Adam Gordon Bell on Attracting Customers through Blogging

Michael: I’m interested in what your experience was like writing at Earthly.

I know when you joined, Earthly was only about four people. And very quickly, Earthly’s articles started getting on the front page of Hacker News, and you were the author of most of those articles.

Within Earthly, did you have to campaign to have the time to write those articles and invest in things that would do well on Hacker News? Or did that come from within Earthly?

Adam: I was a backend engineer who got a job at Earthly in developer relations — didn’t really know what that was but figured I’d be talking at a bunch of meetups. Tried talking at meetups, and it didn’t go super well.

We had an investor, Mitch Wainer, who was the marketer behind DigitalOcean. And he was like, “Yeah, just write things about the problems that your potential customer has. And try to really solve their problem and not necessarily be marketing,” …which is a weird thing for a marketing expert to say.

Michael: Right.

Adam: But that led to the writing. And then some of them were a bit more silly and did well on Hacker News. And then I took your course and then kind of leaned into that.

I think because it saw results pretty quickly, and people get excited when their company’s on Hacker News, so it wasn’t too hard to make the case that this was valuable.

Michael: Were there any false starts? Or did it did it pretty quickly get traction on Hacker News?

Adam: Yeah, I think I wrote things that were more directly like, “Here’s why you should be using Earthly.” And we made tutorials and example apps and lots of things like that.

And I don’t know if our goal with those was to really have them do well on Hacker News necessarily, but I would share them on Twitter, or we would post them on reddit. Anyways, none of that ever worked, I guess.

I mean, it was useful for people who are using Earthly and wanted to know, like, “How would I combine a JavaScript solution with a Golang solution?” Well, here’s an example.

But our problem was people didn’t know who we were, right? We were competing with obscurity.

Adam: I think one of the big things was that in your course you were like, “Oh, yeah, this blog post — I worked for 40 hours on this.” And I was just like, “Wow.”

So like, I’m nowhere near where this needs to be this valuable.

Michael: Yeah. I think one of the things that a lot of people miss is can spend two or three hours on a bunch of blog posts, and they all kind of flop, and you’re like, “Oh, blogging just doesn’t work.” And it’s hard to realize that it would work if you put more time into it. It doesn’t always have to be 40 hours, obviously.

I think there’s a power law distribution where getting to Hacker News on the front page at all is like orders of magnitude more readers than just missing the front page. And same thing with first page of Google results and being number one versus being number 10. I think it’s hard to will yourself to do that.

But then the payoff once you’re at that level ends up being much higher, like every incremental hour you put in, like once you’re close to a high-quality post, you’re getting paid off much higher.

Adam: The power law thing is hard to internalize.

We made hundreds of blog posts over time at Earthly, and I’m sure that the top 10 of them are responsible for the lion’s share of people knowing things, which is wild to think about, right?

Michael: Do you remember the first big success on Hacker News at Earthly and what the reaction was?

Adam: So, I could have it wrong, but I wrote this post that — I think it was called like “YAML and Other Stupid Programming Languages,” and it was kind of relating to Earthly because there’s all these build solutions where you write all this YAML and you’re basically programming, but in YAML.

But the funny thing was that people don’t necessarily read the things you write.

And so as soon as I posted this, there was just like a million comments that were like, “Yeah, well, it’s not a programming language.”

And then other people would chime in and be like, “Well, actually no. If you read this, he’s saying something more complex.”

And it took off just because… it was basically a reading comprehension test. And everybody was failing it.

Michael: Right.

Adam: But then after that happened, we had somebody message us who was interested in Earthly, who had found that blog post. And they were like, “Oh, what’s this?” And they were a pretty big company.

And so we were like, “Hey, this is a thing.”

It’s hard to think of a less business-oriented thing than writing an angry rant about YAML, but it was bringing business to us.

Michael: Right.

And so how did you decide what topics were going to match the kind of customers? Was there strategy in figuring out, “Okay, somebody who would agree with us about how annoying YAML is is somebody who would like using our product?” How did you figure out from the topics what would attract customers that matched you?

Adam: It’s easiest to write — especially if you’re writing something that’s explanatory — it’s easiest if you don’t know it well. And you’re kind of teaching yourself as you go.

Michael: Right.

Adam: In Earthly, you had to do a lot of things using bash and using the shell and gluing together these tools.

It’s not like I spent a long time as a Unix system administrator. So, I knew that awk was a thing, but I didn’t understand what it was, except people said it was useful. So, I’m like, “Okay, let’s explain that.”

So, that was certainly helpful. If there’s something you don’t understand, and especially if there’s something you don’t understand that you have to use, and you’re pretty sure nobody else really has mastered either.

Michael: Yeah, I think that’s often a good source for ideas.

You tried to find resources on it, and you couldn’t, and so you can just be like, “Okay, what’s the resource that I wished that I found 20 hours of tearing my hair out ago?”

Adam: Yeah.

Michael: Were there topics where you tried to post to Hacker News or topics that didn’t even get to the point of publication?

Adam: Yeah, there’s lots of things that were like posted to Hacker News and didn’t really take off. One of the key things I learned was just to get a lot of feedback and spend a lot of time editing.

So, I would write things, like — okay that snarky thing about YAML was funny, so I tried to write an article where it was like — what’s the word I want? Like, it was funny by me being directly straight about, “I’m going to solve this problem to this solution.” And then the way I built it was just using complicated YAML on top of YAML and templating YAML.

I shared it internally, and I built some horrible tool and nobody really got that it was a joke. They were just confused about what I was building, and I was like, “Okay, well, I don’t think this works. The sarcasm is not coming through.”

Michael: Right.

Adam: So there’s lots of that: getting feedback and seeing what people thought.

Michael: What was the feedback process like within Earthly?

Adam: I would just write something, and then either I would write it directly in Markdown in our static site generator on a PR branch and then the CI would make a preview URL on Netlify, and I would share that and get feedback that way. Like I would just message a couple of people, and they would provide feedback.

It was actually easier to get feedback if it was a Google Doc because people could just leave comments. The formatting was sometimes an issue with like, “I want to show code stuff, and then people don’t understand it.

Michael: Yeah.

Adam: But the thing that I found very valuable was I was always asking people for feedback and then the harder part was I would always try to take action on their feedback.

I think there’s a — maybe it’s just me — but like somebody’s like, “Oh, I didn’t like this part,” and you’re like, “Well, no, but that part’s good.”

And so I always had to remind myself: if I’m asking for somebody’s feedback, I should actually be taking action upon it.

And so that was hard because oftentimes I just… didn’t agree or — it’s hard to get out of your own head sometimes.

Michael: Yeah. I think that’s a hard balance because sometimes — I guess it’s like knowing your reviewer?

Because I think sometimes people give notes just for the sake of they wanted to give a note, and they maybe haven’t really thought through the advice.

And so sometimes you do have to reject the note and just be like, “Hey, I think you’re the weirdo. This actually is good.”

But I think if you’re doing that too much — if you’re just rejecting everybody’s notes and saying like, “No, everything I write is gold!” then you probably also have a problem.

Yeah, it is something that’s difficult to learn: which feedback to take and which feedback to- when you have to be kind of confident in your own instincts.

Adam: I know Michael Lewis was saying that he gets a lot of feedback and that he — trying to think how he phrased it… that they’re always right about something that they point out, but they’re usually wrong about what’s actually the problem.

Michael: Yeah.

Adam: He takes it as like, “There’s a problem here,” but the person might say, “Oh, you should do it this way.” And he ignores that. He’s the writer. He knows what he’s doing.

Michael: Right, yeah.

Adam: Something drew their attention.

Michael: Yeah. I have that a lot with Grammarly, where Grammarly will flag something for me and be like, “You should write it like this.” And I’m like “That sounds stupid.” But the fact that you flagged it is making me take a second look, and usually I’m like, “Yeah, that’s not the best way to phrase it.”

Adam: Yeah, I’ve had that for sure.

Michael: Are there any tools for writing that you found useful?

Adam: Yeah, I did use Grammarly. I had like a ChatGPT prompt I would use somewhere. I found it pretty helpful for editing.

So there’s that book. It’s possible I heard about it from you, but it’s like On Writing Well by William…

Michael: Zisner[sic], I think.

Adam: Yeah. So, he has a lot of things about removing unnecessary words or overly complex words.

So, I have a prompt that’s something like, “Hey, pretend you’re him. Go through this sentence by sentence and make me a list of all the words that are unnecessarily complex or the sentences that could be simplified.”

I found ChatGPT is actually pretty good at that. It’s providing specific editing advice.

But you also have to ignore it, like just like we were discussing. Sometimes it will be like, “No, you should say it this way.”

Michael: Right, that’s interesting.

And do you have any rituals around writing itself? Is it like you have like a specific hour of the day when you write or you do certain things to put yourself in the mindset for writing?

Adam: I think my main ritual would be I have YouTube Music on my phone and I plug these headphones into that and put on something that’s instrumental because the words distract me.

That’s the main thing. I don’t have too many rituals around it. But I hear that’s very important for some people.

Michael: Yeah. Did you find interruptions from coworkers would interrupt flow? Like if you would get an email or get an IM?

Adam: I don’t know. I could just ignore those.

Michael: Yeah.

Adam: Dealing with my own interruptions was more of a struggle. So internally-driven…

I feel like there’s this thing where you’re trying to figure out a way to say this thing. And then you’re like, “Oh, I should go grab a coffee or something.”

It’s like I hit a hard spot and my body prompts me: maybe you should do something else. “Have you checked the mail?”

I think those are more of a struggle for me.

Michael: Yeah, mine is I always have the intense desire to clean my keyboard when I’m trying to write something hard. Like, “It’s really important that my keyboard be really clean. right now.”

Have you found any techniques to overcome that?

Adam: No, but I forget where I heard it, where somebody was saying that basically if you hit a certain level of mental difficulty, it can prompt you to be like, “Oh, I should do this other thing,” which seems more immediately rewarding.

Once I had that idea that like, “Oh, it’s not actually about checking the mail. It’s that this is hard and I need to just think about this.” That was helpful.

Michael: Yeah. I think awareness is really helpful too — just knowing that that’s happening and knowing, “Maybe I don’t really need to check the mail. Maybe that’s not really the most important thing right now. I just feel that way because I’m doing something hard.”

And I think you can train yourself to get more comfortable with thinking hard, especially for writing.

Adam: Yeah.

And editing… it’s harder and it’s easier. In some ways, I find it takes less concentration, but in other ways it’s very boring.

Michael: Do you find the first draft more fun than the edit? What’s your favorite part of the writing process?

Adam: The people telling me that it’s awesome.

Michael: Yeah, that’s the same for me.

Adam: Everything between, no — yeah, the initial ideas are pretty fun and titles are super fun.

Michael: Oh really?

Adam: Yeah.

Michael: Oh, I get stuck on titles.

Adam: I always feel like no matter how much time I spend coming up with the title, it feels like you could do more. It’s actually so important. Especially when you’re talking about on Hacker News, it’s all people see: this little title.

Michael: Yeah. I was curious about one of my favorite titles of yours was “Green Vs. Brown Programming Languages.”

I feel like it’s such a catchy title, and it’s so evocative. Even before reading it, you have that graphic, right? Before you even get into the text of a brown field and a green field and it’s like, “Oh yeah.”

Even before you get into it, it’s like I know what this is about: older programming languages vs. newer ones. And I feel like, like that kind of phrase just sticks in my head.

Does that just come to you? Or do you spend a lot of time thinking of — what’s the… it’s not exactly a soundbite, but I think certain authors are really good at introducing a phrase into the lexicon.

Do you have a process for that? Or do you just kind of get lucky and be like, “Oh yeah. That one really works.”

Adam: No, but I think it’s important. But I don’t know where that one specifically came from.

There’s two things at work there. One is: if you can put together a concept and explain it to somebody — especially if maybe they didn’t see it before, but it was there. That’s very powerful. There’s like, “Oh, here’s this thing. Maybe you didn’t notice, but here’s this thing.” Super powerful.

And then you give it a name. And that holds together. And if you do it really well, it can become like the canon. I feel like there’s canonical things covering topics, and if you are able to like coin a phrase and it becomes something that’s used like that, you’re now part of the canon of software.

I don’t know if I’ve succeeded at that, but I’ve made attempts.

Michael: Yeah.

Adam: And then the other one is: I love when a nonfiction book — because I like to read a lot of nonfiction books, but I love it when they can give you a title that, once you’ve read it, you’re like, “Oh, the title is actually the summary.”

Joel Spolsky had his hiring book called, Smart and Gets Things Done. That was a great example. You could read the entire book, or you could understand these are his criteria.

Michael: Right. Yeah, I love Joel Spolsky and he’s, he’s one of the first people I think of as somebody who’s really good at, at coining phrases like that, that just always stick with you.

Michael: I was also interested in the article, “When to use Bazel?” That did really well on Hacker News, and it was a pretty atypical article as far as what succeeds on Hacker News.

You interviewed seven different experts about Bazel, which is a system for building complex monorepos. It was derived from Google’s — I mean, obviously you know this — but it was derived from Google’s internal Blaze system.

Earthly is, in some ways, a competitor to that.

What was the process of interviewing those people and putting together the article?

Adam: I posted on Twitter that I was looking for people who had done migrations that I wanted to talk to them. And then, I had like a Calendly link. And then I just interviewed them. I have experience interviewing people for my podcast.

And then actually writing it took me a lot of time because I couldn’t figure out how to structure it all. And so I ended up printing out transcripts of everybody and circling sections and then grouping it.

Michael: Oh, wow.

Adam: It became very complex. And then I threw that away and I was just like, “Well, what’s the most interesting and important things people said?” and grabbing them here and there.

Michael: I’m just imagining you like a serial killer wall- or how a detective has red string…

Adam: No, it got a little wild, but it turns out it’s actually better to just read through and find the most interesting parts that people say. And then just throw those all in a markdown file and then start, “Okay, how can I make this into something people would read?”

Michael: But the themes appeared to you as you read them?

Adam: Yeah, exactly.

I think that one reason it did well is that it is high-effort, and it’s apparent that it’s high-effort.

And I think another reason it did well is because it contained people’s stories about software migrations. Like taking down their CI and moving to this new system and having it take a long time.

And people really like stories, but also the stories don’t exist out there in the world. There’s a lot online about build systems, about Blaze, about Bazel. You’ll be hard pressed to find somebody saying, “Hey, we did this, and it took us three years, and it worked in the end, but at what cost?” or whatever the story is, right? “It was awesome.”

Michael: Yeah, you’ll see like random comments on reddit or Hacker News of like, oh, we used Bazel and it was terrible, but you’re not really getting the context.

I think the thing that’s really interesting about your article is you’re setting up who these people are, and they’re very credible. A lot of these people contribute to Bazel, and they have a lot of hands-on experience. They work at these very well-known companies.

And so there’s a lot of context behind their opinion. And it’s not just a throwaway comment. These people have clearly spent a lot of time thinking about the technology and what went well and what didn’t.

Adam: Yeah, and it’s also the reason I read Hacker News a lot is oftentimes for these comments.

Like somebody’s like, “Oh, there’s a new Kafka feature launching.” And then there’s somebody in the comments who writes two pages about this three-year Kafka migration, and you’re like, “I want to know more. I don’t know what happened at this place.”

Michael: Yeah. But what do you think the incentive was for the people who you interviewed? Why did they want to talk to you?

Adam: So, when I started podcasting, I did guest interviews for The Software Engineering Daily podcast, and then I was going to start my own podcast because he was like, “I’m the host. I don’t know why I need you.”

But I was like, “Okay, well, I’m going to do it on my own.” And he’s like, “Yeah, go for it.” I was like, “But how do I convince people to talk to me?” He’s like, “Oh, people just like to talk about themselves.”

Michael: Yeah, I think that’s a really good lesson.

And especially knowing that not only that you’re going to talk to them, but you’re going to do something useful with the result. It’s not like you’re going to hear their story and just write it in your journal. You’re going to feature them and make them look pretty good, too, setting them up as this expert in this technology.

Adam: A lot of people don’t know, actually, journalists will often audition people.

So they will reach out to, say, a whole bunch of people who want to talk about a certain topic and say, “Hey, I’m writing about whatever — Biden’s mental health. What are your thoughts?”

They may talk to lots of people and throw most of them away because there’s that one really interesting person who has an interesting perspective on it.

I’ve never gotten good at that, but I think it just shows that you shouldn’t be afraid to reach out to people and see what they have to say.

Michael: Yeah, the other thing I thought was interesting about that post was that when I read it, I was like, “Oh, this is interesting that they’re writing about a competing product, and it’s not in bad faith.”

Like you see a lot of companies that say like, “Oh, hey, we have product X. We took a look at our competitors… or we make calendar software and we rank the best five calendar software. And wow! We’re number one!” And all of our compliments to our competitors are kind of backhanded compliments.

And I thought it was interesting that when you wrote about Bazel, it felt very fair and even-handed.

Was there a concern at Earthly that you would make it look too good? How did you come up with the tone of presenting Bazel in a positive light?

Adam: It wasn’t hard, but I think it just speaks to being small and our communication strategy or something — I don’t know.

I think that the longer a place is around and the more insular it is, you can be like, “Oh, our product is the best!” and there’s kind of an internal - everybody ignoring other things.

I feel like because I got to set the direction on what our communication strategy was, and I felt like it was really good because I remember Gavin who was marketing, he had written this thing for this feature that we were building. It was a self-hosted platform and, in it, he was like “Hey, here’s the thing: at home I have this Windows server that I stream things from and whatever, and I have some open-source software that I run on it, and I’m never gonna pay for any of it because it’s just like my home thing but you know, I work at a tech company and if they want to know like what’s a great thing for streaming this, then I already have experience with this.” And so, hence our strategy.

You can use it free for home use, and the secret that usually nobody says is: we’re just hoping that you work somewhere, and eventually they’ll adopt this.

I felt like we have this strategy of just “let’s speak directly to the person.” Everybody knows this when we’re giving away a free product — we’re hoping that there’s some way that it gets adopted, but nobody’s saying it.

Michael: Yeah, that’s a really interesting strategy because Bazel is so complex that nobody’s using that for their home server. But somebody might use Earthly for their home server.

And so if you can get them to pay attention to Earthly, there’s a lot more people with home servers than there are making product strategy decisions for companies that are at the scale where they would need Bazel.

Adam: Yeah. Yeah, I just meant we weren’t afraid to talk about competitors or about what our go-to-market strategy is for pricing.

Michael: Yeah, yeah, that was something in my business as well.

I feel like people, in general, are very cagey about details, and I feel like you gain so much more by being open. People trust you a lot more, and they feel like they can relate to you a lot more if you’re just open and honest about like, “Here’s when you would use our product.” Like, “Here are the scenarios where you would use our product, and here are the scenarios where a competitor would be better.”

I think people are so used to seeing the way giant megacorps compete with each other. Google is never gonna say, “Oh, here are some situations where you would use an iPhone instead of us.”

Adam: It’s funny, right?

You are really good at it. You are constantly open about what’s going on, like the sale of your business. It has huge value.

Michael: Oh, thank you.

Coming back to blogging in general, are there any turning points in your blogging career that influence the way that you write or the topics that you write about?

Adam: Yeah, there’s this Austin Kleon book about writing. I forget what it’s called. Maybe Show Your Work! or something like that? Anyways, it’s super good.

But the thing he always said was — or maybe he had also one like… Steal… Artists Steal or something.

Anyways, that you, should find things you like and try to make your own version of that. I found that very powerful.

I don’t have to invent a new style of communicating. If there’s just things that I find speak to me, maybe I should try to adapt that style. I found that very powerful.

I learned through the podcast that people’s stories have a lot of resonance, like, “This happened, then this happened, then this happened.” Whenever you give somebody an objective, and then things get in their way, people pay attention.

I learned from you to put a lot of work into the content.

Those are some of the big lessons that come to mind.

Michael: Were there things that you learned to avoid?

Adam: Yeah, I wrote an article once about wanting people in Zoom meetings to have their cameras on. And it did take off, but like just everybody saying how stupid it was. And like, “Just because my employer pays me, why do they have the right to see into my home?”

It really generated a negative reaction, and then I learned that… I guess the tone of my thing was a bit like shaking my finger. So, I learned that that wasn’t a great idea.

I learned that tutorials don’t really work for that type of viral content but they can be very powerful just for Google searches.

Michael: But going back to the webcam one, I think that’s an interesting lesson because that’s something I’ve found as well.

There’s certain kinds of posts that can get a reaction, but maybe it’s not the reaction you want. Like, maybe it becomes popular, and you just don’t really feel good.

I’ve had posts like that, where the reaction is very combative. I wrote this article that was critical of Stripe, and I showed how their library was, kind of doing more tracking than they let on. And the reaction on Hacker News is one of the most negative reactions I’ve gotten, where everybody’s like, “You’re an idiot!” Like, “You don’t understand: Stripe is doing that for good reasons because they’re doing fraud protection.”

And I was like, “I know why they’re doing it. I just still don’t like that they’re doing it.” But it was such a different experience from, like, “Oh, I sold my company,” and it was like, “Yay! Great! You sold your company!”

I do think it’s important to call out bad behavior, or if you have a strong opinion on something, I think it is really useful to do it, but know going into it: this is going to be a fight. And are you ready? Are you going to feel good if this turns into a big fight?

Adam: Yeah, yeah. No, it’s true.

That one surprises me that there was a negative reaction because usually, big company doing something bad seems like it would succeed.

Michael: Hacker News loves Stripe. I think it’s because the founders are so active.

For people getting started with their blog, what advice would you give them?

Adam: Make a list of writing that you like. We had a channel at Earthly called #great-writing or something, and we would post them here and there. And then just try to do it yourself.

And it’s actually a really good way to get that initial ramp up to see what it takes to be at that level is to be like, “I really like this piece, and now I’m going to write something like it. And when I finish, let’s go back and compare and see, okay, well, maybe what I have isn’t quite there.”

That can be a way you’re almost getting a feedback loop, but internally generated by just comparing yourself to an external source.

Michael: Do you remember examples of articles you’ve done that with?

Adam: Yeah, so there’s this article that’s called, like, “A Brief and Incorrect History of Programming Languages [sic].” And it’s hilarious. It’s just made-up stories about programming languages. I think it has the years right, but the backstory is just humorous.

So, I wrote something called, like, “Mostly Incorrect DevOps Glossary,” because there’s just all these DevOps terms like, “cloud engineering” and whatever. And I just made them all up: “A cloud engineer is just a developer who writes Go.”

And so it was like can I make something as funny as this history example?

Michael: Oh, nice. Yeah, that’s a really interesting technique. I’ve never tried that, but I see how that would be a good exercise.

Adam: Yeah, and also there’s that… works for Fly? Used to work for Tailscale…? Xe – she writes with narrative bubbles, like her –

Michael: Oh, yeah.

Adam: Yeah, I tried to imitate that once when I was writing an internal engineering thing. I interviewed two of the engineers that worked on it and kind of interspersed their dialogue with me, describing the feature. So, that was me trying to copy her.

I think that it’s a powerful technique. And you can take inspiration from more than one place, right?

Michael: Yeah, definitely.

Adam: If you copy one post, people might be like, “Oh, that’s like that!” But if you take inspiration from two, they’re like, “Oh my god! It’s something new we’ve never heard of before.”

Michael: Yeah, I think that’s such a powerful thing. I think people really underestimate the value of being good at coalescing information.

I don’t think I’m especially good at coming up with original ideas, but I think I’m pretty good at coalescing the good parts of a few different sources.

And yeah that often gets a good response because it’s like, “Oh, you saved us the trouble of scouring the Internet and finding these three good ideas. And you put the good parts in one place.”

Adam: Yeah, totally. And you have your cartoons. I think those are solid.

Michael: Oh, yeah. Yeah. Yeah, I think lots of things where it feels like a signature move that you do, I think it will help you stand out.

Adam: Yeah.

Michael: Cool. Thank you so much for talking, Adam.

So, you host the podcast CoRecursive, which I really like. I’ve been listening a lot the past week, but I’ve been listening for years.

Adam’s website is adamgordonbell.com, and he’s on Twitter at @adamgordonbell.

Is there any other place people should find you?

Adam: No, that’s probably it. But, yeah, write things. It’s fun. I think you’ll enjoy it.

Michael: Yeah. Okay, cool. Thanks, Adam.

Adam: That’s my advice.

Michael: All right. Good stuff.