In episode 5 of Developer Love, Patrick Woods speaks with Knut Melvær of Sanity.io. They discuss building developer tools, promoting psychological safety, and nurturing connections within the DevRel community.
About the Guests
Patrick Woods: Awesome, Knut. Thanks so much for coming on Developer Love today. I'm really excited about having this conversation.
Knut Melvaer: I'm excited to have the conversation, too.
Patrick: Awesome. Let's start out by talking a bit about who you are and where you're coming from.
Knut: I'm Knut, and that's not a usual English name because I'm from Norway, where I'm now talking to you over Zoom, like everyone else is doing these days.
I run developer relations and support over at Sanity.io, the platform for structured content.
Patrick: That sounds great. So, tell us more about this concept of "Structured content."
Knut: Sure. We are all used to dealing with CMSs, and what CMSs has typically done is to force your content into posts and pages, and make it a bit hard to get to it in meaningful ways if you want to integrate it with apps or other services and stuff.
That's where structured content comes in. We try to let you treat content as data.
We type it, we make it nice to interact with it for APIs, and we build a lot of infrastructure and tooling to make that a happy thing to work with.
So, that's the gist of it.
Patrick: Cool. What are you focused on as the DevRel at Sanity?
Knut: Since we are mostly making developer tools, that means that we have to make sure that developers actually manage to use them and are happy doing so.
My main task is to make that happen, to listen and observe how they interact with the various parts of the Sanity platform and to make the materials and the guides and stuff they need to understand it all.
Patrick: Cool. What would you say is your favorite trick of the trade?
Knut: I think the realization that people are way more approachable than you think.
Both the people who are using whatever you're building, but also the people you are looking up to.
Especially in the DevRel world, just reach out to a person that you look up to and there's a pretty good chance you will get a nice response and learn something.
Patrick: Yeah, that's a great insight. Any favorite examples of when that worked out for you?
Knut: I entered this field two years ago not even knowing what a developer relations person was, and I guess I realized that I have done the things before.
But coming to, for example, the JAMstack world and meeting all the great people over at Gatsby and Netlify, and learning a lot, I guess that would be the example.
Patrick: Cool. You mentioned that you got into this world just a couple of years ago.
What brought you into the world of DevRel and community building?
Knut: My formal education is actually within the study of religions, which is the non-confessional academic study of religion as a cultural and social phenomena.
For reasons, I pivoted into technology and user experience and design, and so on.
I was a consultant and during that work I got access to the Sanity beta version and it was a revelation.
I also got to know the team and they asked me to join two years ago, and I did. Now I'm here.
Patrick: That's an awesome story. So many of us have had divergent paths to where we are today.
What do you think is the secret to building things developers love?
Knut: I hope it's not a secret, because the really dry answer is probably what you will get in any sort of design process.
You need to observe how people interact with whatever you're building. Where do they struggle, where do they succeed? Fix the former and do more of the latter.
Patrick: Cool, tell us about that. So many of our listeners are super interested in tooling, so I'd love to hear the details of your Slack tracking apparatus.
Knut: We are now 3,000 people in this workspace.
It's not like 3,000 people actually communicating every day in the inside of this thing, but there's enough stuff that happens and enough people who need some pointers and some help.
It's not easy to stay on top of, so we needed something.
We weren't quite ready to move the whole operation elsewhere, so we used APIs and we used Sanity to make it easy to tag a message and follow whatever happened in the thread. We made our own Freshdesk, in a way.
Which sounds bad, but was actually awesome because we can also have an archive of what is happening in this free Slack workspace. Don't tell Stuart or Phil about this .
Also open source, actually documented, so you can go and try it yourself.
Patrick: Awesome. What has been the impact of having that tool available for you and the team?
Knut: The impact is twofold. First of all, we can actually keep track of the questions and follow up and so on.
And the other side of it is that we are not only saving the conversations and we can look back at them and you can also tag them and we can make statistics like "How many of the questions was related to a bug or a feature request or just the lack of documentation ?"
And then we can look at the deltas from week to week and, "Are we doing better or worse?"
So there's your KPIs and stuff that managers love, right?
Patrick: They certainly do.
To that end, I'm interested in understanding how do you demonstrate that what you're doing is working?
I guess two questions, how do you know it's working? And then, how do you demonstrate that to your stakeholders?
Knut: Now we start to have these super scientific numbers, pie charts and stuff.
You should never make pie charts, by the way.
But I think the most impactful and most important thing is the testimonials that we get, like the happy people that burst out in joy that they sold the thing because we helped them.
Those things are what I share with the team, and that we take with us.
Patrick: Yeah, that qualitative storytelling is a huge part of the role.
How do you do that? Do you just copy and paste stuff opportunistically, or do you have any alerts set up?
What's your toolkit look like?
Knut: It's mostly opportunistic copy/pasting in the channel called Press in the company workspace.
So, yeah. Also I select a few of them and have them in the slide on the weekly all-staff meeting.
Patrick: That's awesome. Are there any other more quantitative measures that you're looking at right now?
Knut: We are trying this MBP of this new promising product called Orbit .
Patrick: Tell me more.
Knut: But actually, we are. It allows us to keep track of what is happening in GitHub right now.
That's good, because a lot of what we do are actually open source and there's a lot of surface to start to keep track of what's happening there.
The activity that the community are putting into it is kind of neat. Not just "Kind of," it is neat.
Patrick: That's cool. Other than Sanity and Orbit, are there any other tools that you have really been enjoying working with lately?
Knut: Yeah, that would be spreadsheets.
Because I've been the only DevRel at Sanity for two years, and I've been the head of developer relations for two years.
Go figure. Now I'm actually hiring and building a team both for DevRel and support, which is super exciting.
But that means that I have to embrace the spreadsheet, the ultimate project management tool. I couldn't be without it, actually.
Patrick: That's how you know you're making the transition.
Knut: I'm being serious.
Patrick: I know, I love spreadsheets. All right, so you're building your team and you're becoming a spreadsheet master.
What are your plans for the next 12 months? What are the big things you're excited about and interested in moving the needle on?
Knut: Becoming a manager, and hopefully a great one.
Because a lot of what I'm doing and have been doing for these two years, I have to make it possible for other people to do, and in a great way that makes them and their audience happy.
This is what keeps me awake at night.
Patrick: What have you learned from your experience running the community that you think will apply to building a team?
Knut: I think being a bit mindful and conscious about how your behavior--
This is like modeling, if you are nice and if you greet people and are open and try to make a safe environment to ask so-called "Dumb questions" or "Noob questions," if you manage to do that it will resonate within the community.
I think that's probably the most important thing that I've been doing, and I think it shows.
Patrick: It seems like psychological safety is such an important hallmark of healthy teams and healthy communities.
What have you seen to be effective ways to engender that type of safety in the groups you've worked with?
Knut: I'm a beginner, I feel, in this field as well. I feel I have a lot to learn.
But a very simple thing that I've been doing is for everyone that is in start, I greeted everyone that joined with a "Hi, it's awesome to have you here. You should feel totally free to ask anything in the help channel." That was hard to keep up with when more people are joining every day, but we still have the "Hello, introduction" channel. I make sure to greet everyone who posts there.
I also automated part of this, of course, with a bot.
Don't know how effective this is, but I think also people are observing what kind of questions are asked and how we approach them.
So hopefully that will make for a safe environment.
Patrick: Cool. So, you've got 30,000 people in your Slack, and it sounds like people are asking lots of great questions.
What's on your mind as you continue to scale up the Sanity community?
Knut: We see that a lot of these people, they actually want to contribute back.
They want to make plugins, guides, whatever. It's not always that easy to figure out where those efforts should be directed, so what I'm working out now is to make it easier for people to contribute if they want.
Some of it is tooling to publish a plugin, or guides for how to make a guide. That's meta.
Also offering a space, a place for people to show off their work on our website. That's what I'm doing now.
Patrick: That's really cool. What has been your biggest challenge as a person building a community?
Knut: Again I guess this is a bit of a startup cliche, but to differ between what's urgent and what's important, especially when you are the first line of all the feedback and all the things people are struggling with.
It's easy to get lost in that stuff.
So, to take a step back like, "What is actually important that I'm spending time on?"
That's probably the biggest challenge.
Patrick: I think that's a great point. This question around prioritization of something every one of our community members thinks about and struggles with.
What are some tools or tactics you use to make those prioritization decisions about your own community?
Knut: I think it's more of a practice, so when you find yourself a bit lost in the weeds or busy in a way, that's the time you have to stop and think about what you're doing.
Write lists, and think about the prioritization of those, and talk with your colleagues and team.
Like, "Is this what we should be doing?" And so on.
Again, a bit of boring answer, but that's actually what's working for me. So, there you go.
Patrick: Do you have a productivity app or a to-do list app that you like to use?
Knut: Don't get me started. I probably, like many others, try them all .
A bit like the spreadsheets, I'm now looking at Apple Notes and I use Apple Reminders because that's the things that are there.
I can talk to my phone and there's a new to-do thing , so I haven't got time for doing the proper omni-focused stuff.
When I was a student, I had time for that.
Patrick: You used omni-focus as a student?
Knut: Yeah, I listened to all the podcasts.
Patrick: We talked about your biggest challenge as a DevRel and as a community builder, what's been your biggest success or your proudest moment?
Knut: I think my proudest moment is every day, because we have a couple of thousand people in the Slack community and only once I had to kick one out.
Because this person wasn't being nice, I expected that to be a lot more, but it hasn't. That's my proudest ongoing moment.
Patrick: That's cool. This is one of my favorite questions, what are you reading?
Knut: I'm reading Twitter now.
Patrick: A true DevRel.
Knut: I actually do, but for books I read a good one called Resilient Management by Lara Hogan, which teaches me all about management.
There's loads of management books out there, and this one is one of the good ones.
For fiction, the last book I read was Agency by William Gibson, and I read that while I was in San Francisco. So, that was fitting.
Patrick: That's the new one, right?
Knut: That's the new one, and I actually also was at the book event with the guy.
Knut: That was also great.
Patrick: That was on my calendar but I wasn't able to make it, unfortunately.
Knut: He probably predicted DevRel somewhere back in the 80s, right?
Patrick: Yeah, I think he did. He said "DevRels are unevenly distributed,"or something.
Knut: They are.
Patrick: DevRels are here, they're just unevenly distributed.
Knut: Most of them are working at Netlify, I guess.
Patrick: And Microsoft. So, do you think that developer communities are different than communities as a concept more generally?
Knut: They probably are, or else we didn't need this taxonomy, but good question.
How are they different? It's definitely about craft, just like the larger conversations about "What makes a community?"
I guess having this community that's tied to a product is different than having a community tied to something else, and I guess ours is more focused around building with Sanity. So you won't find long musings about the movie you saw last weekend, and so on.
It's more skewed against professional concerns, probably.
That doesn't need to be less or a bad thing, but it's another thing, a different thing.
Patrick: Yeah, Jed O'Bakan makes a distinction between communities of consumers, champions and collaborators to articulate the depth of focus and technical proficiency.
I think most product-focused communities are in the champions bucket, which was people that are focused around a specific toolset and leveling up on those tools.
Knut: That may be the case. I feel I could see all of these in our community in a way, but there you go.
Taxonomy is hard, right?
Knut: Always a difficult person tries to solve it.
Patrick: You mentioned earlier that developer communities seem to emphasize craft more so maybe than more general interest communities, what are some ways that you and Sanity tap into that innate interest from your developer community?
Knut: We have this, I guess it's an ideal or ethos even about structured content, how you actually should think about whatever content is in an organization .
That comes up a lot because we are trying to challenge some conceptions, and understandably so people are bringing old problems to the community and we feel this urge to educate.
We're walking that line between telling people, "No. You should think about this differently," and actually just trying to solve what I'm trying to do. Something in that area.
Patrick: I feel like you're holding back some really spicy take about content and content management.
So, if there's anything I haven't asked that you just really want to make a point about as it relates to CMS or headless CMS or content management, help me read between the lines here a little bit.
Knut: Yeah, sure. So you want to blow up on Hacker News?
Knut: If you're using Markdown as your content management thing, you're doing it wrong.
There you go. People are furiously typing on keyboards.
Patrick: So, you're telling me Sanity doesn't support Markdown?
Knut: It does. Begrudgingly. It's only text strings, but--
Patrick: What's your beef with Markdown?
Knut: At university, I had this sweet setup.
I was running Markdown, it compiled , I had a PDF preview. It was great.
But when you are taking over that project and you have to sort through endless Markdown files in whatever Markdown spec was on top of mind at the time that they built the thing, and validating YAML stuff.
It's not great if you want to use a modern web framework to redesign that site, or have a service augment that content in any meaningful way and have to parse HTML with RegEx or whatnot. It's not a great developer experience, even though some people have tricked themselves into thinking that.
I'm spicy now, right?
Patrick: This is great. Spicy Markdown takes.
Knut: Now I'm that person, that's great.
Patrick: This podcast is called Developer Love and we talk a lot about creating products and tools and communities that people really enjoy using and building with.
So, I'm interested. Knut, what's one thing that you're loving right now?
Knut: I guess that would be my wife, but I guess you mean professionally.
Patrick: You can have two types of answers, sure.
Knut: I guess it would be my vocation.
I love being a DevRel and all it brings with it .
To elaborate a bit, perhaps what I love about being a DevRel is the variation and all the things I get to learn.
That's the thing that drives me to learn new stuff, and DevRel is a great place to do that.
Patrick: That's awesome. I couldn't agree more. Are there any communities that you really enjoy participating in?
Knut: I think the JAMstack community is a great one. It's also quite friendly, I think, and quite open.
Patrick: Cool. Thinking as a community member yourself, whether it's JAMstack or DevRel Collective or whatever, you mentioned the JAMstack community is very friendly.
What are what are some hallmarks you look for as a community member? Like, how do you know a place is friendly?
Knut: I think everyone in those communities are not friendly .
We have people that are out to get you. I go to the orange site, like Hacker News or something and sift through the comment sections, there we have the opposite thing.
Being met with openness and curiosity and having people being interested and engaged in whatever you bring, that would be the indicators for a friendly community .
Patrick: Cool. Everyone loves Markdown, and yet you're telling me it's terrible for all these reasons. Is this really a sustainable position?
Knut: It's the right tools for the right things. I write a lot of Markdown.
Everyday in GitHub I will write markdown and that's fine , but if you are trying to actually capture the most valuable thing you have in our organization, which is your content and descriptions of what you are doing and the products that you're offering and what not, you should do that in a sustainable and resilient manner.
That's not Markdown. It's HTML, it's super great for browsers to look at and make sense of.
It's not super great for a toaster or a voice skill, or an AI augmentation service or whatever to figure out.
So, that's the professional take on that question.
Patrick: Can you talk a little bit more about content resiliency?
That's a concept you've mentioned a few times. It seems kind of interesting. What does that mean?
Knut: That means that it is easy and actually possible to know what content means, what the intent behind it was.
That may sound lofty in a way, but it's mostly a matter of the specification and the properties that you are giving it wherever you have it.
So, marking something within rich text as a thing. Like, "This is a footnote."
I can query that, I can know I can query my product descriptions for products that have more than five footnotes in the description, because you want footnotes in your product descriptions.
Stuff like that, or I want to know how many of my services has an image illustration where the flashlight has gone off when they took the photo.
Stuff like that is pretty useful when you are trying to keep your content tidy and nice and up to date.
Patrick: We've talked a lot about structured content and content resiliency, these are what seemed like almost philosophical constructs.
Is there something under the surface here that's driving your passion around this topic?
Knut: I think having been a consultant and having been in numerous projects where you are entering a situation, and you're trying to build something great.
You're trying to build something that enables their users to be successful, or get whatever they need to get them done.
Having a tangled mess of HTML in the way is frustrating, so I think it comes from that.
Also, having experienced all the creativity and innovation and possibilities that comes out from just having structured content, being able to mold it however you like it to a presentation, that's pretty powerful and great.
That's why I'm passionate about those things.
I'm not sure, I'm not completely sure I'm willing to die on the "Anti-markdown" hill yet.
I didn't expect this movement either, but there you go.
Patrick: There's this idea that we're trying to tap into with the Orbit brand, starting to tap into, this notion of "The people behind the pull requests."
In that there's lots of tooling and dashboards out there that will show you charts and graphs about code coming and going, but that's actually--
Code coming and going is a second order effect of building relationships with people and understanding them.
So a big part of what we're doing from a brand and product standpoint is trying to service those people inside, which is why if you look in the product we've put the people almost front and center, more so than just charts going up and down showing how many lines of code are added or deleted.
That concept seems to be in the water with the people we're talking to, so I think it's a pretty exciting time to be thinking about the space.
Knut: I think it's a great perspective, and it ties into what we talked about, facilitating inclusive communities and so on.
Other people that opines that code should be this nonpolitical space where you don't have to think about social equality and stuff, but it's people all the way down.
There's IDs and biases and so on, so putting more focus that it's people. That's great, I think.
But it's way more complex than monads or typed arrays or whatnot.
It's hard, so I can understand why people don't want to think about it. But having tools that make that easier, that's great.
Patrick: I just had a thought that code is actually inherently political, in the sense that it's--
When you're working with a team of larger than one there's influence and decisions that are being influenced--
Carried out and decisions being made based on things beyond just the objective realities, as if there is an objective platonic notion of "The best way."
Knut: If there are, you can bet that it's encoded in structured content.
Patrick: That's good. Knut, this conversation has been really enjoyable.
I've really learned a lot and appreciate your thoughtful and intentional approach to building communities, so thank you so much for coming on the show today.
Knut: Thank you for having me.
Patrick: If people want to learn more about you, where should they find you online?
Knut: I'm on Twitter. I have the worst Twitter handle for actually saying on a podcast, it's @KMelve.
You can also try to Google Knut Melvaer, but good luck with that if you're not a Norwegian native .
Go to Sanity.io and take it from there, I guess .
Patrick: Awesome. Thanks, Knut.