Disha - Reo.dev (00:05)
Hey everyone, welcome back to another DevGTM conversation. And today we're diving into something that every DevTool marketer eventually obsesses over, which is how do you build a content engine for DevTools? And to unpack that, I'm joined by someone who's probably seen more DevTool content programs than anyone else, Karl Hughes, founder of Draft.dev. So Karl, before we dive in, we'd love to understand about your journey, what are you doing at Draft.dev, who are you serving, and how did it all begin?
Karl Hughes (00:34)
Yeah. Yeah. Thanks for having me on Disha and happy to share. ⁓ Like a lot of people who got into developer relations or developer marketing, I was before this a software developer. So the first 10 years of my career. I ⁓ started with startups, I managed teams, I hired engineers and ⁓ most relevant to what I do today, I did a lot of picking of software tools that we were going to use, whether it's hosting or CICD tools or
IDEs for the team, productivity tools, all that stuff. So ⁓ one of the things I realized as I started doing that was that content played a pivotal role in my decision around what kinds of tools I would use with my team. And every year, that sort of intuition has been reinforced by data. Now every year, the Developer Marketing Alliance kind of puts out a survey on where is...
where developer marketing and relations teams focusing their efforts in content marketing, both written and video, are always at the top of that list. And so it just makes sense. so when, long story short, during the pandemic, I was kind of between jobs and I was thinking about what I wanted to do next. And I realized that this need that all these companies had around creating really good technical content was something that was just kind of always going to be there. And I really liked the
solving that problem. So what we do at draft.dev has evolved a little over the years. mean, in the early days, it was me and a couple freelancers writing blog posts. ⁓ And now it's a little more mature and we build, ⁓ like you kind of alluded to, we build these repeatable, scalable content engines for developer tools and developer platforms. so we can kind of dive in, obviously today we'll dive into what that means and what it looks like and how to do that.
But that's kind of where we are today. And you've mentioned we worked with a ton of the best brands and developer tools, everything from tiny startups just getting started up to large publicly traded companies. And so it's been really fun to see so much of this market and see how much has evolved in the last five years.
Disha - Reo.dev (02:44)
I think that's an amazing, amazing story card. And I think I love the fact that you've had this engineering background because I think it gives you a very unique perspective and a very grounded view of what actually resonates with developers. So I'm sure it's going to be a super interesting conversation.
Karl Hughes (03:02)
Yeah, yeah, definitely. think
it's very hard in this market to, it's very hard to do developer marketing effectively without some level of understanding of what the day-to-day looks like as a developer. It doesn't mean you have to be a software engineer, but you certainly should work closely with them and understand their needs and challenges. But yeah.
Disha - Reo.dev (03:21)
Yeah, I think that's fair. I think that's why dev marketing is so, you know, it's difficult. It's not easy. Great. So we would love to start with the fundamentals. I think every marketing team talks about building a content engine. But what does it actually mean in practice? What all does this include?
Karl Hughes (03:39)
Yeah. So I'll start with the results. The end result of a working content engine should be predictable, ⁓ new leads for sales to actually have conversations with or predictable new trials. It depends a bit on what your sales cycle looks like. But essentially, your content engine should be a known inputs into the machine should give you known outputs of
reproducible customer acquisition. So ⁓ that's what it should look like when it's done. Now it takes time to get there. And this is one of the trickiest parts about content. ⁓ Probably the number one pitfall that companies run into is that it takes longer and you need to produce more content than you think. ⁓ Very few companies, if any, in this space can get away with publishing a blog post a month, a thought leadership piece a month.
⁓ a tutorial here and there. It's just super hard to do that because the space is crowded. mean, developer tools are now a well-known market and there's a lot of people producing a lot of content out there. And so you've got to kind of get to a cadence that allows you to stay competitive. ⁓ when you kind of starting with the end result in mind, we want to get you to a point where every piece of content you publish, you know, will drive results in aggregate. And every month you should be getting
a certain number of page views that translates to a certain number of trials or sales calls booked that translates to a certain number of closes. And we want to use content to help push people down that funnel from that top and sort of just awareness building stage all the way down into consideration and buying. In practice, the way that this works is you start by building a roadmap of content that gives you a view at what it is we're
trying to do in the first three months. And I like to break it down into three to six month cycles, depending on the size of the company. For smaller startups, you probably want to start with these shorter three month cycles. For larger companies, it might make sense to go on six or 12 month cycles because you're not going to be able to change things quite as quickly when you're a massive company. But you want to take this three months at a time and say, all right, so what do we need to do to start to reach the target audience we want? And what are we optimizing for?
So there's two ways to approach this. ⁓ Almost every developer tool company comes to us and says, well, we wanna capture the ⁓ whole hosting market. Like in other words, we wanna be the new AWS or something like that, right? And we say, okay, that's great, but you're not gonna get there in a year with a budget of a couple million dollars. It's just not gonna happen. So where do we start is the question. And so the very first thing that you have to do when building out your content engine is really the first thing you have to do in...
defining any go-to-market strategy, which is narrowly define who that customer base really is and who you're really trying to get in touch with. It's going to change over time, and that's OK, but you have to have some working thesis to start from, or else it's too hard to, as they say, boil the ocean, meaning do everything all at once. ⁓ So the first thing is picking that unique target buyer and then building a content plan that can actually get pieces of content in front of them. And that includes everything from
the introductory or top of funnel pieces that kind of ⁓ introduce people to your solution, all the way down to the things that help them see how you compare to competitors and then actually implement the solution once they get to that point. So, ⁓ happy to break that down further, but I'll kind of leave us a pause there because there's ⁓ plenty of ways we could go.
Disha - Reo.dev (07:20)
I think it's a very refreshing thought that when you are starting with content, are starting with outcomes because I think that's something that I rarely hear people who are providing technical content talk about. You start with leads and then you work backwards as to how do you get there. So, Karu, we'd love to understand. were saying that there are different types of content you can create. So, how do you decide what content to create, how much content to create,
the frequency and what is the type of content that you should create. Does it depend on the size of the company, the stage of the company?
Karl Hughes (07:57)
Yeah, depends a little on the size and stage of the company, as you mentioned. It depends on how mature their content operations already are. And it depends on who their target persona is that they're trying to attract. So I'll give you some examples to make this more clear and kind of concrete. DIf you're a really big company and you've already covered all the ⁓ basics, ⁓ then what you're trying to do with content is continue to grow the top of your funnel.
awareness so that you can capture more and more ⁓ of the total addressable market. I'll give you, again, I'll take it really, really concrete. Several years ago, when I first started Draft.dev, I worked with Okta just as a freelance writer. And I talked to some people over at Okta at the time and the way their content program worked then, which it's evolved a lot. So this is not relevant and not, don't take this as how they do it today. It's totally, you know, I don't, I don't want to speak for them, but how they did it then was
They knew that every blog post they published would get, over the course of a year on average, it would get about 10,000 visitors or so. And of those 10,000 visitors, they knew the percentage that would convert to trial users of Okta, and they knew the percentage of trial users that would then convert to paid users of Okta. And they knew the lifetime value of an Okta user. Putting all those numbers into equation, my contact at the time said something like,
Disha - Reo.dev (09:18)
Yeah.
Karl Hughes (09:23)
Every blog post we pay you, Karl, $1,000 to write is worth at least $10,000 on the back end to us, and sometimes much more. Now, of course, sometimes a blog post doesn't do that well and it's not worth that much, but in aggregate, this content program was so successful for them at the time that they couldn't produce enough content. In other words, they would have paid to get even more if they could find it and keep it at the high quality. It's just hard to do. So, you know, when early on at draft.dev, that was one of my signals that this is probably a really interesting market to...
Disha - Reo.dev (09:35)
Yeah.
Karl Hughes (09:52)
to be part of, but also ⁓ when you think about what a really good content engine looks like, that's a good example of it because you've got a really tight ROI that's very connected to the ⁓ inputs that you're putting into the machine. Now, the other thing I'll say about this is when you're at that stage, you want to...
basically write as many pieces as you can. I think they were publishing multiple pieces a day most days and it was different types of content, lots of different things from tutorials and entry level things all the way up to deep dives into their product and releases, new product releases, just tons of content. And that's why they were so successful of it, partly is like they were able to invest a ton in it. On the other end of the spectrum, we work with a lot of early stage startups and the challenge they have is where do we start? Like we can't be Okta, we can't write.
you know, 500 pieces of content this year, we can probably write like, I don't know, 50 or something, like one a week, that's pretty good. So with them, what we try to do is figure out what, again, starting with the goals and persona, what are some unique angles we can take with this content that would be hard for the big competitors to replicate? And where can we start to augment what's already working? So again, I'll take it really specific here, I was talking to an early stage startup.
last week that targets ⁓ engineering managers and managers and directors. And for them, we don't need a high volume of content to be perfectly honest because managers and directors are a smaller group than just general purpose software developers or DevOps teams or whatever. So there's fewer of them and they were even more targeted into very specific size company. And so we could say, okay, let's deal with the, let's,
get enough information to say there are five to 10 kind of key problems that engineering managers and directors at that size company tend to face. All of our content is gonna be centered around those five to 10 problems. And we're gonna talk about different approaches to solving those problems. We're gonna talk about where your tool fits into the problem solution life cycle. We're gonna talk about examples of different companies that have those problems, industry examples where they have those problems.
We're gonna talk about customer use cases. We're gonna talk about like an individual implementation tutorial that shows how your solution actually solves those problems. And we're gonna compare your solution to two or three others on the market. And we're gonna produce different types of content that achieves all those goals. So some of these might be landing pages, onsite pages. Some of these are gonna be third party pages like things you publish on Dev2 or DZone or what have you. ⁓ And it's a combination of all those things. But basically,
Disha - Reo.dev (12:28)
Okay.
Karl Hughes (12:31)
What this allows us to do is build out a roadmap of the most high impact pieces of content we can produce ⁓ with the client for the first three to six months. And then from there, we'll take a look at the analytics and we'll iterate on that over and over again. So that's kind of the go-to-market strategy for content these days.
Disha - Reo.dev (12:49)
Can I also follow up question on that? So how do you figure out what are the top five problems that your ICP is facing? Is it like some research that you do or do you generally see your customers already have some idea about this?
Karl Hughes (13:03)
Yeah, it's a mix. So often the customers will have some idea because they've had some conversations with their customers or they did some market research before they launched and things like that. I mean, there's a reason they're building this startup and usually it's to solve a specific problem or set of problems. The other ways you can kind of come up with some problems that are relevant are using Ahrefs or SEMrush to determine keyword volume.
While SEO is ⁓ currently in this weird state where everybody's unsure of where it's gonna go because of LLMs, it's certainly still valuable for a data point ⁓ as to what engineers and developers and technical people are searching for. ⁓ So for example, ⁓ again, when you think about some keywords that engineering managers and directors are gonna look at, things like ⁓ developer productivity, that's.
certainly something like a problem space they're gonna be investigating. So we can look for keyword ⁓ traffic around developer productivity and developer engagement and developer ⁓ results as far as like measuring productivity and results and things like that. And so we kind of build out these maps of keywords to target. And similar with the LLMs, some of the tools now are giving us some insight into what kinds of queries ⁓ people are asking, chatgbt or Gemini or Claude.
around different spaces. That data is still pretty limited. So we're talking today in October of 2025 and my guess is it will get more access to that data as they open up advertising and some other kind of analytics and insight platforms in the next couple of years. But today there's starting to be some platforms that can give you some sense of that, like those search volumes and what kinds of queries people are using and things. And so all those can help guide your content as well as.
just the market research that you would hopefully do on your own anyway.
Disha - Reo.dev (14:58)
Understood. And then you also mentioned that the startup that you were sort of working with, they were targeting engineering managers. But in DevTools and specifically DevGTM, I've seen that there are a couple of different personas. One is your developers who are actually evaluating your tool and then there would be your economic buyers. So how do you divide content between the both of them? Do you target them together? How does that go?
Karl Hughes (15:21)
Very good question, super hard, it's hard. This is where, and it gets even harder, I'll just, I just had a call with another client that is in the open source space. So here's a really complicated problem for you. You have an open source tool with a user base who all wants free things, a lot of hobbyists and people doing this inside projects and things like that. And then you have this paid enterprise tool that's sold to CTOs and directors of engineering.
Those two audiences don't really cross over, unfortunately. Like, yes, the tools are fundamentally very similar and they solve similar problems, but the types of people who buy them are very different and the reasons they buy them are very different. The open source developer buys quote unquote, but they use this tool because it's free, they know how it works, they have a support community around them, and it solves their problem. The CTO though might buy this tool because it either saves them money or time somewhere in the chain.
or it gives them analytics or insight that they don't have into how the tool's being used or where it's fitting into their software chain. ⁓ It might give them more better security than they would have if they use the open source version. There's lots of different reasons, but again, now you've got this really challenging marketing question, which is, who are we building our marketing pipeline for? Is it for the CTOs, the economic buyer over here who wants to use the enterprise platform, or is it for these open source developers that want the open source community driven platform?
Or do we have to target both? Do we have to spend like double money basically and try to do both and have different, you know, it's super hard. So I wish I had like one answer for you, Disha. Like this is not, you asked a good question because there's no one answer. ⁓ What you have to do or what I try to do with clients is figure out where can we show some business results that really matter so that we can get buy-in to...
the content budget and tackle more and more of this. So in other words, ⁓ you know, often I'll be talking with say a content manager or a mid-level marketing manager and they're telling me about this problem and I'll say, if you went to your CEO, your founder or somebody, know, whoever your boss is and said, I want to get us more of X and that X was either these CTO leaders or whatever or these open source providers, which one would they clearly gravitate towards and be like, yes, get us more of X and it's like, let's go.
Disha - Reo.dev (17:32)
Mm-hmm.
Karl Hughes (17:41)
chase that first and let's take our first three or six months to really show that we can drive a little bit of that. And then we'll ask for the budget to go increase the scope and go aim at the open source side too. So you've got to just figure out the priorities and focus on what matters most of the business. So you can open up more budget to hopefully grow into the sort of the nice to have portion of the business.
Disha - Reo.dev (18:03)
I think that's fair. And I wanted to do a follow up on Okta as well. So how much time does it take for an Okta to reach that maturity wherein they know what each piece of content is actually doing for them? Is that really possible? Can you get that level of clarity? I understand that Okta did, but is there like a rare occurrence or is that something which you see happening with companies?
Karl Hughes (18:28)
So I've seen companies that commit to content very early on figure this out by the time they raise a series A or series B. It's rare and I don't think that's going to be something that every company should shoot for. But for example, I worked with a founder a few years ago early on at draft.dev. He knew content well. He was less of a dev tools person, but he was building a dev tool with a technical co-founder. He knew content marketing and SEO really, really well. He knew that it would work in the space. He just needed somebody to do the
like creation of technical content. That was where I came in with the team. But he wrote in the first year or so we worked together, at least, I think they published at least a hundred pieces of content. So that's two a week consistently ⁓ in this very technical space, ⁓ very niche kind of product and a very niche set of content pieces. did it all. He and I would do the research and come up with topics and all that stuff. And then the team would do all the writing there.
And very, like within that year or so that we worked together, I don't remember the exact numbers, because this was years ago, but it was, went from a standing start to at least 50 to 100,000 page views a month on the blog. And so those, that level of volume of readership, you can start to test like what conversions kind of work and how do we get people into paid, like free and paid trials or into sales calls. And so he could actually gather that data by the time he got to a series A stage.
Now, ⁓ that is again, exceedingly rare and I don't expect most companies to go that hard into content that early. Typically what I'll try to tell founders is look, you're in that first seed to series A range, like you're probably testing five or six channels, you probably need to spread your budget around, you're not gonna be able to produce 100 pieces of content in that first year. But let's try to really focus on your core pieces that are fundamental that we need to get.
regardless of where your company goes, regardless of how hard you go into content later. And let's try to build some of these high impact pieces out and kind of clusters of content out so that we can get something started and show some value early on. So yeah, the answer is it really depends. And then I've seen, again, like big companies that don't have any clue on the ROI of a piece of content. I mean, this is your world, Disha, with Rio. So I think you're.
Part of reason I like what you're building is like this is needed. ⁓ So many developer tools companies run into situations where they don't even have conversion tracking. Even huge companies we've worked with don't have conversion tracking. They don't have ⁓ basic like a real CRM that tracks who are their users and who are free users and paid users. They're just querying a database and hoping they get it right sometimes.
And there's a huge gap there with some companies. Now again, not every company, there's plenty that get it, but it has shocked me to see how immature some developer marketing operations can be even at large companies. And so there's a huge need for kind of better tracking and awareness that tracking is possible and a best practice, I think.
Disha - Reo.dev (21:39)
Uncle, you've worked across companies. You said startups and enterprises, etc. I'm sure open source and well-funded enterprises. So what separates the teams that actually succeed with content from the ones that struggle?
Karl Hughes (21:54)
Consistency over time. Like, there's this ⁓ kind of fundamental ⁓ belief that I have and that I've seen over and over again that you will get good at things that you do consistently. You won't get good at things that you do once a month. ⁓ So whether that's, ⁓ like, I actually believe this is true of lot of marketing. Almost any marketing channel can work. ⁓ If you're a developer tool, podcasting can work as a channel. It's hard, takes a while, but like everything is hard and takes a while, so it can work.
Content can work as a channel YouTube can work as a channel But you need to be consistent so that you can build the reps to actually gain enough experience to do it Well, there's no substitute for that now You can hire an agency to help you kickstart things or get you on the right track But fundamentally you have to devote resources to it Like even if you work with draft dev I'm gonna ask for some of your CEO's time to do interviews and figure out what the vision is and what things they're learning about the
the customer base, it's not even, and then you're gonna pay us money too. Like it's not free, it's not just an instant snap your fingers and we've got a content engine that works, it's gonna take time. Same is true of any marketing channel, but I think it's where most people go wrong in content as they expect. Well, we could write a couple of really good blog posts this month and that'll be it, that's it, we'll see how that channel works. And it's like, it just doesn't work that fast, unfortunately. ⁓ Now the advantage there for people who stick with it is,
because so few people stick with it, you can actually do really well on a relatively small budget if you just keep doing it. ⁓ But you might have to do it for two years before you really start to get a positive ROI on everything you publish. And that's where it is hard to convince people who are skeptical of it. ⁓ I think one of the big challenges in this industry is a lot of founders that start developer tools companies are technical and they under appreciate and under invest in marketing. And this is something like if
Disha - Reo.dev (23:44)
Thank
Karl Hughes (23:46)
You you had asked me 10 years ago, Disha, like, do companies spend too much on marketing? I'd be like, yeah, I don't see why anybody spends any money on marketing. You it's all about the tech, right? Like, I've clearly turned ⁓ my opinion on that because I've seen companies that do really good, like have really good technical products just fall apart because they never get it in front of the right people. And it's just gotten harder and more competitive in this space. ⁓ So yeah, your marketing has to be consistent, it has to be done.
Disha - Reo.dev (23:57)
yet.
Karl Hughes (24:16)
repeatedly over and over and so building up the ⁓ practice so that you can get really good at it is part of the goal here.
Disha - Reo.dev (24:25)
Understood. think being a marketer, is great to hear. Very reaffirming. So, Kaal, you also spoke about budgets, right? So is there like a percentage that you generally advise that startups should keep aside for content or does it depend on the category? Does it depend on where they are in the stage?
Karl Hughes (24:44)
That's a really good question. I don't have a benchmark of like, is how much you should spend. I've worked with startups who were 10 people that spent $20,000 a month on content. And I've worked with startups that were a thousand people who wouldn't spend $1,000 a month on content. It's crazy the difference there. ⁓ yeah, I think the way I would approach this, and I think I always caution, ⁓
startup founders like, hey, we don't need ⁓ a $20,000 a month budget to start to show some results in three to four months. There's certainly things we can do with $20,000 a month that we can't do with $5,000 a month, but there's a way to get something started to at least build your engine, your content engine, and your discipline to publish regularly, early, without a lot of money. And that's what I would typically, if a founder's skeptical or they're not really sure,
Like I would say, let's build some pieces that help you close sales and deals that you have today. And then we'll show you some of that value and then we'll try to scale that up little by little. So that's the way to approach it if somebody's skeptical. On the other hand, if somebody says, I know content is gonna work and I'm ready to invest in it, then you need to get yourself to a publishing two to four times a week as fast as you can. And that usually takes a budget from 10 to 20, $25,000 a month depending on the.
how technical the stuff you're publishing is.
Disha - Reo.dev (26:10)
Understood.
I think that's a fair benchmark because I understand there is no one right answer but you you've figured out. Understood. You also spoke about you know there's a lot of uncertainty because of SEO and because of AU etc coming in, LLMs coming in and I think how everyone consumes content specifically developers that is changing a lot. For example even product docs right they're not really going to product docs and reading you can just sort of query it now you're getting that in your ID so how have you seen?
Karl Hughes (26:15)
Yeah.
Disha - Reo.dev (26:41)
content evolve because of the change in which people are consuming content.
Karl Hughes (26:46)
Yeah, it's interesting. So on one hand, people aren't coming to your docs directly. And what's challenging about that for us as marketers is the tutorials we write now for clients, sometimes they don't get a lot of traffic anymore because instead of humans reading that, you have LLM scraping it and then regurgitating it back to people, which I don't know if that's bad. the end, the result is the same in a way.
Disha - Reo.dev (26:54)
Mm-hmm.
Karl Hughes (27:14)
but it really is harder to track. That's been the biggest shift. There's this kind of term going around called zero click, and we're in this zero click era, meaning that people are getting the answers they want without clicking through to search results anymore. And developers, we adopt technology earlier and faster than a lot of industries. And so I think this is especially true in the developer market. So behavior has certainly changed. It used to be two or three years ago, we built lots of content clusters that were focused on
⁓ top of funnel topics. remember at one period we wrote four versions of what is a data warehouse because that term data warehouse was so competitive that everybody wanted to have an article that spoke to what is a data warehouse. And now I would never, I would never suggest that a client write that what is a data warehouse piece because no one clicks on that. It doesn't really need your unique perspective unless you have some really unique perspective, but most of the time it was just like very basic. ⁓
But what you can do is ⁓ shift your thinking around this. in other words, trying to get clicks from every result in LLMs is a losing battle. They're not gonna give it to you. It's not gonna happen. But training the LLMs on the advantages and benefits of your product, your platform, and then how to successfully integrate your platform into other tools is super essential.
So it's harder to track, but it's even more essential than it was. I was just writing a LinkedIn post because this came up in a conversation. Somebody was asking like, yeah, ChatGpT does a pretty good job explaining our tool, but it seems to always favor this competitor of ours. And I was like, well, do you have a comparison page on your site that compares you and the competitor? And they're like, no, we haven't written about it much. We've only kind of had a launch post and that's it. I was like, well, your competitors have a bunch of comparison pages and one of them is you versus them. They've fed ChatGBT.
exactly what they want to tell it about how they're a little bit better than you and you guys are better for some thing down here that maybe they don't want to have anyway. And so my point is if you're not feeding the LLMs with what our tool is good for, why we're better than some competitors and how to integrate it, your competitors will do that. And now you have no control over what's coming back to people in LLMs. And that's kind of a scary place to be. So I actually think the urgency around content is higher ⁓ than it ever has been because
Disha - Reo.dev (29:28)
this.
Karl Hughes (29:39)
this behavior change is happening so fast where devs are moving over to LLMs to ask these questions instead of Google, instead of Stack Overflow. And we've got to be ready for that. And we've got to a lot of content that trains these LLMs or else they're going to pull our competitors content and use that to guide them. So, yeah.
Disha - Reo.dev (29:55)
Is there also, so while I understand that you need comparison pages and all, but for example, if you are creating a content for say devs, would there be certain differences because you would want LLM to be able to read it or does the same content work? What are some of the technicalities there?
Karl Hughes (30:12)
Yeah, so this is changing a lot every day. like I keep ⁓ it reminds me of 15 years ago. I worked for a company that did a lot of. ⁓ Did a lot of WordPress sites and we did we would do all sorts of hacks to get people better SEO right? Like that was a thing that worked 1520 years ago. You could hack Google to rank highly for current key terms. You could hide text in the footer. You could have these like.
big backlink farms that you generated automatically. It was just trash stuff that worked. So we're in that same, and it's not quite as bad right now, but we're in that same phase of LLMs where there's some hacks that kind of seem to work right now, but my guess is a lot of the doors will close to these things. So when I say this right now, this is kind of seeming to work right now. I would caution everyone in thinking like, if you listen to this in six months, it may not work anymore. So keep it in mind. There's a few things that matter though, or seem to help right now.
Disha - Reo.dev (31:04)
Mm-hmm. That's true.
Karl Hughes (31:10)
One is having accurate and appropriate metadata on every page in your site. Again, it's a simple thing. It's a lot of people have been doing as a best practice for SEO. So just like descriptions of articles, keywords, topics, things like that, interlinking between articles and ⁓ how your brand fits in, accurately portraying your brand is all important. ⁓ Another thing that seems to help right now is...
kind of putting summaries at the top of your articles and kind of key takeaways at the bottom in bullet point form. The LLMs seem to love citing bullet point, like summaries of things. And again, this is kind of like, you and I probably know enough about how they work to understand why they're essentially these text engines just trying to remember what combination of text should, it should output based on the input. they're not as smart as they look, is my point. So there's still these like little hacks that can kind of work. ⁓
Disha - Reo.dev (31:54)
But it's real.
you
Karl Hughes (32:03)
You know, Reddit has gotten a lot of talk lately, like everybody's like, oh, Reddit gets cited all the time. So we want to like go hack and throw ourselves into Reddit a million times. yeah, that works right now. But in six months, they're going to close that loophole. They're not dumb. They know these hacks are happening. So anyway, my point is there might be some things that work today, but I would be cautious to like implement them at wide scale and put a ton of money into them. What I do think, again, like fundamentally how they work is they're trying to gather
text data and accurately give answers back to people. That's hard to do and they're getting going to get better at this over time, but you've got to feed them good input data. So my best advice at a high level is feed the LLMs really good input data. ⁓ One way that this works right now, and again this may continue to work, is right on your site. That's great. Write blog posts, write tutorials, write docs, ⁓ write case studies, write comparison pages, but then also write on partner sites, write on third party sites.
⁓ Right on LinkedIn right on reddit and all these other places are giving the LMS more sources and context that feed it data about your product and so What we've gotten into doing with clients more than we had in the past is is basically building Content plans that involve a lot of external publication as well as internal publication That's a big shift from the SEO days when you know two or three years ago We we published almost everything exclusively on client sites and now it's shifted a lot towards, you know third-party syndication and publication as well
Disha - Reo.dev (33:28)
I think very exciting times because you know it's an evolving space.
You're just sort of figuring out what's going to work, what's working now, what's going to work maybe sometime in the future. I think very interesting times. CreateCast, I think we've discussed how do you actually plan, what are some of the topics that you should actually sort of write on, which persona to target, how much content to write, et cetera. I would love to understand some executions. So for example, once you finalize the topics that you'd like to cover, what actually goes behind creating high quality content
developers like any processes or any frameworks that you use.
Karl Hughes (34:07)
Yeah. So we the way we approach this, we've we kind of categorize each piece of content into a different type that we are going to and each different type has a different workflow. So, for example, thought leadership pieces, a lot of early stage startups, this is one of the fundamental pieces of content that you need some basic pieces that share your vision for the future. So.
⁓ If I'm working with a very early stage company and they don't have any content out the first few pieces we're gonna start to interview their CEO and founders to get our What we're building why we built it our past experience that led us to this our first customer use cases and stories ⁓ Integration guides that show how to use the product successfully And there's kind of a laundry list of things and each of those types of content has kind of a formula for how we produce it It's interviews tend to be like an interview piece or sorry
A thought leadership piece will be like an interview between us and their founder. And then that'll turn into a bunch of notes, which will turn into an article. ⁓ We use a blended approach, as most marketing teams do, having an AI with a knowledgeable human in the loop. You one of the reasons that we've stayed so focused on developer content is that all of our writers and tech reviewers are software engineers and developers and data engineers and engineering managers. And so we don't... ⁓
Disha - Reo.dev (35:22)
I just did.
Karl Hughes (35:31)
We don't rely on LLMs to be the subject matter experts. We rely on them to just help us generate, fill in the gaps between concepts and make our knowledgeable SMEs better writers, because they're not writers naturally. So like they need help with that. So that's kind of our hack, if you will, or just our method for doing this. So again, there's thought leadership pieces which are revolve around interviews. There's tutorials which will revolve around documentation and ⁓ user guides. There's ⁓
comparisons, which will revolve around feature sets and internal documentation. ⁓ And then again, mixing all those different ⁓ types of content and the formulas we use for them with knowledgeable SMEs to actually make sure that everything works. A big part of the process is quality assurance. And this has always been true. Even when we wrote everything by hand from the ground up ⁓ in the early days, ⁓ it's still always been a multi-part.
quality assurance process. we have like an engine, another engineer runs the code to make sure it works, test the tutorials and test the any code samples. And then a editor goes through and revises all that to make sure it makes sense and flows and sounds in the right voice and all that stuff. And then we have the client, you know, review and approve things and then get it published. So it's, it's important to, ⁓ it's important to have a high quality ⁓ production process so that you can get high quality output.
but ⁓ you can't let that get in the way of publishing regularly. And one of the big things I'll say about this is ⁓ a lot of people make the unfortunate mistake of, ⁓ developer relations team says, hey, I've got an idea for an article, they write it, and then they send it to their engineering team to review, and the engineering team says, well, this isn't how I would do it, so let's not publish it. And it's like, that is such a counterproductive attitude, right? This is like, ⁓ it's kind of the reason that...
Disha - Reo.dev (37:03)
Yeah.
Karl Hughes (37:25)
getting your engineers involved in that workflow can be ⁓ a mixed bag as far as like how it can be a little bit dangerous because they can really shut down ⁓ the publishing consistency because they're busy and they got this article thrown in front of them and they don't really want to review some random blog post and they have an opinion on why they should use TypeScript instead of JavaScript. It doesn't matter, know, that wasn't the point of the article. ⁓ So anyway, whatever the process is you build, needs to involve like
technical sign off and the articles have to be good, but they don't have to be perfect. It's really a mistake to let perfect get in the way of consistent publishing cadence. And that's where I do a lot of like pushing clients towards how do we get you on a weekly cadence? Even if not every article is perfect at the moment we hit publish, we can always revise. We can always come back to these things and refresh them in six months. We can always.
ask your engineer to go make their own edits. know, like if they really think this should be done a different way, write the article yourself and we'll republish it in six months when they finish it. It's like there's lots of ways to solve this problem and you just have to have the the discipline to consistently publish regardless of ⁓ your process being high quality and having a lot of people involved. anyway.
Disha - Reo.dev (38:39)
Understood. And right now, one piece of content, how much time does it take for you to produce it? Is there an average that you see?
Karl Hughes (38:47)
Good question. ⁓ It depends a ton on the type of content. And then we do a ⁓ lot of things to try to get as much runway out of each piece of content as we can. So for example, ⁓ it might take eight to 10 hours to write a really good tutorial. ⁓ That said, once you've written a really good tutorial, you've tested the code, you know it works, you know ⁓ all the documentation has been correctly cited.
Disha - Reo.dev (38:51)
and.
Karl Hughes (39:15)
⁓ It's actually pretty easy to go build two or three other versions of that one. So for example, you write that first version for how to integrate our tool in Python and blah blah blah. And then you go ask an LLM to convert this Python into Ruby and then write the Ruby version of it. Now you have to have another engineer test that and actually make sure it works and make sure they didn't miss anything that's Ruby specific. But it's way less work than eight hours. You're now talking two or three hours. So we can take one tutorial and spin out three, four, five different versions of it.
in the same amount of time it took us to write the original one once we have that source material. So we do a lot of things like that to help our clients generate more content ⁓ in an efficient way rather than try to do every piece from scratch, you know, like you've got a handwrite, you know, and do all the discovery and all the testing, you know, from the ground up. So my encouragement to people is like,
try to figure out ways to either use content in multiple ways. ⁓ So maybe it's again, take that tutorial example. We wrote up, you know, how to integrate this tool in Python and ⁓ on a Linux server. And now let's do one that's like a Python and a Windows server. And let's do one that's Ruby and a Linux server. let's see, you know, and again, you just iterate through these cycles. And then you can publish versions of those on third party sites like.
write a different version that's slightly altered for Dev2, write a different version for D-Zone. And again, the LLMs can help you with a first pass at it, and then you can get a human to test it. Super efficient to get a lot of content out that way. We do a lot of things like that to help our clients get to a publishing schedule of two to four articles week without spending quite as much money as you would if you had to write everything from scratch.
Disha - Reo.dev (40:53)
So basically repurposing and making sure that whatever you've written, put thought in, you're making the most out of it. How do you repurpose it in different ways and make most? I think that makes a lot of sense.
Karl Hughes (41:06)
Yeah,
another good one that we haven't touched on is refreshing content. in our space, content gets out of date within six to 12 months almost all the time. So you don't have to keep every single piece you've ever published up to date, up to the minute, you know, correct. But you certainly want some of the core fundamental pieces to stay that stay correct and updated. And so with bigger clients, often half the publishing they do is just republishing older pieces that needed a rework. And that's lot less work than publishing.
scratch. there's a lot again there's a lot of ways to get creative and keep your publishing cadence consistent without having to publish purely net new content every single week and I think that's where people people get too lost in like how do we publish every week it took us eight hours to write one blog post I'm like yeah that one blog post though has three or four offshoot pieces we need to pull out of it and so let's do that and then we can get ourselves up to publishing two or three a week without nearly as much difficulty as it would have taken otherwise.
Disha - Reo.dev (42:02)
I think that's fair. And now that you know you have content, think the next big piece is how do you actually get developers to see it? So what are some of the most effective promotional playbooks that you've seen and are there specific channels that work really well for DevTools?
Karl Hughes (42:17)
Yeah, the wish, the dream we all have with our content is that you will just write it and it will magically attract people, right? And SEO used to, to be fair, SEO kind of worked that way if you gave it enough time. And it still does. I mean, like we've always practiced what we preach at draft.dev and we've always published consistently and our blog drives new sales calls all the time. So I know it works because we see it working in our space.
Disha - Reo.dev (42:27)
you
Karl Hughes (42:47)
⁓ So what I'll say is like yes SEO does work but it takes 6, 12, 18 months to really pick up steam depending on how much you can publish and how much you can attract attention. So what you need in the short term is some way to drive early traffic so you can test things, so you can test CTAs, you can test lead magnets, you can test ⁓ the type of content that is resonating with your audience, what's attracting the right kind of user, etc. ⁓ And so there's two ways you can do this. One is the org.
Organic approach where you basically try to you know share this content across many different channels like slack communities or discord communities or reddit subreddits hacker news like You could pitch it to newsletters all this stuff that all works and like it is hit or miss because some pieces will Like you share a link on reddit and might get either down voted or blocked like that's gonna happen sometimes
Disha - Reo.dev (43:41)
I
Karl Hughes (43:41)
You
share in a community and they say, this is overly self-promotional. You can't put this here and like whatever that happens, but it certainly can work. You can use the founders organic LinkedIn posts, although LinkedIn devalues, you know, linked off content. So you probably have to write like a summary of it and then link to it in the comments, do all this hacky stuff. Like that all works. So that's certainly one way to do it. I will say that rarely do you get massive bumps or big viral hits off of that kind of strategy. If your strategy is purely organic, it's just going to take time. That's the...
Disha - Reo.dev (43:59)
Yeah.
Karl Hughes (44:11)
pure and simple answer. If you can, on the other hand, if you can pay to get certain pieces, like say your high value pieces, your ⁓ really high impact landing pages or lead magnets ⁓ promoted in things like paid publications, newsletters, places like that that already have an audience of hundreds of thousands of visitors, you can drive thousands of actual users or clicks or downloads, depending on what your metrics are.
in three months easily. that's something we do all the time with clients. know, set aside a three to $5,000 a month budget on promotion and you can guarantee hundreds of thousands of impressions on your content. Whereas if we're just gonna write the content, I can't guarantee you any number of impressions in three months, it's gonna take time to build it. So I think that's something that a lot of people miss these days. it didn't, you know, it used to be easier ⁓ to just get.
Disha - Reo.dev (44:56)
Good.
Karl Hughes (45:05)
more viral hits and to get stuff in front of people through these little niche communities. But now everybody knows that is a thing and so everybody does it. It's competitive. There's no free pass here anymore. So you really do, if you wanna drive results in a shorter time period, pay for them. If you're okay with waiting, then wait. And that's, you get this direct trade off of time versus money. And you just gotta make that call. You can't get it both.
Disha - Reo.dev (45:30)
For paid promotions, are there any channels that you would really recommend? So you mentioned newsletters. Is that something that ⁓ gets a lot of results? ⁓ Is Stack Overflow also, you run ads on Stack Overflow? I'm not sure. Is that something that you do?
Karl Hughes (45:47)
Yeah, yeah,
we don't directly. So the kinds of promotions we run with clients are through publications and ⁓ newsletters. We don't do any with Stack Overflow yet, although I think that's certainly viable. ⁓ Reddit has paid promotion as well. I haven't tried that personally. I hear mixed things and it's probably just, my answer to this is a lot of it comes down to who's your audience and where are they and are we actually at a channel that they're spending time? ⁓
You know, if your audience is CTOs, are they hanging out on Reddit? I don't know. Like CTOs at Enterprise? Probably not. Are they on LinkedIn? Probably. Are they like watching, you know, are they going to trade events? Probably. Like there's places where they probably are, but you'd need to do some research and talking to figure that out. So one of the things that I encourage clients to do as we're getting spun up with them is ask your five to 10 customers that you know best or whatever, your kind of archetypical customers, ask them,
what kinds of places they're hanging out online? Like are they in subreddits? Are they in Slack groups? Are they going to live events? Are they not doing any of that? Like what are they doing with their time to sort of further their own learning and development? that the next obvious question is can we get in front of those places or pay to get in those audiences? ⁓ We put on our website a big list of like developer newsletters. There's hundreds of them and they're all different little sub niches and
subscriber level. like, there's a lot of them out there. You just need to find the places that are most appropriate for your audience. ⁓ you know, I think newsletters are hugely underrated form of getting in front of developers because there's such a high trust medium. Developers don't say yes to putting something in their inbox lightly. And I think it's really cool to get in front of that group like that.
Disha - Reo.dev (47:33)
I think such a beautiful and simple hack, know, ask, figure out your top five sort of audience and figure out where they're hanging and then you know, double down on that. Any DevTools that you know, you look up to who's been creating good content and sort of distributing it beautifully that you see?
Karl Hughes (47:51)
Ooh, that's a good question. ⁓ I've always taken a lot of inspiration from DigitalOcean. Their writing program back five years ago was top of the heap, and it's different now. I don't know as much about it internally, so I'm not sure. I think they still do some things really well, so I think they're good. ⁓ We have done work in the past for a company called Fly.io. I feel like they've always done really good content. Their founder,
He's in Chicago as well. I've talked to him about that. He would always write these really detailed, long blog posts in the early days. And they were just like very esoteric, but very fun to read and interesting. ⁓ you know, it's funny because those both the, you know, those examples are very different. Like DigitalOcean has always been a high volume, ⁓ lots of search focused traffic and ⁓ lots of long tail traffic building over time.
whereas Fly.io has been low volume, very focused and very niche and nerdy in a way, ⁓ but both are successful. So I think there's something to be said for this. There's a style thing here that not one approach always works for every company. of it ⁓ is ⁓ your style. But what I'll say both of them do that it comes back to what I've always emphasized with clients is they're both consistent. They haven't stopped publishing. They haven't stopped.
producing and refreshing content. They haven't stopped ⁓ putting ideas out there in the world. So, whichever your style, wherever your style falls for your brand, that's fine. But do it consistently and do it for years and eventually it does pay off. But you just gotta hold on and keep going.
Disha - Reo.dev (49:34)
Understood. I think a beautiful sort of note to end the session, Karl. I think this has been really, really insightful. I'm sure a lot of companies which are struggling with content will get a lot of inspiration and understand how important content is and hopefully, you know, ⁓ invest in content and see the results. Thank you so much for sharing. Yeah, thank you. Thank you so much. I think this was this was super amazing.
Karl Hughes (49:55)
Yeah, yeah, well thanks, Tisha. It's always fun to talk about.
Great. Well, thanks. Yeah, thanks for having me. I love doing this kind of stuff. I'm happy to share more, Karl at draft.dev if anybody wants to follow up with me too. ⁓ And yeah, thanks for having me on.
