Disha Agarwal - Reo.Dev (00:00)
Hello everyone.
Welcome back. So today we're digging into how developer marketing is evolving. Over the past couple of years, developer marketing has really gone through a pretty big reset from how teams invest to how developers are actually discovering and evaluating tools, what channels are working, know, everything has gotten a reset. And today we're joined by Karl Hughes, the founder of Draft.dev, who recently ran a developer marketing survey to actually understand what's happening across DevTool teams today. And we'd love to unpack the insights, talk about what was most surprising.
And also translate this into takeaways for anyone from the DevTool team listening to us today. Karl welcome back. So good to have you again.
Karl Hughes (00:38)
Hey, it's good to be back, Disha. I feel like we just spoke. I mean, it was probably six months ago now, but it was a lot of fun and I'm glad to be back to talk about some new stuff we're working on,
Disha Agarwal - Reo.Dev (00:47)
Yeah, and think the last conversation was so insightful, Karl, so really looking forward to this one as well. I would love to begin with understanding what was the goal of the survey and why did you decide to run it now?
Karl Hughes (00:57)
Yeah, there have been companies in the past that have done developer marketing and relations surveys. And I noticed in the last couple of years, some of them were not doing them as much or maybe didn't release the answers publicly. And I kind of just realized like, this is good information for us to have and obviously to share. So with draft.dev, because we do content for exclusively for developer tools companies, we work with a ton of developer marketing teams
So we're in this nice position where we get to talk to these people every week or month, and we figured, why not ask them a few questions? the way we did this was send out emails to a lot of our former clients and current clients, former prospects, people we've had sales calls with, people that are in our network that we've just met along the way. We did this in person at KubeCon as well, so some people listening may have seen us there. I know we got to meet in person at KubeCon, which is fun.
So anyway, it was a mix of things that we wanted to ask about. But essentially, just like you said, Disha, this industry is going through a lot of change. AI is very much a big part of every conversation and every product is being positioned around AI now. And so ⁓ that means a lot of the marketing and channels that people are using have kind of shifted and budgets have shifted.
We like to get a sense of that and then we like to share it because then other people kind of know, we on the right track? Are we following what everybody else is doing? Are we doing something completely different? Is that good? Is that bad? I don't know, that's up to them to decide, but we can at least get the data and send it out there.
Disha Agarwal - Reo.Dev (02:18)
Understood. Makes sense. And before we move to the results and our findings, we would love to understand the lens here. So who participated in the survey? Was it small companies? Was it big companies? What kind of DevTool teams did you had the opportunity to talk to?
Karl Hughes (02:31)
Yeah, fortunately, because this is such a ⁓ specialized market that we work in, we work with teams of all sizes. So we had several, I mean, ⁓ over a dozen respondents, I think, were in that like big corporate 2000 plus employee range. And then a lot of them were in that mid-size company and then a long tail of smaller startups as well. ⁓ Some of the findings we broke out by those different categories where there was an ⁓ of interesting difference in how spend or how
⁓ how budgets were allocated or how people were focusing things, which was interesting in some ways. ⁓ And in a lot of ways, a lot of this stuff becomes kind of universal as you look at it and you realize that there's not that much different in the way that bigger companies do things. It's just maybe a sense of scope and scale that they have that smaller companies don't. ⁓ yeah, breaking it down by size was helpful.
We didn't do, and this would be a bullet point for me next year, ⁓ quite as much breaking down by audience or the type of product that the company had. And so we didn't wanna pry too deeply into what industry specifically do you work in, although I will say that given the base of
people we talk to, a lot of them are in pretty deep tech. So a lot of data and DevOps and platform tools, Kubernetes tools, things that are pretty deep and technical.
Disha Agarwal - Reo.Dev (03:49)
Make sense. And in terms of size, understand it's a broad cross-section of DevTool companies, not just focused towards early stage or big enterprise.
Karl Hughes (03:57)
Yeah. And another thing we asked was about like how up the chain is your decision maker? Because some of these tools are developer first, very bottoms up, you know, they want to attract lots of individual developers who will then bring it into their boss later or maybe buy their own small subscriptions. And other tools we talked with were very top down. They had a C-suite type decision maker that they were trying to attract. And then everything else they do in marketing was kind just to enable that. So that that probably does.
that's going to impact your marketing channels and your mix for sure. But at the same time, a lot of the fundamental core things still stay pretty relevant, you know, with some exceptions.
Disha Agarwal - Reo.Dev (04:33)
And when you looked at the data, what were some of the 2-3 themes that really jumped out immediately? And anything that really surprised you?
Karl Hughes (04:39)
Yeah, so I'll start off with like last year, 2025, when we didn't do this survey formally. This is our first year of actually making, like pulling all the data together. But when we did this anecdotally, we did it a little more informally with clients. So in 2024 to five, when we did this, budgets were very tight and everyone was nervous. Just again, like not 100%, but was much more the vibe was we're gonna wait and see. Part of it was in the US, we had a new administration coming in with a lot of changes around tariffs and that was like,
causing a lot of uncertainty. was also, know, the LLM sort of Zeke Geist was at its peak maybe in early 2025. And there was a lot of talk about what that was gonna do to this industry and everybody was trying to reposition around that. A lot of marketing teams were getting pressure to use AI to basically replace themselves and spend as little money as possible on marketing. And, you know, over 2025, what I think people realized was...
You know, some of the work that LLMs can do absolutely will help you in marketing, but it isn't a replacement for good cohesive strategy or very strong senior marketing talent. And so that that fear kind of started to go away. Tariffs kind of normalized and seemed to be in a place where we kind of know what's going on. ⁓ Then ⁓ there's there's just been a little more confidence overall in markets and in tech. So that may change as the course of this year goes on. But as we surveyed people, the.
Why I mention all that is that one of the bigger findings was that 62 % of teams are increasing their developer marketing budgets next year, which is pretty significant. And the rest, a lot of those were either staying the same or uncertain, and only a very small percentage, under 10%, were saying they knew they were decreasing their budgets at this point. So that's a good ⁓ sign as far as the healthiness of this industry and of this space. I
And then if you're a marketer, it's obviously, it's good to know that your competitors are gonna be spending more money next year. ⁓ And just keep that in mind, if your plan was to cut budget and all your competitors are increasing budget, you're gonna be fighting an uphill battle.
Disha Agarwal - Reo.Dev (06:36)
So a lot of positive sentiment and a lot of ⁓ propensity to spend when it comes to developer marketing. That is what came across.
Karl Hughes (06:42)
Yeah,
that was one takeaway. And you asked about others. Another big one that I was really interested in was the usage of AI in marketing, because it has completely changed the way we work at draft.dev, to be candid, ⁓ and in a good way. I mean, it's gotten us to where we can produce content faster. We're helping ⁓ keep up with data and analytics more easily. It's automated some of the parts of our sales and onboarding process. I mean, it's been huge.
Now it's taken some work to get there and a lot of experimentation, but my general sense from talking to a lot of people was that ⁓ AI is still, LLMs especially are still in this nascent phase. So 85 % of the people we talked to were using AI for things like content ideation and maybe first drafts at things or helping them clean things up here and there, but they were still pretty skeptical of its ability to completely like, you know,
work across the whole marketing stack. And then another thing that was really interesting was a lot of these companies that we talked to, they didn't have a standard set of rules or ways to use AI. So for example, there might be five people in the marketing department. One of them likes chat GPT and they're using that. The other one likes Gemini and they're using that. Another one likes Claude and they're using that. So there was a ton of like people using different tools within the same company and none of it was really like officially sanctioned, which
has a couple of implications. One is it's a little bit scary if you're worried about security and data, know, and internal plans getting leaked, you know, through some of these tools because LLMs sometimes use this for their own trainings. You have to be careful with that if you're not aware. And then the other one is that companies aren't necessarily giving people how to use these tools yet. They're just telling them, hey, we got you a chat gpt subscription, go wild. And that's not really that helpful for a lot of people who aren't like,
Disha Agarwal - Reo.Dev (08:31)
is next.
Karl Hughes (08:36)
maybe like me who aren't geeks, they're exploring these things every day. So anyway, my takeaway there is if you're a marketer and you are using LLMs and AI tools, really start to lean into them and train your team on them because it's gonna make your whole team more effective. And if you're one those people who's not learning these tools, you're just like, well, chat gpt didn't do a very good job, so I'm not gonna use it. ⁓ You might wanna take a second look at it. ⁓ The agent tools are getting better, the LLMs are still...
They still have limitations, but there's ways you can use them productively. So anyway, I encourage people in this space, like there is a lot of mileage we can get out of LLMs and AI, but it does take a bit of work still. It's not the easy button. It's not something where your company is gonna give you a subscription to some AI tool that's gonna make writing a thousand blog posts super easy and high quality. It's gonna take a little bit of work on your side to manage. anyway, something to think about and something that I thought was really interesting.
in looking at this data.
Disha Agarwal - Reo.Dev (09:33)
Makes sense. I think totally agree, Karl, because I think every every LLM or every AI tool has has its own strengths and weaknesses and it takes some time to figure out what it is good at. And, you know, if you're letting each of your member figure that out, it's just time wasted, right? You would rather, you know, have this collective understanding of what you're using what tool for and then everyone is doing that. And I totally understand on the AI workflows as well.
setting up an end-to-end workflow because that's tough. That's not an easy thing to achieve.
Karl Hughes (10:00)
Yeah,
yeah. Another thing that stood out to me too was the comeback for events is huge. And even last year, there was still little bit of skepticism around events, live events in person. There were still a lot of events that were kind of on hold or not coming back to full force like they were. But when we talked to people this year, 73 % said that they're increasing their investment in event marketing this next year. And I think that's pretty telling.
73 % is a lot and it was even higher or it was the highest in the corporate enterprise. So if you're a smaller company, just know that there's gonna be more people, more larger corporates out at events. If you're a big company, know that a lot of the big guys are also gonna be spending more at events. If you're an event, ⁓ like person who running events right now, it's a good place to be. And it's ⁓ true, I think what's happened is there's more flooding of the market with digital.
marketing and advertising and digital content. And so what that means is it pushes people out to where is there less competition So I think it's really cool to see them coming back. And I hope that that kind of trend continues for a bit until it levels back out.
Disha Agarwal - Reo.Dev (11:11)
Understood. And why do you feel that the importance of events is back? I think you mentioned that one is that we're coming out of COVID. Second is since there's so much digital clutter, it just always is better to connect in person. ⁓ Is there anything else that the teams are getting offline which they struggle to get online?
Karl Hughes (11:29)
Yeah, mean, I think a couple other things that are all related to those two points. One is during COVID, a lot of companies went remote and a lot of engineers built their lives around working remotely. it's nice to see people in person coming back is actually kind of fun. And so some companies have certainly done the return to office thing, but a lot of engineers have stayed remote since this. mean, I'll say anecdotally among my friends,
Almost all of them moved out of the, you know, center of the city. Only one or two still have an office they go to, especially in startups that we work with. Very few have an actual physical office space. So I think people, we'd like to connect with other people, right? And this is like a human thing. And we like to connect with people who are like us that we get to learn from and meet in real life. And so events are a way that that's coming back and an excuse for people. Even if you work remotely from a small town, you can go meet your
your coworkers, can go meet your community members, you can go meet the vendors you work with. It's just a, it's a unique thing. The other thing that, you know, we didn't include this in the survey, so this is more anecdotal, but a lot of people talked to me when we talked about events, they started talking about micro events as well, which could mean anything from getting five or six directors of engineering into a room together to have dinner, all the way out to like doing meetup groups.
And I used to be a big fan of meetups here in Chicago. I was a part of several, I ran a couple. ⁓ And I'm glad to hear them coming back, because they had a really hard time, obviously, during COVID and even last couple of years. So companies are coming back and sponsoring meetups, starting their own little community meetups and events. More companies are doing the sort of world tour type events, like mini conferences that they do at lots of cities to meet with customers. So anyway.
⁓ All that to say that events are back. There's a lot of different formats people are trying. They seem to be very strong from an ROI perspective as well. They rank among the top two channels for ROI, measurable ROI in developer marketing. They probably always have. ⁓ So anyway, it's a good channel and I like that they're coming back and it'll be interesting to see how they continue to evolve.
Disha Agarwal - Reo.Dev (13:34)
Understood. Any DevTool team with limited budget listening to this, do you have any advice on how they should think about events strategically?
Karl Hughes (13:42)
Yeah, that's a good question because they can be expensive. mean, getting a corner booth at KubeCon is like $100,000 or more, I think. You guys had a smaller one, so I'm sure yours was not quite that much, but it's expensive is my point. ⁓ And when you're a smaller, we've had some of our smaller clients that are Series A startups maybe ⁓ have booths at conferences, and some conferences can still be worth it. ⁓ If you think about it, it really just depends a ton on who your audience is and what the
buying cycle is like and how many customers or prospects you have at that event that you can set up calls with or set up meetings with in person. So when you're a small company though, you want to leverage, you basically want to get the most bang for your buck as we say, but the most leverage out of every event you do. So if you do one and you have to get a booth to be a sponsor, also try to get a speaking slot so you can get in front on stage and talk about the product or the advantages it has. Also try to set up meetings with current customers.
partners, anybody you can in person, a lot of our clients and you guys included, but a lot of our friends and clients there were having their own little side events too. I think that's, again, you got to kind of do it all because the expense of flying your team in to go to this one in-person event and buying all the tickets and buying the booth, it's pretty high. And so you want to get as much mileage out of it as you can.
The other way to do it is to try with like smaller local events that might be cheaper, you know for the price of a couple boxes of pizza if you're in San Francisco you might be able to get 20 or 30 users together and at least ⁓ you can kind of get some people in person to give you feedback and ⁓ build stronger relationships with so really depends on your product, but certainly you want to think about how do we do this cost effectively and then how can we measure its impact over the course of time and I think this is something that
I'll kind of, well, we can talk about ROI and stuff later because that's a big conversation point and there's a lot of things there.
Disha Agarwal - Reo.Dev (15:39)
Understood. No, Karl I think agree. You know, once you meet offline, ⁓ there's a different kind of connection that you can make with the person. So we also experienced this at KubeCon. The people that we were able to meet offline, we've had such strong relationships with them wherein we've been able to involve in a lot of initiatives with them. So I do understand the importance of events and it's great to see that it's making a comeback.
Yeah. And you also mentioned that, know, events is one of the top two channels. So, you know, zooming out based on the survey, which other channels seem to be working right now?
Karl Hughes (16:10)
Yeah, so for ROI, developer marketing has kind of always had these, what I would call two foundational channels that continue to be the top of the list for ROI, for investment, for consistency, for predictability. And those two are community and event stuff, and then content marketing. Content marketing has many different shapes and sizes. could, anything from SEO and top of funnel type stuff, now AEO or GEO, ⁓ LLM optimization, essentially. ⁓ It's also helpful in
product adoption, so anything down mid to bottom of funnel when people are starting to compare their options, content has a huge role. We've talked about this. there's just so many different ways to use content that in reality we should ask like 10 questions to clarify what do you mean by content marketing because it's a huge category. those two categories have always come through in these surveys that I've seen with us and others as the most kind of cornerstone foundational places to invest. For smaller companies, content is great because
It's ⁓ very low cost to get started. You can literally just start writing content, writing content on LinkedIn, writing content on Dev2 or whatever, and get it going at least. ⁓ For bigger companies, their footprint becomes so huge. In other words, the kinds of things their product touches on and can do and the types of audiences they need to serve that they just have to keep pushing more and more content out to make sure that they're covering it.
You never want to be a big company doing an enterprise sale and then realize halfway through that you don't have any collateral that speaks to this client's language or framework they're using or something like that where they just immediately cross you off the list because you didn't have the support and collateral to help their engineers adopt the product. So it's just a cornerstone thing and it has all different price points and scale you can do. And again, same with events. And this is kind of like, we're extending that same conversation.
So those were the two that like everybody basically, a large percentage, think it's something like 60 plus percent are investing in those and seeing positive ROI. ⁓ The next tier down are like the situational channels. So this would be social and word of mouth kind of marketing. ⁓ So social media influencer type stuff Paid marketing tends to be in the second bucket of like,
situationally very good. And then email, newsletters, and kind of that long-term relationship building through life cycle marketing, that was also in this kind of situational category because some products, they really benefit from building up their own internal community or newsletter. Other products, it just doesn't make sense. If it's a ⁓ more enterprise solution,
engineers may or may not want to actually sign up for a newsletter about it. It's just not, may not be engaging enough or something they care about. So that's kind of the next category and all those score somewhere in the like 25 to 35 % of people say they get a good ROI off of it or it's one of their best channels for ROI. So that's like, you know, kind of middle of the pack. And then you've got these experimental channels, which influencers and external social community members.
Disha Agarwal - Reo.Dev (19:13)
Mm-hmm.
Karl Hughes (19:14)
and ⁓ long form video and multimedia, both experimental. A lot of people we talked to were putting a fair bit of their marketing, ⁓ experimental marketing budgets at least, into those channels. But ROI was very hard to measure. So both cases, the problem, as you might imagine, is like you put a video on YouTube and let's say you get 10,000 watches. How many people came to our website because of that? I don't know, because sure, some of them clicked through the description and clicked a UTM link, which we can track.
But the majority don't. They go to the URL and they type it in directly, which is really, for marketers, tough. Like now I don't know where that came from and I'll never know unless they tell me. You can, you know, obviously you could try to fill in some gaps and guess, but my suspicion is that video will always be hard to attribute. And that's just because of YouTube being the dominant channel and we don't have full access to visibility on it. ⁓ Same with influencers though. And again, I don't think it's a bad channel. I think it can absolutely work. I've talked to several people who've said,
this has been great. But a couple of things have happened in the last couple of years. One is there are a fixed number of meaningfully influential developers out there who actually have audiences of 100,000 plus followers that speak to the specific technical niche that you're gonna be in. And those people are getting gobbled up right now by the really big spenders, the big companies that can spend 20, 30, 50 grand on a series of videos. again, earlier stage starts really hard to get that same
ROI when you don't know your whole funnel hasn't been developed yet. So bigger companies seem to be like raising the price essentially on the really top tier influencers. And then, you know, if you're a smaller company, what that means is you have to go reach out to the people with 10 or 20,000 people in their audience. And that's OK. But now you have to cobble 10 or 20 of them together and now you have to manage them. the ROI like it gets more expensive to manage. So while again, we didn't have a ton of granular data on like
where influencer spend goes or how much influencers cost. I think that would be really interesting to try to get next year ⁓ if people can share it because I'm very curious about this and I do think it's a channel we'll see more of as content continues to get easier for LLMs to produce. How do we stand out? We've got to find people who have their own followings. So that might be a place to just keep our eye on for the future.
Disha Agarwal - Reo.Dev (21:35)
⁓ That's such a great framework, Karl, because it really makes you understand how to look at different channels in terms of ROI So thank you so much for sharing that.
So one thing that you mentioned is making a comeback and is giving decent ROI as social and it's becoming one of the more serious channels. What do you think is driving that shift?
Karl Hughes (21:55)
Yeah, that's a really great question. I think a couple things. ⁓ One is consolidation of social platforms. There's kind of fewer of them in a way, and they're getting more well known in developer marketing. So I have calls with people in developer marketing all the time, some of them, ⁓ it used to be a few years ago when I would have these calls, people would say, what's Reddit? Like GitHub, like we need a GitHub. And like they didn't even, they'd never heard these terms.
But now more and more marketers are coming in with an understanding of there's Discord, there's Slack, there's GitHub, there's Reddit. They may not know all the nuance on how to use these yet, but they certainly know they exist and they know those are channels developers use. So I think there's just familiarity with what are the channels. There's kind of been some consolidation around them. And then to things like you guys are doing, the traceability is a little better than it used to be ⁓ for some of these things.
Now will say that at the end of the day, social is often ⁓ like one of the touch points. It's rarely, and this is true in just broadly in developer marketing. There's almost no channel where it's single touch point, they bought your product. Like that's just not the way this works. Like developers take a while to make these decisions because they're usually very impactful decisions. I always like, you know, think about it this way. Like I wouldn't go ⁓ buy a new refrigerator just because.
Disha Agarwal - Reo.Dev (23:01)
this.
Yeah.
Karl Hughes (23:14)
Somebody wrote a really good blog post on a refrigerator unless I was already in the market to buy that thing, right? Like it's a so a developer might read your great blog content They might read your great LinkedIn posts and they now know your brand exists But until they have budget in a real need to actually go replace their existing solution You kind of just have to wait and continue to build that relationship So again, most developer tools require five to ten touch points and over the course of 12 months to really get something sold
Disha Agarwal - Reo.Dev (23:19)
place.
I'm
this.
Karl Hughes (23:40)
And so social is a great touch point and it may be hard to measure that like, yeah, somebody bought because I post on LinkedIn, but here's what happens. Say you have a sales call or a developer likes your product, they follow the founder or they follow the sales engineer they talk to on LinkedIn or they connect with them. And then that person is just posting content every week or two that shares more about the product, how it works, some use cases, some wins, some things like that.
and the person following may never be doing any interacting, may never really be thinking about this much other than the moment they see that post and they just heard from their boss last week, they've got to get a new edge hosting solution in place by next month, great. I've been talking to this guy, I've been hearing this stuff on LinkedIn, I know who he is. So that's why social is so powerful is that it is a great way to stay in someone's feeds. Same with X, same with, it could be TikTok. Again, I don't.
⁓ The short form video is still a little experimental, but I certainly think all these things are gonna be like the way that we have a touch point on the developers and build the relationship and trust for the moment that they're rated by.
Disha Agarwal - Reo.Dev (24:42)
Understood and within social are there any specific channels that are really working well? So you mentioned that know, Linkedin is a great way to stay top of mind because you keep showing in your prospects feed So whenever they actually have a need and actually have a budget they think about you But any other channels that you see are sort of gaining traction
Karl Hughes (25:00)
Yeah, it's very product dependent. So for example, if you're an open source tool, really want to look at, mean, you got to focus on GitHub. Like that's the place where your biggest advocates are going to be hanging out. They're going to be asking support questions there. They're going to be adding PRs. Like that's your probably your bread and butter. ⁓ For a lot of, again, like let's say more bottoms up developer tools where we're trying to reach a large, you know, large pool of developers and hands on keyboard people, X seems to be working fairly well.
Reddit can work in certain communities. ⁓ And then for the more enterprise or more top-down sale tends to be more LinkedIn maybe. But again, these are all kind of just broad tendencies. I would say there's nothing in the data that tells me you can't market a developer tool that sold directly to developers on LinkedIn at all or that has no value. I think it probably does. So it's probably more about finding where your current and
future customers are already hanging out and just testing different social channels. We typically want to, we'll start with the ones that we think are most relevant or that our client already has the most following on and then expand from there as it makes sense and as there's budget and time and resources to do it.
Disha Agarwal - Reo.Dev (26:15)
Make sense. So ⁓ we do understand that content is one of those channels that will continue to give good ROI and people continue to invest in. ⁓ But how should DevTools think about optimizing their efforts on content now with AEO coming in, with AI coming in? How do you feel that this should change?
Karl Hughes (26:32)
Yeah.
Yeah, this is one of the big areas where I want to ask more questions about this next year, like AEO specifically, which is right now there's, we don't know what to call this, but answer engine optimization is one of the terms that gets thrown around. And the idea is how do we show up more often in chat gpt or Gemini or Claude or whatever your tool choice is. ⁓ So we've been talking to clients about this for almost a year. And in the last six months, we've started running more experiments with this with clients and starting to see what works. so,
I'll say that fundamentally, ⁓ you need content in order to get visibility in LLMs, which makes sense. All they do is read text from the internet and then index it and then figure out how to spit it back out when people ask them questions. ⁓ They aren't as smart as you might think. And so I think a lot of times what people assume is that they actually understand the information they're giving out. But in reality, they're just regurgitating information in one way or another.
One thing you want to think about is how do we feed the LLMs information that's easy for it to digest and push back out when people ask a question. But two of the big things that stood out to me as we were running some numbers lately with clients are refreshed content. So content that's kept up to date.
tends to perform better. LLMs seem to be indexing stuff more frequently. so they're ⁓ basically they're pulling recency and saying, giving a strong bias towards recency. So in other words, if you have strong content that's two or three years old, refreshing it should be your number one goal for LLM optimization or the easiest win you're gonna have with every client when we do this. Yeah.
Disha Agarwal - Reo.Dev (28:09)
But what does refreshing, sorry,
sorry, I just don't understand. I've heard this a lot that, you know, you should refresh your content, but what does refreshing mean? What percentage of content do you need to change to make it say refreshed?
Karl Hughes (28:18)
Yeah.
Yeah, that's a great question. And I don't know that I, we don't have a scientific answer for that, to be honest. The way we approach it is we go look for what, so we'll look for a piece of content that could conceivably show up in an LLM prompt that we're monitoring. And then we will figure out like kind of backwards engineer, how do we make sure this piece of content is going to answer that prompt and any of the other related ones that we're tracking around it? So in other words, usually it's
pretty minor changes. And then the other thing you do want to watch for is like correctness. So again, LLMs are not smart. And so if you're posting about content that's say a product feature that was deprecated five years ago, LLMs think it's current because you just posted about it. You know, they don't know how to parse the version numbers and things like that very well. So you do have to be like, you want it to be technically accurate and correct and up to date. And then you also want to make sure it answers the kinds of prompts that people are asking around this topic. Now, LLM like,
Monitoring is a black box, right? They're not giving us any real data yet. So we're using tools to try to backwards engineer it, and it sort of works, but it's a guesstimate. Nothing is perfect in this space yet. So take everything with a grain of salt. But anyway, so refreshed and up-to-date content seems to be working for increasing presence and traffic from LLMs The other big one is external content. And so this may play into that whole influencer, where we get the ROI from influencers. But what we're seeing is
with each new client that comes on, we also get them ⁓ publishing content out on Dev2 and Medium and don't know, Hacker news and a bunch of other places. We'll get them publishing on other partner sites. We'll get our clients cross collaborating on things, if it makes sense. ⁓ And so what this does is it sort of gives the LLMs a greater surface area to find out about your product. So in other words, they're not just reading about, let's take Reo.dev. We don't just read about Reo.dev on your homepage and your comparison pages and your product pages.
Disha Agarwal - Reo.Dev (30:08)
Thanks,
Karl Hughes (30:16)
Now somebody wrote a really good article about how they're using it over here and how this feature was really meaningful to them. So the LLM takes that as, oh, this is more credible because now we're not just getting one domain it's talking about. We're getting one, two, three, four, five, and a bunch of different domains and people talking about it. So those are the two things we're focusing on right now mostly with AEO. And I think, again, it seems to be working, but all this with a grain of salt that this could change anytime. We don't have perfect data. This is directionally probably correct, but who knows about the specifics.
Disha Agarwal - Reo.Dev (30:39)
Yeah.
Karl Hughes (30:46)
It's a fun evolving field for sure.
Disha Agarwal - Reo.Dev (30:49)
And
the evolution is happening at such a pace, it's really hard to catch up because I think, so we were also discussing AEO at Reo and we realized that Wikipedia was one of the main sources and then Reddit became one of the main sources. But then, they stopped overindexing on Reddit and then we had to change our entire strategy. I think it's quite fun to figure out what might work.
Karl Hughes (31:07)
Yeah.
It is.
like anything in marketing, you have to be ready for it to change. also, if you're fundamentally, and this is part of why I love draft.dev and what we do, fundamentally, if you are putting out useful content for people on a consistent basis, you will drive more interest in your product through SEO, through GEO, through ⁓ social media. Like all these things will benefit from it. It's really just a question of how quickly can you
Disha Agarwal - Reo.Dev (31:25)
with.
Karl Hughes (31:38)
push those results, you know? And again, like these little micro optimizations, like Reddit's more indexed, so we're gonna do more Reddit. And then you realize, they kind of backed off that, now we gotta do more of this. Like all this stuff's gonna change. like, I think it's hard to chase those little things all the time, but you can certainly find some ⁓ overall, if you've got consistent content production going, you're gonna be in a better spot than if you just let that languish.
Disha Agarwal - Reo.Dev (32:01)
Make sense, make sense. And what do you think about SEO? So do you feel SEO, AEO, do they go hand in hand or should there be a separate strategy for SEO? What are your thoughts there?
Karl Hughes (32:09)
Yeah,
man, great question. ⁓ This is definitely evolving too. So we've been using a blended approach where we're kind of taking keywords and prompts and sort of figuring out how we can merge to get the most overlap. So we get the best kind of bang for our buck for every topic we're pursuing. But I don't know if that's the right long-term strategy or not. I think that there's gonna be, ⁓ I think certain kinds of content will move towards
LLMs and certain kinds of content will stay relevant for the traditional search engine. So I'll give you an example ⁓ when I was looking into You know a new webcam not too long ago I literally typed a huge prompt in the chat GPT about my use cases and my environment and my lighting and my you know cost area whatever and I asked it for like recommendations on on webcams right and like it gave me a list but had all these like
I mean, like silly errors and things that like I wanted to double check because they just didn't seem quite right, you know? So it did give me a starting point, which is great. So it answered my top of funnel. What's the best webcam for me in a really complete way? Whereas SEO wouldn't have done that. I would have gotten listicles of just like random things, you know, that weren't relevant. So it did that first top of funnel, but I still had to go search each of these webcams individually and look at things like, oh, what are the actual specs here? What's the actual cost if I go to Amazon? Is that the right cost in my country?
Little things like that. So I was ended up Googling a bunch and doing traditional search as well. So I think that's what we're seeing more and more is this blended approach that people are using. And I see it in my sales calls. We get now close to a quarter, maybe third of our sales calls every week from LLMs. Like people say, we heard about draft.dev through an LLM. That's great. And often those people come in and they've,
Disha Agarwal - Reo.Dev (33:54)
⁓ yeah.
Karl Hughes (33:57)
then they went and started in the LLM, but then they also had to Google around to find some of our maybe past work we've done or clients we've worked with or case studies or whatever, and they come in with a bunch of information there. So I think we're starting to figure out what are the patterns of usage going to be? And maybe over time, people just move towards LLMs for everything, but I do think there's certain cases when you have certain specific knowledge that just might be easier to use traditional search. So we'll see.
Disha Agarwal - Reo.Dev (34:25)
So your recommendation right now would be to do both and try and have as much overlap as possible.
Karl Hughes (34:32)
I think that's the easiest way from like a, again, like looking at an ROI perspective. I don't think you need to build a separate piece of content for an LLM and a separate piece for SEO that talk about the same topic. those, that's a little bit, that's probably just gonna confuse things because they're ultimately both getting indexed by the same machines and getting put in the same databases. So you don't wanna have like this duplicate content that, you know, they can't parse apart because one's optimized for LLMs, one's for SEO.
⁓ Fundamentally, I think you want to have a single piece of content that answers the prompts that are relevant and the keyword that you're targeting or whatever and is actually useful content for humans and is technically correct.
But again, it's situational, because if you're really focused on the bottom of the funnel and conversions, you might not care as much about SEO. So it's just, it depends.
Disha Agarwal - Reo.Dev (35:19)
very interesting. And I think the confidence in content still remains. Understood. So Kyle, if you had to distill the entire survey into three practical takeaway for DevTool teams, know, planning GTM in 2026, what would they be?
Karl Hughes (35:25)
yeah, for sure.
I love it. ⁓ That's a tough one. ⁓ I think ⁓ the first one is understanding, like understand your place in the market and where your biggest competitors are showing up. Do you want to be in the same places as them or different and just know that? what I see a lot of people missing is they say they don't have competitors or
We are just in a class of our own or we're creating a new category. But the reality is everyone has competitors and those competitors might just be the home-built solutions or the hacked together solutions that aren't great. Either way, you know your competitors, know where they are and make sure you're showing up in the same places where it makes sense and different places where it makes sense. Then the other thing from this ⁓ survey is, I kind of talked about this already, but like marketers, think in dev marketing need to continue to explore what AI can do for them.
Don't approach it with this fear of it's going to take my job, but instead ask like, how can I get it to make me 10x more productive? Because the marketers that are going to like stay relevant and stay, keep jobs in the next 10 years are the ones who learn to use these tools really well. ⁓ The ones who are fighting it right now. I mean, this is, you're going to get left behind. And again, I'm not like AI is going to solve all the world's problems or be perfect. Like it's not.
But it is useful for certain things. And once you bake in its own checks and you build more robust processes with it, it can do a lot. And then the other is understand your experimental versus your foundational budgets. I'll give you an example that illustrates this. At an early stage startup, technical founder come to me not long ago and they just said, well, it seems like TikTok is really the place where people my age go.
We just wanna market on TikTok. look, for an early stage startup, you wanna have maybe one channel, fine. If that's what you wanna do, fine. I'm not gonna fight you. But it's the highest risk, lowest chance of guaranteed reward kind of channel to pick. It's great for experimental budget. Take 20 % of your marketing budget and say, we're gonna go experiment with influencers. 20 % is gonna go experiment with crazy videos on TikTok. Cool, let's do that, but.
it's not really good place for your foundational budget. Like the sort of core things you have to do to keep this compounding growth over time. So just understand the differences and try to build out your sort of like plan based on what is really foundational to what we do and needs to get out there, benefits all stages of the funnel and compounds versus what is out there that is like kind of cool but probably a trend that may or may not really work for us.
Disha Agarwal - Reo.Dev (38:03)
Understood. Great takeaways. Just one question. So you mentioned using AI to help you become 10x better. Just want to understand, are there any workflows that you are using which is really helping you your work today?
Karl Hughes (38:16)
It's a ton. Yeah, mean like so many I can't even ⁓ So we started working with so this is a to disclose this is a client but also I really like their product called zo computer ⁓ And it's similar to Claude bot, which I think has been getting a lot of play lately so ⁓ similar idea though where ⁓ You essentially have this little operating system or virtual machine out there that you that runs LLMs and runs agents for you and then has its own file system But you can get it to do some really complex things. So for me, it's like
hey, somebody asked for a meeting, okay, I chat via text message to my Zo computer and I say, hey, set up meeting invite for Disha and I, ⁓ you've got her email in my contacts, you've got my calendar, go book it for us, you know, right? And it just does that. ⁓ Yeah, so I think things like that, like anything that two years ago I was using a virtual assistant for, I've slowly been migrating those workflows over to ⁓ using AI. And I think what's cool about that is it gives everybody the potential to have their own.
Disha Agarwal - Reo.Dev (38:57)
All well, I'm just...
Karl Hughes (39:14)
virtual assistant once you get it all kind of tweaked. It does still take some tweaking. So I do admit I spend a couple hours in the weekend trying to get all that stuff right and make it, you know, build the skill that it needs to actually do all this stuff. yeah, there's a little bit of time investment there, but anyway, so that's one personally with the company with draft.dev, we use LLMs at a number of different places upfront during research. There's a phase where we pull data from Ahrefs as well as some of the LLM monitoring tools. And then we
put that helps us with ideation and sort of picking topics as well as how we headline things and what we focus on in content. So for certain piece of content, it comes into play. ⁓ We use it on the content production side now where, ⁓ you know, content is, there's always an engineer at the front and back end of the loop, but a lot of the middle part is written out through or created with generative AI now. ⁓ Same with processing comments from, ⁓
clients. So we often get dozens of comments when we have multiple reviewers on a piece of content. And so parsing through those and trying to understand how severe are these, what's the priority order, super hard problem. But now we have LLMs that read all these comments and kind of give us a concise list of here's everything that was said, here's what we can action item, here's the conflicting questions. It's really cool. it's helped a ton with a lot of these workflows. It enables us to go from
A year ago, it took us six weeks to get clients spun up and started to now people get their first piece of content in like two weeks. If you multiply this out to all the different marketing channels you have and try to figure out ways to optimize and automate this stuff, it's really powerful.
Disha Agarwal - Reo.Dev (40:49)
Yeah, yeah, totally agree. We are also seeing the power of AI in our workflows at Reo right from prospecting to doing research to making sure that when we are re-engaging with prospects, we have bots which are able to tell us the entire context or whatever conversation had happened. It's just so amazing to use it, yeah.
Karl Hughes (41:05)
Yeah, yeah, that's so cool. yeah. I mean, I
will say that it does make me a little nervous about all this data going out to the cloud, and you gotta watch your security a little bit, but I'm with you. I overall, there's so many, mean, 99 % of the conversations I have are mundane and have, don't care if Google knows them at this point, but yeah, I would say that the benefit it's gonna have on like,
Disha Agarwal - Reo.Dev (41:15)
Yes.
Karl Hughes (41:32)
being able to pull together tons of information that's out there and deliver it in a concise format is great. And so, you know, from my perspective, it could be that in a year or two, more of our content is read by LLMs than is read by people. I don't mind that. It doesn't mean it's bad. It's just a, that's just the way things go sometimes.
Disha Agarwal - Reo.Dev (41:51)
No, think it's such an interesting perspective, Kyle, because, you know, we were just discussing this with the CTO of Reo.Dev He was also telling us the same thing that, you know, evaluation, like DevTool evaluation is also shifting towards a place wherein it's going to be more agent-driven, wherein we're already seeing that a lot of documentation is being read by agents. So, you know, this evaluation is also going to be done by agents. So you'll need to make assets which are more agent-readable. And then, as you said, security comes into picture. How do you keep secure and how do you make sure that, you know, it's the agents who are
reading it and still the security and privacy is maintained. So I think very very interesting times ahead.
Karl Hughes (42:26)
Totally.
Yeah, totally. I'm 100 % agreeing with that right now. There's certainly some content that's gonna be read by people, but like a lot of it will probably be LLMs and agents reading stuff, consolidating it, putting it into a big report for me when I'm making my next purchasing decision as well.
Disha Agarwal - Reo.Dev (42:34)
Yes.
I think exciting, exciting times and exciting things to look forward to. All right. Thank you so much, Karl. This was really, really insightful. And thank you so much for breaking down your survey and sharing the results with us.
Karl Hughes (42:48)
Definitely.
say link put a link into the survey directly draft.dev. You can see it in the home bar for the next few months. We'll have it up there or just email me if you ever want to copy and you don't want to put your email address in. That's fine too.
Disha Agarwal - Reo.Dev (43:08)
Definitely, we'll link it in the comments and know, people can reach out. you.
