do encourage folks to get in touch with you.

Amy: I appreciate that.

Jorge: All right, Amy, thank you so much for being with us.

Amy: Thank you, Jorge! It was my pleasure.