Pax (02:04)
Kirsten, thank you so much for joining us today. Super excited to talk with you.

Kiersten Gaffney (02:09)
Thank you Paxton, it's great to be here. Really excited.

Pax (02:12)
So let's talk about Reddit. ⁓ There's so much going on with Reddit. It's had a massive increase in visibility, massive increase in search traffic overall, and huge increase in users over the years. When I started in digital marketing, I actually don't know that Reddit existed. in the early days of Reddit, it was a very different place than it has become in the past five years or so. ⁓

You have ⁓ been using Reddit a lot. It's our favorite place to learn about audiences because so many conversations are happening there that are super organic and it really helps you like learn an audience. I wanna know from you like what in your opinion is different from Reddit or different about Reddit from other social platforms that makes it really valuable.

Kiersten Gaffney (03:05)
Yeah, well first of all, I think there's probably 500 million monthly active users. It's a very, very big pool of active users on Reddit. ⁓ I think the thing for me is that it's got these threaded conversations that really add a lot of depth to whatever the person is trying to solve for. And even though you can do Twitter threads and you can do LinkedIn posts, ⁓

they're not nearly as interactive as Reddit. Reddit has a tendency to bring the honest opinion out from people and that is where the meat is when it comes to marketing. that's.

Yeah, that's.

Pax (03:50)
Yeah.

And, and so we, we've already given your intro, ⁓ which obviously includes ⁓ a lot of experience in tech. think it would be useful for this audience to understand your background, especially as it relates to Reddit, because I think that it makes it particularly useful. So yeah, tell us about your background in marketing.

Kiersten Gaffney (04:17)
Yeah, so I am a deep tech CMO, which means that I focus on AI data infrastructure, open source software. So really marketing to the technical audiences and it's Reddit is a great channel to find those technical people. Yeah.

Pax (04:33)
Yes.

Yeah. I mean, that's, ⁓ it's almost like the stereotype at this point of reddit and who's there like these engineers and a lot of just like very tech forward people. So that makes it a great platform for you, for you. ⁓ and I think the rest of the B2B world is starting to catch on that their audiences are there, you know, ⁓ when that's not always been the case. How have you seen, have you seen reddit change over the years as you've, you know,

worked in it and ⁓ for listening and content.

Kiersten Gaffney (05:05)
Yeah, for sure. For a long time, Stack Overflow was really popular, and that's kind of faded out. And Reddit has ⁓ kept to its grounds of being technical, although it has evolved. I mean, even my husband, who is not a technical person, loves to read Reddit. And I think that ⁓ the audience has grown just because when you start with a more technical audience and that audience really believes in it,

then you start to see the peripherals start to capture onto that because they then see that, this is actually a really good educational platform, ⁓ community, yeah. So that's what happens naturally, yeah.

Pax (05:44)
Yeah. Yeah.

Yeah, some, some cool stats and then we'll, leave this point, but, ⁓ about like the current state of Reddit, ⁓ number one, ⁓ think historically Reddit was predominantly kind of male oriented, but that has changed the, the, percentage difference between male and, and, and female users of Reddit is within one percentage of point now. So it's like very, very split. ⁓ yeah, isn't that, isn't that interesting? Yeah.

Kiersten Gaffney (06:12)
⁓ okay. Yes, that is. That's surprising to me.

Pax (06:19)
And then on top of that, there are a shocking number of people who use Reddit and use no other social platforms. So 30 % of Redditors are not on Facebook, 45 % of Redditors are not on Instagram, and a whopping 68 % of Redditors are not on LinkedIn. Which means this is so important for B2B brands. There's a big chunk of your audience that is there that is not on LinkedIn.

Kiersten Gaffney (06:32)
Hmm.

Wow.

Pax (06:48)
And LinkedIn is where all these B2B brands flock to. But like there's big opportunities here if you can just like broaden your understanding of who who is on Reddit. So anyway, lots changing there and lots to do. So let's let's get into what to do about that. I feel like you've been doing some really cool things when it comes to both like listening and then turning that into great content. Let's start at the very beginning. So how do you find your audience on Reddit? Like how do you know where to go?

subreddits that are applicable. mean, there's the obvious ones, but like, how do you get deeper than that? I'm interested in your process from the beginning.

Kiersten Gaffney (07:23)
Yeah, so I will typically start with keywords, ⁓ but you know versus going straight to a thread. I'll usually I have a listening tool that I like to use called Octalens and it will scout it'll you know syndicate all of the content from reddit YouTube. ⁓

Stack Overflow, Twitter, LinkedIn, and it'll take all that content, but Reddit is really the area that I focus the most on. so I'll go ahead and add in those keywords to Octolens and we'll then look at that every single day when it comes to those keywords. And I'll be surprised by new thread, new Reddit threads that I wouldn't normally see and ⁓ track it through that way and then ⁓ start to follow those.

folks and those

Pax (08:15)
Okay,

so you find the thread or the community from there. ⁓ What are you doing? Is it like a manual process of just reading through it or are you using any tools to kind of assimilate all that information?

Kiersten Gaffney (08:31)
Yeah, so usually what I'll do is I'll see that there's multiple conversations going on there on that one topic. I'll take that topic and I'll take the whole page, copy it, put it in Claude or chat TPT and say, hey, give it a prompt and say, hey, what is the most important takeaway from this topic? ⁓ I ask a few more questions about like, are they missing anything? Should I talk about that? Should I write about it next?

then I'll go ahead and do that. So that's, it's pretty simple, but ⁓ that's what I do. Yeah.

Pax (09:08)
Okay. ⁓

I, so do you have a preference of clod or chat? I mean, the models are always changing these days. ⁓ currently like, what is your, favorite?

Kiersten Gaffney (09:16)
Mm-hmm.

Yeah, Claude is great for copywriting and for any, really any writing task. ⁓ It's excellent for that. Chat GPT is, I find is best for research and data and it's less good at copywriting. So I tend to use both for different things and sometimes I'll compare and contrast and see what works better, but that's historically ⁓ what they're good for. Yeah.

Pax (09:46)
So, so you're taking these like a thread and then you will drop it into Claude. Well, let's say, are you having it produce when you say like, what are they missing? What are they not talking about? Are you then using that in order to find out ways that you could contribute to this conversation within the thread? Or are you saying what are they missing and not talking about so that you can then go turn that into like an article or like what's the, ⁓ that the destination of this insight.

Kiersten Gaffney (09:53)
Mm-hmm.

The destination is creating a pillar piece, a pillar blog post, yeah, for the business. Yeah, so it's not, I won't go and engage in the thread, no. ⁓ But I will go and use that as leverage to create a piece that will be captivating to that audience that I'm focusing on that keyword for.

Pax (10:39)
Okay. And how has this, how has this worked for you? you have any like, I guess what's your batting average on this? ⁓ when, when you do this process and you produce this piece, does it have like a very high likelihood of resonating with and, and, ⁓ like hitting home with this audience? I'd imagine so. ⁓ like what's the success rate of this kind of approach?

Kiersten Gaffney (11:04)
Yeah, I would say 70%. know, not only, so you're not, ⁓ it depends on what the end, so do you want SEO or do you want a paid campaign or, you know, what are you trying to get at? But usually you can nail it pretty well with that. ⁓ But there's always room for failure, you know, it's hard to know. ⁓

Pax (11:09)
Wow, okay.

Kiersten Gaffney (11:34)
really with anything if it's gonna work until you try it. But it's a much higher guarantee of getting some good eyeballs on that. ⁓ Sometimes what I'll do is I'll have a CEO or founder test it out just in a LinkedIn post to see if that resonates with their audience first. And then if that works, then we'll create a longer piece to support it or we'll create a webinar or we'll create ⁓ more of a campaign approach.

⁓ The topics are good. Yeah, they come out pretty strong and going about it in that way is ⁓ kind of a crawl walk run approach that works well.

Pax (12:15)
So are

you using the AI just for topic identification or are you using it for content production once you have the topic identification? OK.

Kiersten Gaffney (12:25)
Both. Yeah,

so we'll do the topic identification and then we will have Claude. So we'll take the research. we'll have Chat do the research. I usually have Claude do the research too, just to see compare. And then I'll take Chat's research, put that both in the Claude thread that I'm already in, and I will give it some additional prompts and ask it to develop content based on the channel that I wanted to.

Pax (12:38)
Yeah.

Kiersten Gaffney (12:55)
create the content for. And there's a lot of massaging that will go on after that. It's never gonna be a slam dunk. It requires a product marketing expert eye to go ahead and look at this. And then it also requires a copywriter's eye to look at it and to make it more focused on the brand voice and usually air it up a little bit and give it some personality that is not cheesy. So there is massaging that goes on, but it's...

Pax (13:20)
Mm-hmm. Yeah.

Kiersten Gaffney (13:25)
getting you, you know, it's getting you 75 % there. Yeah.

Pax (13:30)
Yeah.

I'm really interested in the technical aspects. So you, you work in deep tech. ⁓ and I had imagined that the likelihood for the AI content to go astray is high. mean, in my experience with a lot of AI generated content, it's like, at best it is best practice for whatever subject matter you're talking about, because it's kind of like regurgitating a lot of what has already been said. And especially in the.

Kiersten Gaffney (13:54)
Mm-hmm.

Pax (13:59)
The more technical you get, the more it often needs to be like, what new are we adding to this conversation or how are we thinking about this in different ways? So I'm curious. Obviously you said it needs to be massaged and looked over by various people. If you had to say from zero to 100 % coming straight out of Claude, where does that content sit in readiness of like, or new ideas or, ⁓

interesting ideas to the audience versus like where you then have to take it from there. Does that make sense?

Kiersten Gaffney (14:33)
Mm-hmm.

Yeah. Okay. So new ideas, it can come back sometimes pretty strong, you know, where you're like, I didn't think about that. ⁓ But with the actual meat of the piece that you're trying to create, that's where the massaging really comes in. And that's where you have to be, you already have to be an expert in creating content before you can guide Claude to.

create a good content piece. as you know, and this might be a little off topic, but as you know, ⁓ there's companies out there right now that are just creating generative content, right? And they're getting flagged by Google, and it's becoming a mess, especially for agencies like yourself, where you're like, no, that's not good content. That's going to flag Google versus Alp. And ⁓ that content is because it is just

copy and paste basically, right? And so it really does require an expert to go in and turn it into a human touch written piece. ⁓ Yeah, does that answer your question? ⁓ Yeah.

Pax (15:43)
Yeah. Yeah, it does.

Uh, I think there's the massaging can happen on the, the, the prompt side where it's like, I'm going to, I'm going to play around with this prompt to get a better output. And then it can continue to continue to happen after the output of like, okay, now I've got at least the beginning of my piece. I'm now I'm going to manually go in and add more. Uh, it sounds like you're, you see the most benefit from doing it on.

like continuing to build on the output versus like going in and playing a lot with the prompt or is that accurate to like how you're approaching it?

Kiersten Gaffney (17:47)
I still play a lot with a prompt too. it's not, I've got, and I've also, I also have prompt templates that I use to try to test my templates out and the LLM is not amazing yet. It's still giving me different outputs for the same prompts. And so sometimes I have to challenge it and say, no, that's not what we want here. This is the goal. So it's not.

I imagine a year from now or so that'll be a bit easier, but in the earlier stages now where we are at, ⁓ there's still lot of prompting finesse, so to speak. yeah, sometimes what I'll do is I'll say, ⁓ come up with a topic and then I'll say, okay, create a brief.

Pax (18:29)
Yeah, yeah, yeah. ⁓

Kiersten Gaffney (18:44)
for me, just a marketing brief that will ⁓ give me kind of the TLDR of what I would want to talk about no matter which type of campaign I run from that. And ⁓ I'll challenge, I'll work on that brief until I'm really aligned with it. And then I can say, okay, create a LinkedIn with this personality, create an email with this personality, et cetera. And those are more successful when I do that.

Pax (19:10)
Yeah, I love that. think the, the thing that everyone gets wrong or a lot of people get wrong about AI is they're saying like, this is my way of not having to think. But really when you don't think through it, it just produces mediocre results. And I found that it really forces you to say, what do I want? What is going to be successful? And you still have to wrestle with those difficult questions and you cannot hope to have a really great output.

Kiersten Gaffney (19:22)
You

Mm-hmm.

Pax (19:40)
unless you wrestle with those difficult questions. And the answers to those questions are based in knowing your audience, which I guess brings us back to Reddit and listening about like how they talk and what they want and all of that. Is that accurate in your experience?

Kiersten Gaffney (19:55)
Yeah, for sure. In fact, one of my founders, I was just training him how to use Claude and he wanted to do a LinkedIn post. And so he had this, I said, take a customer call and feed that customer call into the Claude. And here's a few things that you can say and create a LinkedIn post from that. Give it a try.

It was first, it was very intimidating for him, but he did it. And then he came back and he says, Kirsten, can't, this is not right. This is just doesn't, this isn't good. This is just generally, you this is just AI doing stupid stuff is what he said. And I said, well, what's the purpose of your piece? What's the primary key takeaway that you want? That's what you need to feed in to Claude and ask it. Like, what do you think the key takeaway is from this piece? And

it came back with a nice headline and he wasn't directing it. He was letting Claude direct him versus him direct Claude. And ⁓ when you do that, it's not gonna work. that's, ⁓ yeah. Just to, yeah.

Pax (21:03)
Yeah, I love that's a great way to think about who's who's in charge here. Uh, yeah, I love that.

Uh, I, one, one thing that comes up in discussions that I found as people talk about AI is no one's ever clear. Uh, on scale, meaning they talk about, Hey, here's the right way to do it. And in their minds, they're thinking, and you do this thing once a week and other people are assuming, Oh, I'm going to do this 500 times a day. So.

Kiersten Gaffney (21:20)
Mmm.

Pax (21:33)
In terms of scale, I'd love to know if you had to recommend to a brand who is wanting to implement this system, what is the expectation? What's the assumption here when we're saying this process of listening to Reddit, assembling that into AI, playing with the prompt, and then taking the output and getting in front of these different people? How much should these brands be expected to do this to be successful?

Kiersten Gaffney (21:57)
Yeah, well, ⁓ you know, companies like ⁓ Post Hog and some of the more well-named companies, ⁓ Super Base, those guys are pushing out a lot of content every month and they're listening to their communities as well. And I always recommend to just start with

⁓ one pillar piece and then creating like three to five ⁓ channel pieces from that pillar piece every week. That's my recommendation. But I'm talking to founders for the most part and early stage marketers and they don't have a huge team, right? So I would just say that in a perspective, like as your team grows, then that grows with your team, right? So if you're then hiring on a content marketer or... ⁓

you know, a dev rel person, then those folks can take on additional, let them be the three to five and kind of scale like that. ⁓ That's how I would approach it. Yeah.

Pax (22:59)
Okay, yeah, that's good

good context ⁓ It always helps to have numbers and I think it can make it more doable if people understand like What is expected what kind of scale is expected from this kind of process? So I'm not expected to turn around 500 blog posts Using like one reddit thread, right? but I think some brands are hoping to do that and they're trying and ⁓ I think it's hurting some of Yes. Yep

Kiersten Gaffney (23:16)
No. No.

But that's the ones that are getting flagged, right? Yeah, so ⁓

my rule is like, you have to get at least two pieces out every week and that could be the same one pillar piece and then a LinkedIn piece. But the goal is, you can get three to five pieces out of that one pillar piece and be good. ⁓ But if you're communicating in a few channels each week.

and however you do that is going to be okay. So it's a blog, LinkedIn, email, or a sales outreach.

Pax (23:59)
Yeah. So

one thing I want to get into, especially given the technical nature of your work, ⁓ something that people struggle with a lot in marketing is we need to be assimilated into this community, but it can be very difficult if you are not a member of that community. and I'm talking more like, you know, tech or, whatever the, the, the audiences that you're trying to talk to.

I think if you have the expectation of, and so I'm just going to take a bunch of Reddit posts. I'm going to plop them into, to AI and that's going to produce content and it's going to be good. You're going to be disappointed. Like you have developed a certain taste over years of working with developers and engineers. And how would you like, what would you recommend to somebody who is trying to reach a new market or starting a new business, selling into a market that they may not have.

been in for a long time. think this, this process is, is probably the beginnings or at least a framework for how someone can get into this community. What would you say to them beyond just, you know, browsing a couple of subreddits?

Kiersten Gaffney (25:10)
Yeah. what you're asking is trying to find out how to break into the community from a marketing perspective or how to understand what they're talking about or... ⁓

Pax (25:24)
Yeah, think, ⁓ understanding what they're talking about is going to be huge. That yes, more so that like, you know, the inside baseball, you know what they care about and don't care about. I mean, to some degree it's audience research. ⁓ but this is a little bit more focused on the voice, the language, you know, you know how they talk and, ⁓ yeah, that kind of stuff.

Kiersten Gaffney (25:28)
or how to use the voice that will speak to them, which that, okay.

Okay, well first of all, in a technical situation, where you're marketing to technical folks, you want to speak in plain language. And when I say that, I'm thinking of...

Claude always coming back to me in plain English. I always scratch that out. But in language that is easy for somebody to understand quickly. It's the goal of any type of technical marketing is for somebody to understand it within 30 seconds or less. And I always say, explain it like you're explaining it to your 10 year old. If they don't understand it, then your hero graphic on your homepage is not gonna be consumed properly.

Then the next level of that is ⁓ more on the deeper technical side, like if you're creating documentation or if you're creating a release blog or anything like that, where you do need a technical expert by your side to make sure that the code is correct, make sure that the deeper level of the language is correct and.

⁓ As long as you're still just like you would guide Claude you're guiding that se or that solution engineer You know whomever the or the art the developer who who built the feature make sure that you're guiding them in the same way as okay I don't understand that let me make let me make sure I understand and so being very ⁓ Curious and making sure that you are the guardrail for your audience

to ensure that they understand how you're speaking and what you're trying to say and not being scared to ask questions and say that doesn't quite make sense because most of the time engineers, they're not trained to write, right? Engineers and developers, they're not professional writers and us as marketers are. So they're looking to us for guidance on how to speak to this community and you're looking to them for expertise.

It is a partnership that you have to have with your technical team. ⁓ And the best words of advice I can give you is just make it clear and readable and don't use smarketing. Smarketing is, you know, sales and marketing kind of cheesy, lots of memes. Memes can be fun, yes, but not...

Pax (28:12)
What do mean by that? marketing.

Kiersten Gaffney (28:25)
memes that are going to be a turnoff to that ⁓ technical audience.

Pax (28:30)
Yeah.

Let's, let's take a second and talk about that technical audience. mean, this is, this isn't going to apply to everybody listening, but I find it really interesting. So, ⁓ our team has worked a lot with, with, you know, very technical audiences, engineers, CISOs, ⁓ CTOs. And, ⁓ one thing that we have learned is that audience really does not like being sold to, and they don't like,

Kiersten Gaffney (28:56)
Mm-hmm.

Pax (28:58)
things, cluttering their feeds, ads, like they're very, ⁓ averse to ads, ⁓ more than most groups of people. and I'm interested to know your opinion on advertising to a technical audience. Obviously like content is a big piece of your funnel. ⁓ and I can see, I mean, that obviously is going to resonate with that audience. It's like, I'm here to help you. This is genuine. I'm not here to trick you like

Kiersten Gaffney (29:04)
Mm-hmm.

Pax (29:28)
they're gonna detect if you're just trying to show them a meme that you don't really think is funny just to get their attention and trick them into some sort of funnel. That will repel them so quickly. So what does the role of advertising play with that technical audience in your experience given that they're so averse to it often?

Kiersten Gaffney (29:47)
Yeah.

Yeah. Really making sure that you're educating them. Your job is to help them do their job. Your job is to help them get their job done faster, no matter what. And if you aren't doing that, then they're not interested. Everybody is trying to get their job done faster and more efficiently and still remaining that quality control.

If you can't offer something to them that is not going to do that for them, then don't offer it. It's worthless. So the whole reason why you build a company, a software company, right, is to, because you're trying to solve a job to be done. And if you can't solve that job to be done, then what are you doing? You shouldn't be in existence. And if you can't communicate that job to be done that you're trying to solve, then again, you've got to work on your messaging. So. ⁓

Pax (30:40)
Yeah.

Kiersten Gaffney (30:42)
That's the bottom line, whether it's a certification program campaign that you're launching, or if it's a trial that will give them ⁓ immediate impact into their daily life, or if it's a webinar teaching them how to do something. ⁓ So it's not about you as a product organization, it's about what you're delivering to the customer.

And we all try to do that with any type of SaaS software or any type of marketing, but it's really important for that audience in particular.

Pax (31:17)
Yeah. Yeah. Just

adding genuine value. Um, yeah, I love it. And I think that's true for, that's true for everyone. Really? I mean, uh, yeah, as you said, it is especially true for this audience. And I think often marketers kind of get too cute with it or they, they, um, lose sight of what is the problem we're solving and then how do I help them solve that problem really at the end of the day? Right. Um,

Kiersten Gaffney (31:21)
Yeah, that's it.

Yeah.

Pax (31:46)
I love that. love, think that's, that's really great advice. ⁓ Kirsten, this has been, ⁓ an, really great conversation. I know you have something, ⁓ that you're releasing that I think is extremely relevant to this. Do you want to tell the audience about, what you're working on?

Kiersten Gaffney (32:05)
Yeah, so I have a course coming up on Maven in September that will be focused on how to mine Reddit and turn it into your first pillar piece and then how to take that pillar piece and multiply it into the channels. ⁓ And happy to share more about that. You can reach out to me on LinkedIn and happy to share more.

Pax (32:25)
Okay, great. ⁓ so yeah, please search for Kirsten on LinkedIn and I will, yeah, we'll post the link to that in the show notes of this episode. So, Kirsten, thank you so much for sharing your process with us. Super enlightening conversation. ⁓ we really appreciate it.

Kiersten Gaffney (32:41)
Thank you, Pax. I appreciate being here.

Pax (32:45)
All right.