Checking session availability…
Hang tight while we load the latest updates.
For expert users, the job is not to make your UI sticky, it is to make it disappear. Join Benjamin Humphrey as he interviews Fahad Osmani. They'll break down how platform context changes everything, the hiring profile for designers, the role of domain expertise, and the metric shift from clicks to completion and time to value.
- Diagnostic, five questions to confirm you are in platform territory
- Metrics that matter, how to set up time to value and task completion the right way
- Interface strategy, cockpit patterns, progressive exposure, and safe defaults for expert work
Platform UX: The Metric Shift
Fahad Osmani, Benjamin Humphrey at UXDX USA. Video: https://youtu.be/g_-iiKZUQ-Q
Readable transcript: edited from the recording's captions for readability (fillers and false starts removed, punctuation and section headings added). Wording is the speaker's own. Timestamps are positions in the video. Names marked [?] could not be verified against the audio.
How platform design differs from product design
[00:00:08] Benjamin: Hey everyone. Well, bright lights. Long day, huh? So, last session. Very excited to get into it. Joined with Fahad from Capital One. I'm Benjamin from Dovetail. We're talking about platform design. So we're not talking about AI and AI agents, we're going to talk about old school product design, UX design. So I hope you guys like that topic. I guess my first question is, we're talking about platform design. Talk to me about how that's different to product app design. What are the key differences?
[00:00:42] Fahad: Yeah, sure. You and I talked about this a little bit earlier, but it's one of my favorite topics to think about, both what's similar and what's different about those two things. One of the things I always tell people whenever they join a platform design team, or rather an enterprise platform, is your typical buyer-user dynamics are totally disconnected. So you can no longer rely on the signal that someone has paid money for your product to also equate it with, well, my product must be really valuable for its users, right? Quite often it's a requirement, and so therefore you can't pay attention as closely to that signal. That's one of the major differences between consumer, or feature-led design, and enterprise platform design.
[00:01:24] Fahad: And the other one that I really think quite differently about is, in platform design you're often thinking about designing for the workflow, or designing for the outcome, rather than designing for the feature. So we've introduced this new capability into the product. I don't really care if people are using that specific feature, or how often they're doing it, or how they discover it. What I really care about is, are they getting to done for the thing that the enterprise needs them to do.
[00:01:53] Fahad: And then I think the last thing I always point out to any member of my team that hasn't done enterprise design before, or enterprise platform design, is we're much more concerned in a way about the floor rather than the ceiling. So we often operate in a world in which there are either regulatory requirements or systems requirements, and the company has decided at some level that standardization is the key to success. Therefore getting to a standardized way of doing things is often going to yield a lot of value, both in terms of economic savings, so you're not reinventing the same thing over and over again, but also that you can actually start gathering data that you can start comparing against similar types of data. And so getting to a standardized way of doing things is often, at the macro level, more valuable than a single division or a single organization or a single line of business doing something in a super delightful way. But for its internal users, the entire enterprise benefits a lot more with that standardization.
[00:02:58] Fahad: So those are some of the things I think differently about. But obviously the things that are still the same is, there's still a user at the center of it. The fact that you're still thinking from the user's point of view is very important. The fact that bringing clarity to those users is still the right answer. You're not going to try and shove more complexity than is necessary at them. Although in some cases the expertise of the user, and the platform that they're interacting with, can often lead you to a very different answer about the type of information density or clarity that they're presented with, where the typical way of thinking about, well, let's make it simpler or cleaner, isn't always the right answer. Quite often the user needs a lot of information, because they're the expert in that field, in order to make that decision. So those are the things that I think are different and the same.
[00:03:46] Benjamin: Yeah. I used to work on a product called Jira. If anybody's heard of Jira. As a designer, I'm sorry. I'm sorry. Which is obviously a very famous platform. One of the things when I joined Atlassian we ran into was, everyone was saying, "Hey, everyone loves Jira." It turns out we were just speaking to the administrators, and not the end users. We went and had a look at the end users and how they used the product and did a contextual inquiry, visited them, watched them, and they hated it, obviously. So, I did try. I did try.
[00:04:15] Fahad: It's a very common story across platforms. You think you're the main character in your users' lives, but it's usually administrators.
[00:04:22] Benjamin: Exactly. And I think Jira is a great example of workflow oriented, obviously, and you have to balance the needs of the buyer, the requirements they have, and the end users. But I want to get into a kind of meaty topic, which is architectural design, or sort of conceptual model design. One of the things with platforms is they're extensible. Typically you have an add-on ecosystem, a plug-in ecosystem, or even just an API. A lot of the time these days we're designing APIs. As product designers we're thinking through schema design, MCP tools, tool descriptions, these kinds of things. Sorry, I said it wasn't going to be about AI.
[00:05:00] Benjamin: And when you're doing that, you need to think through the conceptual model and architecture, because a lot of the time it's a one-way door. If you have a platform that has ecosystem vendors, we had a lot of those at Atlassian, they build on top of your platform, you can't just change it easily. So, designing the architecture and the conceptual model, how do you approach that? What sort of ways are you validating this? How do you make decisions also around the terminology?
Naming things, and the cost of a bad word
[00:05:30] Benjamin: How do you decide whether to call something an issue or a task or a project or a space or an insight or a doc?
[00:05:36] Fahad: Yeah, all these terms are hitting deep, because if anyone in the audience has worked on an enterprise platform before, you know, there's like 10 different definitions of what's a project, what's a task, what's a workflow, what's a message. And usually all of these things are overloaded depending on which year this thing got built.
[00:05:53] Benjamin: Confluence went for space.
[00:05:54] Fahad: Yes, exactly.
[00:05:55] Benjamin: Which is quite controversial, because what is a space? It's just...
[00:05:58] Fahad: Is it different than a workspace?
[00:05:59] Benjamin: It's just generic. It means nothing.
[00:06:02] Fahad: And then you have namespaces. So, lots of things to pay attention to. What I've found that works reasonably well across different contexts, and I've worked pretty much my entire career on enterprise platforms meant for either technical builders or people that are experts in their field, is that a combination works really well: a little bit of top-down architectural thinking, just thinking through first what's the entire landscape, what are we trying to prepare for, what are the macro issues at play, but then also allowing for a healthy level of organic invention.
[00:06:35] Fahad: Some of the most innovative things that I think have made their way into platforms that I've been fortunate enough to design for have in fact been invented out of shadow IT organizations in various lines of business or divisions. So it's definitely not the preferred way to find exactly what your users need, unless you're really close to your users. What typically ends up happening is that if you're in a division or line of business that has to cater to those users closely, you're going to get really, really surgically precise about what works for the user and what gets to done for that particular outcome.
[00:07:14] Fahad: And we generally go around trying to gather up those ideas and standardize them into an enterprise platform, because typically you run into, innovation is super high within the line of business but standardization carries a really, really high cost, and then you haven't taken into account the run-the-engine cost. I'll take an example from a prior role that I had. What if your cyber security division has to pay for task management or issue management across your observability and your data analytics platform? Well, that's not a great way to do it. Let's have a standard way to do task management.
[00:07:51] Benjamin: Like an efficiency thing, right?
[00:07:53] Fahad: Yeah, exactly. So some of it you're going around gathering up all the interesting ideas and eventually sort of pulling them back in on a regular seasonal basis. And for some of it, I don't think you have to wait for that innovation to happen. You can think through certain things from the beginning, like what is the one plus one equals something greater.
[00:08:11] Benjamin: I mean, you're describing, I guess, systems thinking, abstraction. You mentioned a little earlier, how do you put something out there and see how people use it. So this idea of emergent behavior.
[00:08:24] Fahad: Yeah.
[00:08:24] Benjamin: One of the challenges I found in my career is naming things. I'm sure people are challenged with naming things. I know engineers are challenged with naming things, and finding the right term which is not too specific that it alienates a whole cohort of users, while also not being so generic that it means nothing. In our product we have a feature that we used to call notes and we renamed it to data. This is a fairly generic kind of thing, a note could be anything, could be a transcript, could be raw notes, but it was specific enough that it wasn't really alienating. However, we didn't call it something like sessions or studies, which would just box us into a particular type of data. And over time, as our platform expanded in use cases, we've slowly renamed certain features in the product. So now we call that data. We used to have another feature called insights that we now call docs, which is a bit more generic. This whole idea of systems thinking, abstraction and naming, that's a skill that designers, obviously very important.
[00:09:30] Benjamin: What are your thoughts on how to look for people who have those skills? Is this kind of thinking going away with AI now?
[00:09:38] Fahad: I don't think it's entirely going away, but it certainly makes it easier to figure out when you're going to have contexts that tend to run into each other. I find that naming is especially important, and especially information architecture is very important, when you can figure out the context within which individual users and teams are working. One of the interesting things about platforms for me is that you typically get multiple different kinds of users and roles working with each other in a team in order to accomplish some outcome. So there's typically some idea of a handoff or workflow that's happening within the platform.
[00:10:12] Fahad: And I think the things that really surface over time, which I hope can surface a lot more easily with a lot of these introspective, context-aware tools like AI tools, is when you can find out that a construct, a first class object, that means one thing to user one in the same platform might mean something completely different to someone else that's interacting with the same platform. But what's more interesting is when you don't even realize that these two users are in fact semantically using the same platform.
[00:10:49] Benjamin: An example of that?
[00:10:51] Fahad: Yeah, so an example of that might be, let's say you have three distinct platforms. One for managing all of your internal data within the company, right? The surface area is like the data lake for the company. And then another platform might be how you provision computational space and train models. And then another might be how do your engineers move through the CI/CD pipeline, right? There's data pipelines.
[00:11:19] Benjamin: Yeah.
[00:11:19] Fahad: There's ML pipelines, and then there's the CI/CD pipeline. Don't call something a pipeline unless you really know what this user is about to be confronted with, because the software developers, the ML engineers and our business analysts, our data analysts, are in fact using the same platform semantically, even if they might go to different destinations.
Walking the tool, and eliminating steps
[00:11:41] Benjamin: When you're doing this kind of design work, if we talk sort of tactically, what kind of things are you doing in your team, or have you done in the past, in terms of exercises, workshops, user research, that kind of helps you validate or learn more about how to design a platform?
[00:11:59] Fahad: Yeah, I think one of the other moderated sessions over here, the fireside chats, mentioned this, and I really, really like the phrase, walk the tool. It's the idea of having our design team really be able to use the end-to-end tools, to put on different hats and to actually be able to get the tasks done, rather than be able to just...
[00:12:22] Benjamin: Like dogfooding, like using it yourself, or like having a journey.
[00:12:25] Fahad: Sometimes that's easiest to do. Come up with a practical problem that can only be solved within the platform. But a lot of it is just trying to solve the kinds of problems that your users solve. So I'll give an example within the current context. A lot of our data and ML and software development platforms end up being used for things like banking decisions and credit decisions, right? Me inventing a problem that's more design specific to solve with those platforms isn't really going to impact that level of understanding. So our designers typically work with credit officers, risk officers, to understand how are you practically using these tools, and we always have as part of our onboarding a way to walk the length of the tools or the platforms that they're using.
[00:13:11] Fahad: But I think things like user research, and understanding what are the company's policies and regulatory stance around this, as well as what do your users need, can often be very, very interesting. The most interesting insights that we find sometimes are less about how the user is currently using the tool, and a lot more about, why are we asking them this question here? We should have asked them that way up front. But you only know that if you've actually walked the entire use case, as opposed to just designed the feature. So we get excited about eliminating steps in the process rather than designing a better feature.
Delight versus outcome in the enterprise
[00:13:46] Benjamin: We were talking about how interface delight is kind of going away in favor of outcome-driven design. I wrote an internal post many years ago about the consumerization of enterprise software. There was this trend that was happening where basically Atlassian was wondering, hey, why is HipChat losing to Slack, what's going on there, and GitHub's running circles around Bitbucket. One of the things we observed was basically this idea that at work, users wanted to have the same kind of delightful experiences that they have with consumer apps. And San Francisco had figured it out before Sydney had.
[00:14:24] Benjamin: However, we all liked the Slack loading, cutesy thing that was quite viral back then. It was like loading and it had these fun jokes and everything, right? Now I feel like, we spent so much time on that. Does anyone remember Dribbble?
[00:14:37] Fahad: Dribbble, three Bs.
[00:14:39] Benjamin: Man, the designers spent a lot of time on there. And so you're going from this idea of delight, and trying to do these loading states and stuff and make the user interface beautiful and delightful and aesthetic, and now you don't think that's as important, right? It's more about outcome and speed to outcome. Is that sort of...
[00:14:58] Fahad: I think it really depends on the context. So for instance, if you're talking about a platform or a tool in which a user is having to spend the majority of their day in order to get their job done, that's really where they're...
[00:15:09] Benjamin: You don't want it to be taking forever, and I don't want to look at a cutesy loading screen.
[00:15:13] Fahad: Yeah, but also I don't necessarily want it to be so functional that the user just sort of views it as a regulatory compliance exercise. So for instance, if you're spending all of your time in your IDE, right? That's where you hope a lot of developers spend some significant portion of their time, is within the IDE. I think it's worth it to meet them where they're at and acknowledge the fact that you've spent hours in the interface. And sometimes that can mean delight, or doing things with expressive design.
[00:15:41] Fahad: On the other hand, if you're dealing with a platform that has you register your repository or your data set, so that we're in good governance or that other members of your team can find it, that's not where you're spending the majority of your time. I'm much better off not designing a platform for you to do that with. I'm much better off embedding that experience within your native tool, right?
[00:16:04] Fahad: And so this is kind of what you talked about, the architectural strategy around embedding versus modularized workflows versus platform rationalization. You could have these three platforms and then realize a team actually uses all three of them simultaneously at all times. So why do we have three separate ones? Why don't we have a single one of those? So sometimes it's better to take away the surface area that the user is not spending productive time in. It's productive for the enterprise but it's not productive for the user. Any time we find an opportunity to do that, that's great. But if it's a tool that you're spending six, eight hours in, I think acknowledging user fatigue, and the sort of duration of time that you're spending with your head plugged into something, a bit of expressive design goes a long way even in the enterprise context.
[00:16:50] Benjamin: I remember, speaking of buyer requirements and end user friction or fatigue, visiting on site and watching a user fill out an issue template, and they were sitting there, every single field was mandatory. And of course they would blame us, the Jira design team. It was actually that IT administrator that had made every single field mandatory. But perhaps we should have put some funnier error copy, you know, like, try again.
[00:17:18] Fahad: Even if you have that friction, just the tool acknowledging the fact that this is painful. I'm sorry. I'm really sorry that you have to do this today, but it's forced by your IT team.
[00:17:28] Benjamin: Make sure they know who to blame.
The SaaS apocalypse question
[00:17:33] Benjamin: And so, do you think that, you mentioned at the beginning that a lot of things stay the same, but obviously we have this whole SaaS apocalypse thing right now, the platform companies like Salesforce, ServiceNow. I guess Figma is an extensible platform as well. A lot of challenges with stock price. Folks thinking that horizontal AI tooling is going to cannibalize these tools and they don't necessarily have a moat any more. Or do you feel like there's some truth to that, in that my example of the fields being mandatory, hey, an AI agent can just do that now. We're ticking the governance box, the buyer's happy, the end users happy, life's great. Or is there still some world in which we have these systems and interfaces, or are they kind of melting away? What's your perspective, I guess, on how platforms are going to change?
[00:18:26] Fahad: Well, I think for platforms the question is pretty unique, because for platforms one of the things you have to consider is the run-the-engine, the long-term ownership cost, right? So can I create a new innovative way to solve this problem within a few days? Maybe, maybe a portion of it. But there's things like migration issues to deal with, and then there's ownership issues to deal with, right?
[00:18:51] Fahad: And I also think that for most platforms, regardless of whether you're operating in a highly regulated industry, the fact that you're dealing with domain experts and the fact that you're dealing with builders carries, I think, an extra load of complexity, because we like to be very precious with our tribal knowledge, right, in these technical builder roles. So it's harder to get things out of the heads of someone who has a cron job running on their laptop from like five years ago, which is the thing that makes the build work, right? That's very hard to introspect and draw out with these tools, because you're not just talking about a user's experience, you're now talking about policy codified as code.
[00:19:28] Benjamin: Yeah. When we sell our software into organizations, I feel like most of the time the sales and CS team are kind of therapists more than salespeople, because a lot of the challenges are not necessarily software challenges. They're more like change management challenges, or bureaucratic hurdles, or what have you. So certainly, at least at my business, I've seen our designers become a lot more commercial. The interface is kind of like a solved problem in a way. It's funny that we kind of keep reinventing patterns and components and stuff, because how many times do you need to design a sidebar or modal dialogue? But they're becoming more commercially minded, and starting to solve, I guess it's almost like service design, right, and you're trying to figure out how to actually onboard customers and transition them and do complex data migrations and all of that, which is enterprise design.
[00:20:18] Fahad: Yeah. But I think what I gravitate more towards is systems thinking in general, which can lead into a lot of these same insights. It's not simply enough, I think, or at least in the domains that I'm working in, to simply map the user's journey. You also have to understand the laws of physics around how the system was built, why it was built this way, and what's currently working, what was tried before, and how did we end up where we are at.
Hiring for systems thinking
[00:20:43] Benjamin: So, I'm conscious of our time as well, because we have some questions, so I'll make this kind of a last topic for us. We're talking about systems thinking and abstraction, we're talking about information architecture design, very meaty topics that I'm sure a lot of product designers in the room, UX designers, have thought about in their careers. How do you hire people and find people with these kinds of skills? And how do you run a design team and a design culture that embraces this? Because I don't know, I've been doing design work for about 16 years, and I guess I've seen the transition from Dribbble user interface, pixel pushing, up to more meaty architectural, conceptual model problems. And we had a lot of those at Atlassian, by the way. There's a lot of complexity in the Atlassian suite of software, which I'm sure people can appreciate, even if they want to make fun of it, which is completely valid. But yeah, I'm kind of interested, how do you build the design team that can handle this kind of complex enterprise design? Talk to me about, how do you hire? What skills do you look for?
[00:21:53] Fahad: Well, I think we are at a kind of a pivot moment right now where the future is not entirely clear. So it's hard to go for specific skills or specific domains that you think will define the future. So the only tool open to us right now, and some of it's going to be kind of a leap of faith, is going back to first principles. And understanding, can you true back to the problem being solved? Can you true back to what creating an experience, or building together with someone who has expertise in engineering versus product management, what that looks like?
[00:22:28] Fahad: And in general, I try to look for cues on, are you going to get hung up on my definition of intent is a vector file, or my definition of intent is a PRD, right? That usually tells me that you're more focused on the artifact of your work and that you're less focused on what's philosophically the step to solve the problem, right? Are we defining the problem? Okay, that might look like something completely different in the future, but defining the problem is important.
[00:22:54] Benjamin: Yeah, going back to customer problems, back to customer insights. We do this a lot at work. We probably don't do a lot of external, I don't know if our design team is particularly in the zeitgeist of design, I would say. I feel like we try to solve a lot of problems from first principles, where we just talk to our customers a lot, but we also dogfood our product a hell of a lot. We use it to power a lot of our internal processes. And yeah, I'm often doing workshops, I guess, where we're just getting really back to basics and trying to figure out exactly what are the top tasks that the user is trying to accomplish. With Dovetail we're trying to enter into some new use cases, like revenue intelligence for example, and it's like, okay, let's go learn all about that particular use case, that type of customer persona, and again just approaching things from first principles in these kind of workshop sessions.
[00:23:47] Fahad: Yeah, because it's all about the context and theme development stuff, right. And I think if you start from those first principles it leads you to a lot of really interesting different surface areas for design. So for instance, what I found in a lot of the enterprise platform conversations is, we're no longer dealing specifically with the web or with native mobile interfaces. Now we're dealing with design at the terminal level. And I don't just mean command line, like how to structure a good command line prompt, or even how to structure a good API, although I think that's also an interesting surface area. But it's how to overload the terminal and to impose information architecture in the terminal, where you can now have mouse interactions, or just having some level of complexity.
[00:24:35] Benjamin: The GUIs going all the way back around, with actual mouse.
[00:24:37] Fahad: I mean, in some ways, right, because for a lot of our users they prefer to stay in the command line. So can I bring the information architecture to them where they're at, rather than, no, you have to jump out into the web-based one?
[00:24:48] Benjamin: I find our designers are really being challenged a lot these days in terms of, to your point, these different surfaces. We just launched an MCP series of connectors recently, and the whole sort of schema design for the MCP tools. We had UX designers that previously, maybe a year ago, would not be out of Figma, essentially just thinking through, well, hang on, Claude's going to orchestrate it in this way, it's going to call this tool, the description needs to be really well articulated so that it calls that tool and not this other tool. Should these even be tools, or should they be just a sort of generic, you know. And so you're kind of designing this totally different surface area. It's definitely all changing. I know we have time for questions.
[00:25:36] Fahad: Yeah.
[00:25:37] Benjamin: So we've got 10 minutes, right?
[00:25:39] Fahad: Yeah.
Q&A
[00:25:39] Host: Quick round of applause. That was great. Really good talk.
[00:25:44] Benjamin: Okay. So yeah, we have eight minutes. So to get into it, just take it from the top when you say caught it[?]. Oh god. Sometimes they're like dueling each other. Well, what is the biggest problem your team is currently trying to solve? That was, I believe, that's an old question. I mean, it could apply to ours too. But you guys want to take that one on when we get the right one?
[00:26:06] Fahad: I'm happy to take a run at that one. What I find is that the biggest challenge for our teams right now, and in general, platform teams, enterprise teams...
[00:26:18] Benjamin: Keep on, finish your thoughts.
[00:26:21] Fahad: This is probably somebody from my team pranking me.
[00:26:23] Benjamin: I have a lot of thoughts on that. Next.
[00:26:27] Fahad: Yeah. I think it's just figuring out what do we do with a lot of the surface area that we've already created in the last three to four years. How do we really pressure test? Does this need to be an interaction we need to have with the user? Can this be automated? Can this be solved with an integration? Because the cost of doing all of those things is much lower now. A lot of, I think, two or three year old questions that are suddenly relevant again, and just bringing that back into the pressure test environment is what we're currently going through. What about you?
[00:26:56] Benjamin: I would say for us, it's probably in the last 12 months been what I would describe as a culture change in the product team. So naturally with AI coding, our engineering velocity has skyrocketed. When you can build stuff in like a few days instead of a few weeks, it's really changed the nature of product management, design roles. For one, all the PMs and designers are now coding in production, which is cool.
[00:27:18] Benjamin: But what we're seeing now is this second order effect. We're shipping changes to customers so quickly that we now have this massive issue with customer discovery, change management, marketing team, product positioning material, outdated website content, and then the CSMs never know what's going on in the product, as well the support team never knows what's going on. So we're now actually trying to patch around the engineering velocity. And it's funny, because I never would have thought this problem would become like a solved problem. I guess the engineers are fast.
[00:27:50] Benjamin: Do you dudes think that AI is a bubble? What happened to it? No, what do you dudes think about this whole AI bubble thing?
[00:27:58] Benjamin: I would say it is almost certainly an economic bubble. There are some useful applications for AI. I think it's very good at coding. Makes sense because of the training data, Stack Overflow obviously, and it's a programming language, it's a language model, makes sense. I do not think it is very good at other things like design. I also think it's very verbose and not particularly useful for other applications. I think the model companies are super incentivized to increase their valuations before going public, and there's a lot of hype out there and a lot of noise. We try to, I guess, build sort of pragmatic AI features in our product, where we're balancing what's cool and hypey versus what's actually practically useful. But there's a whole heap of caveats. So yeah, I would say economically a bubble, however more useful than crypto.
[00:28:52] Fahad: Similar reflections for me. I think one of the things I've personally found these tools useful for is helping me think through a broader lens of what's actually going on in teams and in complicated workflows. Whereas previously I would have to literally walk every use case myself, I find that I can introspect and ask questions about where else am I likely to encounter this pattern, and get back interesting insights that I can then go and pressure test myself. It's actually been very useful for surfacing reuse, standardization and connection opportunities. So my team and I have found it greatly useful for that. On the economic side, I'm not going to make any predictions, but what I do find this is very similar to is in the days when we would have, I don't know, Netflix for $3. You're like, this can't last. Somebody's going to make money at the end of it, right? So we're in those days.
[00:29:49] Benjamin: I bought a bunch of SaaS stocks recently.
[00:29:52] Fahad: It's Capital One underwriting the IPO.
[00:29:56] Benjamin: So, what keeps you dudes interested in and excited about platform design, and how did you get into it? I feel like I've kind of always been doing this. Before Atlassian, I worked on a product called Delicious, which was a social bookmarking tool that Yahoo bought, or sold, actually. And then Jira, and then now Dovetail. I like this whole sort of challenge around abstraction and systems thinking. How do you design something that is horizontal and flexible enough, a piece of functionality that can be used by multiple personas and multiple use cases, without being so generic that it's not applicable? So Confluence has always struggled, I think, with being a bit too generic. We see Notion doing a relatively good job, I guess, at verticalizing use cases and using templates to do that. But yeah, this idea of abstraction, and horizontal software versus vertical software, has just been something I've always been interested in.
[00:30:52] Fahad: Yeah, I think very similar to you. I've always been interested in this, this isn't something I've found later in my career. I've been through a lot of different companies, different economic models, but I've been a software developer myself. I find it fascinating building tools and platforms for people that have heavy, heavy technical expertise. It could be software developers, could be risk analysts, credit analysts. I think technical skill and background expertise really, really fascinates me, and building tools for those experts often looks like building enterprise platforms, and so I'm always fascinated by that.
[00:31:32] Benjamin: Jumping around. Okay, so, being the head of design and the CEO, do either of you have examples of times in your career where you had to advocate for the value of design, and any examples from recently in your career now that you're more senior?
[00:31:46] Fahad: I mean, yes, there's always been a period in which advocating for the value of design has been a topic that's just lurking around the corner, and sometimes you choose to engage in it, sometimes you kind of wait for the puck to come to you. What I found more success with recently is advocating for the value of user experience as an outcome, and as a verb rather than a noun, right? So instead of figuring out what role everyone's in and what role will be called if the tools end up changing things, I think it's much easier to draw a value connection between, this works really well as an experience for our users in the market, or for the enterprise as a whole, where you can drive behavior to affect horizontal change, standardization change, you can gain a lot of value out of it. I always find less success, and therefore I've lost a bit of my appetite for, advocating for design as a role or a competency, and much more as an activity, as an outcome. I found great success with that, and usually the other part follows with that.
[00:32:55] Benjamin: I feel like I was quite fortunate at Atlassian, and had a big push into design when I joined that company from the executive team. I think generally if you say to execs that the customer experience is something that you value, and essentially the business should value that too, and customer centricity is something that can set the business apart from competitors, in my experience it's often hard to disagree with that. Executives typically do want to be quite customer-minded. Then you just have to connect that to design and user experience as a solution to that problem, I suppose. But yeah, at Dovetail obviously we're very customer oriented and design focused.
[00:33:37] Benjamin: I think we'll hit one more, and anonymous question person is doubling down on privacy here, so they're asking for a friend. But how do you define the nuanced difference between a platform and a product team?
[00:33:50] Benjamin: Sure, I'll let you.
[00:33:51] Fahad: Yeah, I'll take a run at it. I touched on it very briefly at the beginning, but one of the things I think about is, are there multiple workflows and multiple roles interacting with this thing that we've created? And is this a long-running thing? Is this going to be a codification of a policy for the business? Is this going to be a way for this team of people to create an outcome? And is the workflow more important than the feature? And is the benefit to the organization kind of an all boats rise together, or an individual benefit to the user?
[00:34:29] Fahad: And usually the former leads me to think this is a platform that we're working on, because the standardization of things and the connecting of multiple roles and optimizing their workflows is going to yield some kind of benefit that won't be immediately obvious to the individual user working on it. You're asking them to give up a little bit of personal autonomy in order to benefit the full team, and in order to ship things together collaboratively as a team faster. Whereas on a product team, or on a feature team, I find that you're really, really focused on an activity, or a single user, or the outcome for that user, without worrying too much about how is the organization or the system going to benefit, how's the whole team going to benefit from this. That's sort of how I take a look at it.
[00:35:10] Benjamin: Yeah, I would say platforms are more complex, not necessarily more complicated. It just comes down to whether you're designing and building for multiple use cases and personas, and then also you get into this world where maybe you have ecosystem vendors, add-on partners, API developers and so on. It just becomes more complex. Product team, I think, is just a bit more focused. They're working on a specific set of features, or for a specific job to be done, or a specific persona. And at some point product companies tend to become platform companies, particularly in SaaS, where they're expanding their use cases in order to expand their TAM, right. So it sort of happens over time.
[00:35:49] Fahad: That's a great way of looking at it. I hadn't considered that, but yes, as you want to grow your total addressable market, everybody kind of becomes a platform at some point, which is just really a complexity, because you're having multiple use cases and personas and different functionality and so on.
[00:36:02] Benjamin: Yep.
[00:36:03] Host: Great. Okay, we're at time. One last round of applause. That was great. Thank you guys so much.
[00:36:11] Fahad: Thanks a lot. Appreciate it.

