Benjamin Humphrey, Co-founder & CEO at Dovetail, talks to Maze about the comparison between design software and research software maturity levels, why customer-centricity is the key business differentiator, and how research repositories can help you build more customer-centric organizations.
The Optimal Path is a podcast about product decision-making from the team at Maze. Each episode brings in a product expert and looks at the stories, ideas, and frameworks they use to achieve better product decision-making—and how you can do the same.
You can connect with Benjamin on LinkedIn.
Resources mentioned:
Follow Maze on social media:
To get notified when new episodes come out, subscribe at maze.co/podcast. See you next time!
Ash Oliver:
Welcome to The Optimal Path, a podcast about product decision-making brought to you by Maze. I'm your host, Ash Oliver, UX Designer & Design Advocate. Great products are the result of great decisions, decisions that deliver value for customers and the organization. In this podcast, you'll hear from designers, product managers, and researchers about the ideas informing decision-making across all aspects of product development.
Ash Oliver:
Today I'm joined by Benjamin Humphrey. Benjamin is the CEO and co-founder of Dovetail, a user research platform for collaborative analysis and the storage of research data and insights. Before starting Dovetail in 2016, he worked as a lead designer at Atlassian. He has over a decade of experience in software development, design, and strategy from time spent working at agencies and software companies throughout New Zealand, Silicon Valley, and Sydney. So stoked to have you here, Benjamin.
Benjamin Humphrey:
Thanks, Ash.
Ash Oliver:
All right. So our topic today is research repositories, clearly near and dear to Dovetail, but also how research repositories help to inform decision making. So I thought we could start from more of the genesis of Dovetail. Maybe you can explain what the inspiration story was behind the company.
Benjamin Humphrey:
Yeah, totally. So we started back in 2017. Brad and I went full time on Dovetail. We left Atlassian where I was a designer and he was an engineer. But the story of the company starts earlier than that. I was working as a designer at Atlassian and was fortunate enough to work with a researcher on a project around onboarding. We were working in the growth team and looking at how people onboard into JIRA. And the researcher I was working with was running a diary study, which is a longitudinal research method, essentially asking people in the evaluation flow to write down their thoughts and feelings on JIRA each day. The way she did this was with a Tumblr blog for each participant. It was very manual and it was very scrappy because there wasn't really any good tool for this. After she collected all the data over a couple of weeks, she took it into Word and sort of formatted it into A4 pages. Then she printed that out and put it on a whiteboard and manually highlighted stuff and did this kind of synthesis.
Benjamin Humphrey:
So at the time, as a designer, I was using Sketch, which had just come out or we had just started using it at Atlassian, and this is pre-Figma, but post Photoshop. And I was like, "Wow, this is so cool. Here's a product that's built for purpose or fit for purpose for my use case of doing vector based UI design." Whereas Photoshop obviously built for originally photo manipulation. So that was pretty cool and I was looking at what this researcher was doing and I was thinking, "man, there's like nothing like this for research." So the original sort of insight, I suppose, for Dovetail was, hey, I think I can build some software for researchers. Not quite sure what it would end up being, but where I started was a low hanging fruit, which was replacing Tumblr for diary studies.
Benjamin Humphrey:
The study I think was called EFI, engaging first impression. So she'd create EFI1.tumblr.com, EFI2.tumblr.com, and then send them the username and password to log in each day. It was very manual and so I ended up building the first version of Dovetail, which I outsourced most of the engineering onto Poland. While I was still working at Atlassian, I did the design work and stuff in the evenings and on weekends and got the MVP up. It was in this research space and started getting some customers and things, but the lesson was essentially after maybe six months, nine months or so. We had crazy high churn because everybody was finishing their diary study and they were canceling. So that's that point that I realized, hey, maybe there's something in not how you collect the data, but rather what do you do after you've collected it. That's how we got onto the analysis sort of feature set that the product has today.
Ash Oliver:
That is so interesting. Yeah. I was going to ask if there was some unaddressed need that you spotted, but Tumblr couldn't be more blatant. That is very resourceful on behalf of that researcher, but wow, what a time and place. Well, it's been six years now since starting Dovetail. What comparisons can you draw between design software and research software maturity levels?
Benjamin Humphrey:
Yeah, I mean, I think that design software is very much in the D side of R & D. It's sort of design and development and execution, and that's a lot more mature as a category of software than research is. I think there are maybe a few reasons to this. My biggest theory is that engineers build things for themselves first. You solve your own problems first. So 30, 40 years of the Bay Area building software for themselves, 20 years of the cloud essentially means that we have very mature tooling and software on the development side. So that's cool. Then the next obvious thing was how do you make the products look good and how do you improve the UX? So it makes sense that we've got good design software. There's this sort of plethora of design software now, which is great.
Benjamin Humphrey:
Evaluative research is also somewhat on the development side, like prototyping and building features and usability testing. What hasn't had a lot of love is the kind of stuff that happens earlier in the process of building a product or delivering a service, which is coming out with the original moment or the opportunity, and then exploring that and doing that most top of funnel research or exploratory research. That's traditionally just been achieved through a huge variety of fragmented tools. For example, survey tools, we've got Qualtrics, SurveyMonkey, Typeform, Wufoo. There's lots of ways to survey customers and NPS products, like Delighted and AskNicely. And you've got in product feedback stuff, you get UserVoice, for example, UserZoom. There's recruitment tools. It's a very fragmented market and there's not really any consistency.
Benjamin Humphrey:
If you go and ask a lot of people these days, it's like, "What do you use for design?" Figma wins. It's almost standardized, and it's the same on engineering. You see TypeScript and React very much becoming the standard on the front end, whereas that wasn't the case maybe five years ago. GitHub is kind of a standard in terms of where you host code. So you go to company to company and you hire an engineer and it's like, "Well, how we work is somewhat standardized and the tools that we use are somewhat standardized." That's great because you see that in other industries too. I assume the manufacturing methods from Volkswagen to Porsche to Ford to Honda are somewhat the same,
Benjamin Humphrey:
If you were an automotive designer and you went from one of those companies to the other, you would have transferable skills. Whereas in research, you go ask companies how they do research and it's always different. They have totally different projects—the way they approach them, the timelines, the tools that they use. The way that insights are shared within an organization varies wildly. Even the roles where research lives, like the organization chart. So I think that this is a very immature space and it's getting more mature over the next five or 10 years. It's just going to go the same way as the design.
Ash Oliver:
I couldn't agree with you more and I think what contributes to the problem, like you were saying, so much fragmentation in the tooling means there's so much fragmentation in the insights or the collection of where these things happen. So what do you think are the most pressing challenges that teams are looking to solve through research repositories?
Benjamin Humphrey:
Yeah. I mean, this is a great question. So again, if I contrast engineering and research, engineering is a cumulative discipline. It's a value added thing where if you join a company and you contribute code and you build features and you write lines of code, that is added to the code base, and when you leave the company, you don't take that with you. Imagine if you left the company as an engineer and all the features you built disappeared. That doesn't happen. So organizations get cumulative value out of engineering and you are making money from features that you built years ago. Whereas research, I think the way that it's conducted and how the insights are stored and shared and disseminated is very non cumulative. I use research in a lower case R here.
Benjamin Humphrey:
So a lot of people do research, right? Product managers are always talking to customers. Designers are conducting usability testing and talking to customers. Sales people are fundamentally researchers. They are gathering requirements from customers. They are feeding that back to the product team. They're learning what sort of use case is going to be applicable for the product. Marketing teams are researching the audience to try and find the best positioning. Customer success and support people are interfacing with customers on the front lines. So a lot of people do research in the lowercase R context and the problem is that when one of these people leaves the company, they take most of what they've learned with them. If you lose a really good product manager, you lose a lot of interesting context about the customer, the audience, how they use the product, the strategic thinking. That goes with them. That's why when a PM leaves a company on good terms, they spend their last four weeks having coffees with every other person in the company.
Ash Oliver:
Trying to brain dump it.
Benjamin Humphrey:
Trying to brain dump, exactly. Literally people call that brain dump. Then when you join a company as a product manager, you do the same thing in reverse. You're trying to learn everything you can about the product and the decisions that have been made in the past. We have decision registers and stuff, maybe document on Confluence and things, but it doesn't really tell you the why behind a lot of the stuff. So ultimately what we're trying to do is build a knowledge base of all of the why behind the decisions that have been made in the product, but then also have it as a library that you can go and reference when you are looking to make another decision. Because another challenge is that a lot of work that gets done, it's done for a specific project and the project has a hypothesis and goals, but naturally those conversations will cover all kinds of things. It will be cool if you could leverage previous research in different contexts or different projects.
Benjamin Humphrey:
This is maybe where academic research might be like, "Oh, this is a little controversial. You're kind of taking the research out of its original project and using it in a different context," but I think exposure to customers drives a lot of understanding, which facilitates faster decision-making across the business and also builds empathy for customers which results in better outcomes, products, and services. So my theory is that development is fairly straightforward these days and so the way that you differentiate is through building the right thing for the right person. That's how you win, not necessarily how you execute. That's getting easier and easier thanks to technology.
Ash Oliver:
Absolutely. Yeah. I think it's a really interesting point that you make about this compounding value of research. It makes so much sense when you think about the learning and insights that were generated, that then could go on and continue to just be evergreen and inform the future decisions and really taking research out of this old guard context that I believe is, what you said, just more of a snapshot in time, because it's all just based on projects with a start end date. It's interesting that you mention of the differences between what it was like to create software in the past versus what it's looking like creating software now. Effective customer focused organizations can really make strategic research driven decisions often, but then there's other teams that are on the total opposite side where they can't even get their research noticed or actioned by decision makers. What do you think contributes to this polarity?
Benjamin Humphrey:
Yeah. I think if we look at the last 20 years, if we go back to the start of 2000s, software companies, cloud software companies in particular, were very much on the edge of technology and that technological advantage allowed them to do things that non-software companies couldn't do. Take banking, for example. It's effectively a non differentiated product that banks sell. Everyone has savings accounts, checking accounts, payments, transfers, interest rates are fairly much the same, mortgage, home loan rates. It can compete a little bit with the numbers, but nowadays a lot of people choose their banks, I think, based on their brand and user experience.
Ash Oliver:
100%.
Benjamin Humphrey:
Essentially it's the UX, how do you interface with the bank? You use your mobile app 90% of the time and then maybe occasionally you'll log in on desktop. Occasionally you'll go to the branch.
Ash Oliver:
If one exists.
Benjamin Humphrey:
If one exists. So essentially banks have almost become software companies. So I think that you go through this transition phase, but before they could work on UX and design, they had to move to the cloud and figure out their technology side. So there's almost this lag time where the customer focused companies are at the front valuing research and valuing customer insights and being driven by customer requirements. Then you've got another cohort of companies that are just trying to work through their technology challenges. Then now that they're sort of getting through that period and the technology challenges are solved, the next phase is to actually just try and refine and build a better customer experience. I think that's maybe why the focus has just been different, but I think it's changing. You can't really differentiate as much on execution. It's more about on positioning, branding, on your research side, and whether you're building the right thing for the right people.
Ash Oliver:
Absolutely. Yeah. I was thinking about the industries also that are at the forefront, I guess, of championing and really maturing their practices within research and industries like in agriculture or in healthcare, for example. Now that same transition is happening, like you're describing. It's like you're building software technology, but for a huge group of individuals and one that really requires the deep understanding of those individuals, especially in their own context. How do you see research repositories or insight libraries being similar to that of design systems?
Benjamin Humphrey:
That's a good question. I think about design systems for a second—what they allow you to do is standardize, scale, give designers the ability to get out of maybe pixel pushing all the time and recreating the same sort of visual design every time they want to focus on maybe the UX. So it's really an enabler for speed and for scale. So I think in that regard, an insight library is the same, right? Maybe more for product decision-making. We have customers who will pull up Dovetail in meetings and do a quick search and see if they can find some useful quotes from customers in the meeting. So I think that in that sense, it also enables you to have standardization in that you can build out this taxonomy or classification system to classify the research work and also the continuous research that's going on.
Benjamin Humphrey:
So you've got NPS data feeding in, you've got social media posts, you have G2, Capterra reviews, App Store reviews as well. I think that being able to capture all of that and do something meaningful with it makes sense. For example, JIRA at Atlassian gets quarter of a million pieces of in product feedback every year. When you're getting that much feedback and it's all unstructured and qualitative, how do you make sense of that and how do you classify it? So perhaps having an existing taxonomy or classification system in a library can help you route that in the right direction and then eventually get that to the right team to improve or fix or respond to the customer, but I haven't come across any organization that has a good answer to this. It's always messy. This information flow challenge, I think, is something that an insight library could help with because you got to think of it like data in, data out.
Benjamin Humphrey:
So with Dovetail, what we're trying to do is build that collaboration layer where that's how the rest of a company interfaces with research. It's a little different to Figma because with design, there's a finite and very measurable group of people who need access to a prototype. What we're trying to do is basically end up with research.amazon.com or something, for example, and then everybody at Amazon can have access to this portal, whether you're a salesperson who's looking for some useful customer testimonials to tailor a demo, through to customer success person who's looking for other examples of maybe similar customers who have onboarded the product, through to research and PM and design who are maybe making a decision around a product feature. Ideally it's this portal, in lack of a better word, that everybody can access. So I think that's how it could help you scale customer centricity. I think that's how it's similar to design systems because it's a scale and standardization and also an efficiency thing.
Ash Oliver:
Yeah. Just thinking about, we were talking about that redoing of the work because perhaps it's locked off somewhere else. Just thinking about all the streams of those insights that are coming through and in comparison, it's maybe multiple tools within one team and that's all siloed off, and then you've got multiple tools on another team and that's all siloed off. Nothing's coming together and it's all on these multiple channels. Yeah. That's very hard to harness and then to even make sense of any of that is an entirely different endeavor. So Dovetail really seems to focus on this very important aspect of not just the UX research, but essentially the linchpin to how a company learns.
Benjamin Humphrey:
It's interesting and it's hard because I feel like knowledge management systems have been around for a while. People do use Google Drive and Dropbox and they use Confluence.
Ash Oliver:
I was going to say, Confluence. Yeah. Did that influence you much in even the beginning part of...
Benjamin Humphrey:
Yeah, I think so. There's some really interesting functionality in Confluence and then maybe some ideas that we had that never got implemented that I think would help for this kind of use case, but there are also challenges. For example, we see this with our customers, a hierarchical folder file structure in Dropbox or Drive or a nested page structure in Confluence. It doesn't really work at the scale that it needs to. So instead what we think about is how do we make it more like a social network? One of the challenges of Confluence is that it needs to be gardened. Somebody needs to look after it, and if you don't do that, then the user experience is really bad. It's kind of the same with Google Drive and Dropbox. If you just drop somebody in to some folder file hierarchy, it's not a very good user experience as a consumer. You can't really find stuff easily.
Benjamin Humphrey:
Whereas if you think about social networks, the onus of the consumption experience is not placed on the users. It's placed on the software. The software and the design is the thing that determines what you see. So it's more about algorithmic feeds and personalization and search that allows you to find unique things quickly, and concepts like hashtags which allow you to follow them or drill down into a specific topic. So you think about Twitter, again, it's an unmotivated experience, right? But then it's all about personalization and there's no hierarchy. There's no folders. There's no files. That's the only way you can really scale when you have hundreds of thousands of things being pumped in.
Benjamin Humphrey:
Going back to my example of the feedback that comes through JIRA, you just can't expect the person to manually classify 250,000 pieces of feedback. You can't even really imagine a user interface. What are you going to do? Put that onto a folder? So I think that the actual approach that we're taking is much more like a consumer grade social network and it's a flat graph linked through relations and topics, like hashtags, and having research reports just floating around that are then referenced by projects and vice versa. One of the things you can do in Dovetail, which is kind of neat, is that as an admin you can create custom feeds for the homepage and then you can name them. So maybe you could have a feed set up which is latest insights. Then you just configure it to show you that.
Benjamin Humphrey:
So you get some control over the consumption experience, but ultimately the software is the thing that should be tailoring it and the end users who are just putting their interviews in there and putting the usability tests and generating insights and generating research reports and findings should not have to think about how they're consumed. That's a little bit of a difference between traditional knowledge management and wikis: they're very much like, "Here's a document editor. Here's a hierarchical structure. Figure it out." Whereas with us, it's like, "Don't figure it out. Contribute your stuff into the ether and then we'll tailor it based on who the consumer is."
Ash Oliver:
I love the approach and I think it's really important for two reasons. One, the mental models like you just described, because I could not imagine, not just the user experience of what you just described, all the pruning that would have to be involved to keep things up, but knowing where to find things. Then the other thing that you mentioned, just maybe going back to that compounding effect that you described in the beginning, is this has a compound benefit to the appetite for people to actually consume the information because it's also being served up and provided in a way that's not just easy to get to, but enjoyable to experience while you're actually consuming it too.
Benjamin Humphrey:
Totally.
Ash Oliver:
That's a totally different approach than I've seen others take.
Benjamin Humphrey:
Yeah. I think that feedback loop's important too. This is the kind of collaboration layer where you want to be able to ask follow up questions on the research. You want to be able to dive in and see the original study. One of the things people like about Dovetail is you can embed a quote from an interview in a insight which is sort of a research report, but then if you come along as a consumer of that report, you can actually click through and then watch the original recording and see the context around the quote. It helps to build trust in the research work that's happening because a lot of people are very skeptical of researchers. They're like, "They're going to just choose the quotes that support their narrative," but if you actually have that traceability back to the original recording, you can go back through and see. Even knowing that you can do that almost builds trust as well, because it works out in the open.
Ash Oliver:
Totally.
Benjamin Humphrey:
So I think that kind of collaboration, showing how the sausage was made, does build trust in the research and therefore I hope that it can be more impactful and then also that the team can move faster with a higher level of trust. So there's some stuff there around traceability. I think it's pretty interesting. Then it goes the other way as well. So once you have that research report and you have some insights and your findings and outcomes from the work you're doing, there's this natural desire for it not to become shelf ware. People want to see stuff through to completion.
Benjamin Humphrey:
Researchers, they want to see the insights they've come up with get actioned, and whether that's like, "Okay, here's some really tangible usability issues with our onboarding flow, for example. Can that get prioritized by a product team? Can the engineers fix the issues?" That's pretty easy because you just can connect your insight to a JIRA ticket and then the engineer picks it up and can see why they're doing it, but then it gets harder and harder the more abstract you become. Once you get into exploratory research and strategy, how do you know that the work you're doing is actually going to get actioned by the business? So there's this whole traceability angle at that level, which is interesting too. Imagine if you could have a pull request in GitHub where the line of code that you're writing is actually connected all the way back to the original usability test, for example, where the insight was found. That's pretty sweet, right?
Ash Oliver:
That's amazing. Yeah. I mean that's user centric too.
Benjamin Humphrey:
Yes, and I think that helps motivate engineering teams as well because a really common thing you hear about is designers and PMs and researchers trying to invite engineers to usability testing sessions and trying to bring them along to consume their research and take them to conferences and meetups and have a community where they're active in the community. It's all to try and increase exposure hours, and the reason is because you're more motivated if you know that you're working on things that are tangible human problems and you can empathize. I remember at Atlassian as a designer, if I was doing a usability test, that moment where you're watching the participant struggle with what you've designed, I think that is very powerful, this customer empathy. I think our mission is to try and leverage that to make organizations more customer centric, which then naturally will result in better products and services for the world.
Ash Oliver:
Amazing. Yeah. That's incredible. When I'm thinking about some of these pieces in regards to research as well, I was wondering maybe you could speak a little bit about atomic research and how it relates to research repositories and specifically in synthesizing insights as a team.
Benjamin Humphrey:
Yeah. That's a good question. I mean, atomic research somewhat popularized by Tomer Sharon from WeWork several years ago with that medium article. I think it does relate to design systems in a similar way in the sense that it's about modularity and reuse and bite sized components, if you will, but I think the point is, you might have heard this phrase, "is this a meal or is this a snack?" I think one of the challenges with research, if I can maybe talk a little bit about the kind of discipline at large, is that a lot of researchers work in commercial settings these days, but have come from academia. There's a different cost and risk equation between working in academia and working in a commercial setting.
Benjamin Humphrey:
So in academia, you're not as time poor and quality of the research and the statistical significance of the research needs to be a lot higher. Whereas when you get into the commercial sense, that equation changes and it's more about moving quickly and it's more about having directional indication than it is about having certainty. You may need certainty in the academic sense, but in a commercial sense, it's all about directional indication. So I think that natural tendency is that maybe academic researchers who have come into commercial settings are used to writing these quite long form research reports. That kind of culture is not really how people want to consume insights. Frankly, I don't think people have the time to read a five or 10 page research report, but that's typically how it works.
Benjamin Humphrey:
So I think atomic research is a bit of a challenger. It's saying, "Hey, can we pull stuff almost out of context and have insights that are sort of floating free form, self-contained? Then how can we categorize those in a way that we could then reuse them?" So one of the features in Dovetail, which I like, is you can do a search for something, and then there's a little button on each card where you can click and add that thing that you've found to a story. It's a little bit like pinning something on Pinterest. You're not necessarily adding entire reports, you can add parts of reports, or you can add specific quotes, or you can add specific support tickets, or NPS responses, or a specific interview to your own little private story, and create that as your own little Pinterest board, but with your context and what you're trying to get out of it and your project that you're working on.
Benjamin Humphrey:
So I think atomic research facilitates that and the opposite, which is maybe lengthy Word docs in a Dropbox. It's just impossible. How can you go and grab a specific quote or something like that, especially because it's all disconnected as well. The original recording is going to be on someone's computer and then they've transcribed it somewhere else. So it's just completely fragmented and broken up. Atomic is kind of the way that things go, I think, but it is a bit of a branded term. I think the point of it is really to break up these long form research reports and allow further reuse of the work, which is essentially just an increasing ROI thing. Then synthesizing insights as a team, I mean, research is kind of tricky. I don't think it's similar to design. There's aspects of it that are collaborative, but a lot of analysis and synthesis are sort of deep flow work. I think a lot of the collaboration happens at the output level with the sort of back and forth and the feedback and questions and clarification.
Ash Oliver:
Yeah, it's interesting. I love what you said about weaving the narrative of the user through being able to surface these more, even story level type, either quotes or small fragments to then pull the thread and then illuminate the rest of the story. So powerful when we think about just why we're actually doing any of this to begin with. Thinking about more of the synthesis side, especially for teams that are now doing this, as you mentioned, more collaboratively and beyond maybe the confines of just the research team, what are some of the mental models that go into creating a place where synthesis can occur?
Benjamin Humphrey:
Yeah. So analyzing unstructured data and synthesizing it is a very meaty challenge from a software point of view. There's spreadsheet people who love the filtering and views that spreadsheets allow you to do. There are physical space folks who need to immerse themselves in sticky note walls and highlight literally an affinity map, literally.
Ash Oliver:
That's me.
Benjamin Humphrey:
Yeah, and just sort of mind meld with the content also digitally, which is more like 2D canvas kind of stuff. People who use Miro and Mural for example. So we've just finished a heap of work, actually, and it's releasing at the moment, called Project Views, which is essentially trying to build in the flexibility to Dovetail so that we can satisfy these different groups of people. So we don't really have an opinion on how you should synthesize, but instead, what we're trying to do is give you the tools to do it the way you like it.
Benjamin Humphrey:
So an example of that is we just launched this new canvas view. Let's say you do your transcript and you do a bunch of highlighting and tagging. You can then look at your highlights and you can look at them in a spreadsheet where you can create filters and you can create sorts and change the columns, or you can use a Kanban board where you're just doing basic 2D affinity mapping, where you're dragging and dropping cards between columns. You can synthesize your highlights into insights that way or you could create a canvas, which is the new feature we just launched and essentially you can do your 2D affinity mapping on a canvas, similar to how you can do it in Mirror and Figjam.
Benjamin Humphrey:
But one of the cool things about the Dovetail canvas is, because we are dealing with unstructured data that's become structured, you can do things that you can't do in those other products. So an example of that is, because you're adding maybe segmentation data—for example, you've got the company size, the industry, the demographic data—you can auto group on the canvas in one click and be like, "Okay, affinity map by age, or affinity map by industry," right, and use that as a starting point. That's where, in Miro world, you only have post-its and so you can't do that kind of stuff. So we want to try and build on that intersection of structured data. The objects are structured and have attributes that are structured, but then embrace the sort of free form 2D space, if you will. A lot of people think we're nuts building a canvas feature inside Dovetail, but the reality is that's how people want to work.
Ash Oliver:
That's incredible. I'm very excited to hear about that. I mean, I'm a big tools for thought and canvas guy and long time Miro user. I'm sure there are a lot of people out there that might have research repository and then use Mural for maybe the synthesizing, but yet again, we're in a situation where we've got now two different tools and having to transfer that information back and forth.
Benjamin Humphrey:
Yeah. Yeah. Honestly, I think that canvas is just another way to visualize lists of data. As a UI designer, I remember working on JIRA and it was one of the early products to have a Kanban board and the implementation from a technical side was really clunky, but then Trello was best in class in terms of interactions for how to do a Kanban board. Over the next few years, everyone figured out how to do drag and drop on the web and so then we saw Asana launch Kanban board. We saw Air Table launch it. We saw GitHub Issues launch a Kanban board. All of a sudden, what was originally a differentiator then just became a table stakes expectation for users. They were like, "Okay, I can view my issues in a list or I can view them in a board." That just became every product has that sort of feature.
Benjamin Humphrey:
I think that canvas is just the same thing. It's just another way to visualize lists of data. This is where I go back to earlier point that the execution is not really the way you differentiate these days. It's more about is that the right thing to build? And when do you do that? The sequencing, the prioritization, and who's it for, and who's going to buy it, and what's your marketing and your positioning and your branding, and essentially you go to market. So I think that the equation is actually shifting a little bit away from development and engineering and more into research. I think that's why everybody has a desire to build a research repository, because it becomes an IP, like a massive differentiator internally. If you can figure that out, if you can be more customer centric or customer focused than your competitor, then that leads to better stuff and then you'll win. So I think that's the longer term thing.
Ash Oliver:
Tremendous. Yeah. That is huge. So looking out for the next, maybe 10 years, what do you see around the corner for insights libraries and potentially learning at scale?
Benjamin Humphrey:
Yeah, I mean, so I have crazy ideas in this spectrum. Maybe this is out of touch founder speaking, but there's a few interesting things. So a lot of people ask us, "Does Dovetail tag things for you?" The answer to that is no, because AI cannot auto classify this type of content for a variety of reasons yet. It can't do it yet. The reason is that you need a lot of context to apply relevant tags, accurate tags, and you bring that context as the person doing the research or the synthesis, but the other reason is that we have such a diverse customer base, we can't train some global model. So it would need to be bespoke for each customer, but what happens is over time I think that amount of data for training and accuracy of AI model equation gets better, similar to what's happened with transcription. We're not talking NLP here, I'm talking AI classification.
Benjamin Humphrey:
So I think over the next 10 years, what would be awesome is this notion of what I call black box research, where essentially you're feeding the software with all the inputs, whether that's support tickets, sales calls and transcripts, interviews you've done and usability testing, survey responses, and it actually starts to figure out some patterns and recognizes some themes and potentially starts to make recommendations for decisions. So that's one thing.
Benjamin Humphrey:
Another crazy idea, which is probably longer than 10 years, but with the work that's going on with Neurolink I think it would be really cool if you're onboarding at a company and instead of sitting through lots of one on one sessions and trying to learn and clicking around Confluence and clicking around Dovetail and watching videos, if you could just plug into the Neurolink and we could have an integration and then just be like, "Hey, you can suck down." Then all of a sudden, as a new PM joining a company, you have the same amount of context and knowledge as someone who's been here for 10 years. How cool would that be? How fucking effective would your company be if you could onboard new hires that quickly?
Ash Oliver:
Just like what you were saying at the beginning, with that PM that leaves, it gives a whole new meaning to drinking from the firehose.
Benjamin Humphrey:
Yeah. You can picture some dystopian training room where new starters are in recliners plugged in.
Ash Oliver:
Plugged in.
Benjamin Humphrey:
Plugged in Matrix style, and then they're just downloading all the research that's happened since the inception of the company.
Ash Oliver:
Where the new hires? Oh, they're downloading.
Benjamin Humphrey:
Exactly, downloading. Yeah. I think there's a couple of things, but certainly, taking that idea to an extreme, how do you maybe build software that allows you to do something like that as fast as possible? I mean, you know co-design, right? It's about trying to get customers in the design process. The next best thing is maybe being able to actually pull up video recordings and clips and stuff really quickly and making that time to value equation as short as possible.
Ash Oliver:
That's really exciting. Yeah. Just thinking about being at the crossroads of any decision, it's almost like having that user in the room.
Benjamin Humphrey:
Exactly.
Ash Oliver:
But you have that right at your fingertips. Yeah. That's incredible. We specifically work in this meta universe where we're designers designing for other designers or adjacent roles within our industry, and as a designer and now a founder, I'm curious, what do you think it takes to build products that users love?
Benjamin Humphrey:
Yeah, this is a great question. Let me think about this. So I think a few things. It's very functional, very straightforward and intuitive and easy to understand. One of the things I don't like is these click through hand-holding tutorials where it points out, "hey, click this button to do this." Those kinds of things are crutches in my mind. The way that I think about designing the software is how do you make the UI itself intuitive, thus that you don't need to explain it? That's one aspect. I think there's a lot to brand and community that we see good companies leverage. Figma is doing a great job. I think Notion does a great job. Slack in the earlier days. Yeah, Webflow. I think embracing the community, just being open and authentic builds that trust and empathy in both directions.
Benjamin Humphrey:
But yeah, I mean, that's a tough question. It's a tough question. I think prioritizing, obviously. We at Dovetail, are very customer centric. The whole company is very much immersed in what customers are doing, whether that's looking at analytics and the quantitative side or whether that's being actively involved in our Slack community. Yeah. Our team's always on interviews and doing usability testing. We have an in-house researcher who facilitates a lot of this stuff. Then even our support person, she makes a monthly recap of what themes and stuff and support and uses stuff to offer that. So I think it's a cultural thing too.
Ash Oliver:
I'm just thinking about what you're saying with the drinking your coffee and just catching up with users and drinking your own champagne in this case where you're using Dovetail to learn about users using Dovetail is quite the...
Benjamin Humphrey:
Yeah, it's very meta.
Ash Oliver:
We've had quite a few meta metaphors in our conversation. So I want to end our session. We ask three more personal questions about you at the end of every episode, just to get to know every guest a little bit better. What is one thing that you've done in your career that has helped you succeed that you feel very few other people do?
Benjamin Humphrey:
Man, this is a really good question. I would say I'm fairly decisive and I think I play out a lot of different scenarios in my head quite rapidly, and I think I'm very pragmatic. So it's always about cost risk equations and sort of like, okay, if we make this decision, this is the cost of that decision, and I'm comfortable with that. One thing I've been working on a lot recently is my sleep. I think people don't put enough emphasis on sleeping and unfortunately, people who are highly productive, they're like, okay, you mustn't sleep very much. You're up late working and so on, but actually it's the opposite. You need to get your eight to 10 hours of sleep a day. I started wearing ear plugs. I have a wind down period. I've double glazed the windows in our bedroom to keep the room quiet. So investing in sleep and making sure that you get enough of it, I think actually has helped me a lot in terms of just being able to be calmer.
Ash Oliver:
Totally. I love that you made that link to sleep because I don't think any good pragmatic decisions get made on low sleep.
Benjamin Humphrey:
Exactly.
Ash Oliver:
All right. My second question here for you, what is the industry related book that you've given or recommended the most?
Benjamin Humphrey:
I talk a lot about Legacy by James Kerr, which is the book about the All Blacks culture as a Kiwi, and it's sort of a humility kind of culture and humble culture. I mean, there's all these principles in there around, which I think is applicable to business. Another book that I've read recently was Super Pumped, the book about Uber. I love corporate gossip books. So it's really interesting learning all about that. Then I'm also a sucker for the drama too.
Ash Oliver:
Okay. My last question for you, Benjamin, what is an unusual habit or an absurd thing that you love?
Benjamin Humphrey:
I can't think of something for an unusual habit, but something that is absurd that I love is this show called the Eric Andre Show and it's not going anymore. I think it was four seasons, but it's just a comedy skit show. I highly recommend checking it out. It's not for everybody. It's very weird. I think it's on Adult Swim and essentially, this guy's just crazy. The opening introduction is him just screaming, running around, smashing the set. You're just sitting there thinking what the heck is this, but I love it because it's so weird. I find it really interesting and creative and I'm just like, "How did this guy come out with these skits?"
Ash Oliver:
I love it. That's amazing. This has been such an incredible conversation. Thank you so much for being with us and for sharing such deep insights. This is incredible and it's so exciting to hear about everything that Dovetail is up to and where you all are headed in the future. So thank you very much.
Benjamin Humphrey:
Also thank you, Ash. Really appreciated the conversation.
Ash Oliver:
The Optimal Path is hosted by Ash Oliver and brought to you by Maze, a product research platform designed for product teams. If you enjoyed this episode, you can find resources linked in the show notes. If you want to hear more, you can subscribe to The Optimal Path by visiting maze.co/podcast. Thanks for listening. And until next time.