Loading summary
Transcript77 lines
- [00:01]
A
Hey, everyone. We are joined by Mike Futilla today and he is a amazing expert at Vibe coding apps for marketing and taking the pain out of your marketing. That redundant manual work that you have to do that you hate. He's going to show you how to get rid of that and we're going to walk through a whole app he's built. We're going to show you how to do better TikTok and Facebook ads. It's going to be awesome. Let's get to today's show. Before we get into today's show, here's a quick word from Cutting your sales cycle in half sounds pretty incredible, but that's exactly what Sandler training did with HubSpot. They used Breeze, HubSpot's AI tools to tailor every customer interaction without losing their personal touch. And the results were pretty incredible. Click through rates jumped 25%, qualified leads quadrupled, and people spent three times longer on their landing pages. Go to HubSpot.com to see how Breeze can help your business grow. Well, Mike, I'm very excited to have you on the show today. I first discovered you on Twitter because you'd built a really cool, like, Instagram trend research tool that I had found. And we were in the middle of HubSpot of kind of rebuilding how we do social and social, organic and ads. And I was like, oh, this is awesome. Like, this is the kind of work that marketing has to do in this new modern post AI world. And so I was like, ah, let's get them on the show and like break it down and, you know, talk about that, but also just talk about, you know, the modern state of AI and building from marketing in general. But, you know, I think where we first connected was around so many marketers are doing manual trend spotting, manual research, and social still feels kind of clunky and old school especially. And like you've been doing a bunch of work to kind of fix that. I'd love to hear how you've been doing that. Like what you've been building and everything there.
- [02:07]
B
Hey, thanks a lot for having me on Kip. I'm happy to get into this. And yeah, the specific use case I think we're going to talk about today, as you said, is scraping social media or searching social media for kind of trending topics in certain niches and having AI not only find that content for us, like going out and scraping TikTok for example, but also having AI analyze the videos. Right. As you know, Gemini now watch actual videos. So can we build a workflow or an internal tool or a system that teams can use where instead of having an employee literally go out on TikTok and kind of scroll the feed and find trending videos and take notes. Can we automate some of that process? And so this is a good use case at least. You know, the people that I work with are E commerce brands who are running their own ad or Facebook ad agencies that want to find, you know, what's popping on TikTok. Right. TikTok's usually the place where stuff goes viral first.
- [03:12]
A
Yes.
- [03:13]
B
So can they tap into to those angles, those hooks, those topics?
- [03:18]
A
Because you're going to want to use those for ads. Right. You might want to use them for content on reels or YouTube shorts or whatever. And so if you can kind of get the leading edge, you can then leverage it for your company and do a very fast follow on all that.
- [03:30]
B
Is that kind of the goal? They usually use them in ads. Right. So find, find what's popping on organic and then quickly test it as an ad.
- [03:38]
A
Right.
- [03:38]
B
Because the big thing in Facebook ads right now is creative volume, testing new hooks, testing new angles, putting out a lot of ads. And that's kind of how this was born out of a client that I was working with who was trying to automate the painful process of finding videos on TikTok.
- [03:55]
A
Show us what you're talking about here so that everybody can understand what we mean by say, like really finding these trends and acting on them really quickly.
- [04:02]
B
Yeah. All right, so this is the evolved version of this tool. It started in N8N. This is kind of version two, which is the vibe coded Claude code version. So I kind of vibe coded this, but just to demo how it works. Right. So say you're a brand like AG1, okay. And you want to find, you know, topics that are popping on TikTok around the keyword of gut health. Right. So we can type in a keyword over here, we can search a date range. Like how far back do we want to scroll or search on TikTok? We can set the number of results. We'll do 20, hopefully it doesn't take too long. And then we're just going to click on find trending video. So what this is doing effectively in the background, not to get too technical, but it's using Apify, which is a scraper. So it's calling the Apify API. It's going out to TikTok. You can see it already pulled in the results.
- [04:53]
A
Nice.
- [04:54]
B
And so apify is giving us this data now. We're kind of Displaying it in a way that's helpful to the user. Right. So went out, it found 20 videos about gut health and a high level. We can see all of the views and all of the metrics. Right? So this video got 38 million views, 4 million likes, 33,000 comments, and you can search for hundreds and hundreds of videos if you want. I kept it a smaller search here. You know, if you see something that looks interesting, maybe for your brand or something that you might want to run. And there's a couple of steps that we can go from here. So number one, if this video has a transcript, it will pull in the actual transcript from the video. And we can also analyze this with AI. So when I click on that button, what it's going to do is send this to Gemini and you know, based on the prompts that I have going in the background, it's going to pull out a couple of elements depending on how long this video is. You know, what's happening in the video, what's the hook, what's the messaging, what's the angle? Are they addressing any pain points, those kind of things?
- [05:59]
A
All right, Mike gave us his insights and we turned it into a playbook with over five AI workflows and 15 prompts that automate everything from scraping trending content on TikTok, Instagram and YouTube to generating creative briefs. No more manual scrolling for hooks and angles. AI will do the research for you. Get it right. Now click the link in the description. Now let's get back to the show.
- [06:26]
B
All right, so you can see Gemini came back pretty quickly and it's giving us things that marketers want to know about videos if they want to put out an ad about this. Right. So we've got the visual hook. Right. The video opens with a close up, unfiltered view of the creator's acne prone skin. What was demonstrated in the video? Usually a product that shows that this problem was solved. So it's pulling in kind of that angle. What's the theme of the video? What's the funnel stage? Right. This is identifying it as middle of funnel. Now all of this obviously can be customized, right? This is just a prompt on the back end to Gemini. We're asking Gemini, hey, pull in these four things. But if you're running this internally as a tool and you want to look at other elements of a video, like if your team were to manually analyze a TikTok video, what do you want it to pull out? You can just tell Gemini and the results will be different. For this video, it's just those four things. And then it also can analyze, which I think is really helpful. The comments, right? There's a lot of gold in the comments, especially when we're talking about products. What questions are the commenters asking about your competitor's product? What are they complaining about? You know, what are the things that they like about? Right. So it's also pulling in the common questions, right? How do you make the ginger shots with Chia water? And then it's also, you know, giving some key insights, right? Users strongly believe in the gut skin connection. And so again, this is just one example of how we can structure Gemini to pull insights from the video and from the comments.
- [07:56]
A
So what I think is very interesting about this, Mike, and I want people who may not be in the weeds of this as somebody like you or me, is that it's not just, hey, it's pulling some information. It's like you are an expert in what makes, in this case these types of videos successful because you're working with companies doing this all the time. And so it's a very point of view driven research process of like, this is exactly the things you should care about, the things you should know. And I think that's part of the modern AI process. It's like, how do you scale expertise? So instead of having somebody like you having to scroll through videos for hours and kind of annotate all this manually, you can basically scale what you think is the optimal process just for you in doing all of this and make it really clear and accessible what you actually should care about. And then potentially, I'm sure we'll talk about in a second what you actually do with this information to actually get real value.
- [08:50]
B
Yeah. Now that's a great segue because, you know, the version one of this build, when I was kind of just building it was all right, we can scrape all the data, right? We get all the views, all the likes, all of the comments. All right, that's kind of like layer one, Layer two is what's actually happening in the video. All right, We've got the hook and the theme and all that stuff. And then, you know, what do we do with this? And this is what clients ask for. We're doing the research, but the output that we want are new assets, right? New ads. And that generally comes from creating a creative brief, right, that you're going to send to your creative team or to your ad team. And so I built in this button here which is Generate brief. And I have kind of AG1 in here as a demo Brand just to demo brand.
- [09:32]
A
Yeah, of course, it's a big consumer products brand, lots of people know it. So it's like a great example to use.
- [09:36]
B
Yeah. But what it's doing is it's going to pull in like a brand bible that I gave AG1, you know, customer pain points. Yeah. What's the brand about? Like background about the brand and the audience. Right. So that's context and I can click generate brief and hopefully it will generate a dummy brief template. Right?
- [09:55]
A
Yeah.
- [09:55]
B
And then from here, you know, whoever's running the ad team can obviously edit this, but now we've gone from research to brief in a matter of minutes. Like literally a matter of minutes.
- [10:07]
A
Kieran and I have this thing called a taste profile that we use, but it's essentially that, right. It's all the data and information on your customers and what they are looking for and then also what's your brand style, what's your brand voice? All of those, matching them together, once you have that, which is basically what you've given, you know, this agent in the background, you can generate a pretty accurate brief and we're going to take a look at it right now.
- [10:29]
B
Yeah. So here is the brief that it generated. You know, we've got a campaign name, an objective. Right. So drive website traffic and increase new monthly subscriptions. Target audience. The educated Googler. Right. They've seen DIY wellness trends on platforms like TikTok and believe in the core principle, but are frustrated by the confusing preparation, conflicting advice and negative side effects of these homemade remedies. They're looking for a reliable all in one upgrade. Right. And then it gets into the pain points. And this is just a template brief that I fed the system on the back end. Yeah, right, right. But it gets into the key message. Right. Stop the DIY guesswork. AG1 is the simple science backed foundation for the gut health that supports clear skin, yada, yada, yada.
- [11:13]
A
Right.
- [11:14]
B
And again, whoever is using this tool can put in a template of their own. Like what creative brief template do they use? And the AI will use that as the template to then fill in the blanks, so to speak, on what you provided. Right. So this is kind of like the net output and this is what teams want, right. They want more briefs so they can create more ads and now they can have AI speed it up very quickly. But it's also based on real data, Right. It's based on videos that are actually doing well on TikTok.
- [11:48]
A
Yeah. Well, what I think is really cool about this Mike, is that it's the perfect marriage of what AI is great at. What AI is great at is taking all this unstructured data, in this case, some information from TikTok videos, and on the other side, the information about the company and the problem is trying to solve and match it together and basically say, hey, this is probably what you should do. And that's literally what most marketers are looking for. They're looking for some clarity of strategy and clarity of focus so that they can go and run that down and start executing it. And I imagine what you find is a huge speed up in acceleration in your ability to like, ship workout.
- [12:24]
B
Exactly.
- [12:24]
A
Because you get to that point of clarity really, really quickly. And I'm sure that any human looking at this brief probably makes a few tweaks, disagrees with a few things. Right. And. But then it's like, okay, it's like 80% of the way there, I made some changes and now it's time to go.
- [12:37]
B
Yeah. And I think it's the team members working with AI as kind of like a sidekick or a side colleague that can speed up the process, maybe get you 80% of the way there. Of course, the human is definitely needed to put the finishing touches on this and make sure that it makes sense. But just equipping your team with certain tools or workflows that can just make their job easier and allow them to just have more output, especially in the ad game. You know, I'm not a Facebook ad person, but the clients that I work with are all about running ads. And, you know, from what I've heard from them is creative volume with like the meta algorithms is like the number one thing that meta is looking for, right Net new concepts, as they call it. And so this is just one part of the process. To be clear. They also actually need to create the ads. And you know, AI can help with that with static ads, potentially video as well. And so, you know, they're using AI all along the process. This is just the beginning.
- [13:38]
A
But yeah, we're talking about like the ad strategy. Right. And basically here in like sub 15 minutes, you've gone from nothing to an ad strategy. And then it just depends on the size and scale of your company and your product, whether you have some automated workflows to generate those ads. You've got a creative team, what have you. But like, we've gone from, hey, we know we have a certain velocity of ad campaigns we need to run to keep our metrics high. And this is allowing us to do that at the level we need it.
- [14:05]
B
Yeah. If you were running this for clients, Right. You could have, you know, multiple projects per client. This is what I was talking about, the template that I gave it. Gemini wrote this for me. And then like the brand bible, tone of voice of the brand, what are the target audience? What are their pain points? And as you said, AI does a really good job of just layering on context.
- [14:25]
A
Right.
- [14:26]
B
Arguably you could have another section here that says, what does a winning ad look like for our company based on our historical data, running ads? So that's another piece of context. And throwing all of the context at Gemini, it can watch the video, we can take all of this and it can just synthesize it into an output very quickly. I think is the big thing.
- [14:47]
A
Yeah, I mean, I think you're right. I think we're telling people practically what to do here is like you would take the output of this brief, you would match it with past campaign data. You'd give that to Gemini or Claude or someone, let them basically match those two things together to say, oh, here's some conceptual ideas to build and test. And then you could put that into Weave or Nana Banana Veo, or depending on what type of ad format you needed to be to actually get a first rev of the ad creative itself.
- [15:17]
B
Correct. And I mean, there are people who are very ambitious who want to, you know, go even further. Right. They want to automate that part from. You know what I mean? So part one. But they're like, what can you do for me next? How do I actually get a static ad image from the brief? Right. And so if you were truly building out a real solution, you could potentially think about that. All right, this is phase one. All right, this looks good. How do you tie this into Nano Banana so I can get 20 static variations that we can quickly test, all as one workflow or one tool.
- [15:52]
A
Yeah. And I guess the question I'd ask for the audience here is as somebody's going through a workflow, in this case, we're talking about ads. It could be almost any part of marketing.
- [16:00]
B
Yeah.
- [16:00]
A
Like, how do I determine where to draw that line of automation versus, like, hey, I, the person really need to be deeply involved here because this has got to really be right. And I don't think I can automate it.
- [16:12]
B
You know, I think at least the teams that I work with, anything that is text based meaning research. Right. Deep research. It's very good at that. Writing, you know, briefs, writing, first draft ads, all of that can be automated very well at this point with certain tools, like N8N or if you're building a custom solution inside of Claude code. So like research writing written content as I would call it, right. AI is just unbelievably good at that, especially research. Like if you think about, like if you're taking on AG1 as a brand new client and you need to do a deep dive into their brand, their customer reviews pain points. Like what are the comments left under their videos? Like that manual process is like a huge job to do. AI is really good at the research, pulling in ideas, distilling those ideas into whatever format you want. Where I would caution, you know, automating certain things is when you get into the more creative side of images, especially video.
- [17:20]
A
Yes.
- [17:21]
B
I can't tell you how many people come to me and they want an automated way to create AI UGC videos for whatever reason. You know, maybe it's ads, maybe it's TikTok, maybe it's TikTok shop. Because AI video, you know, depending on what use case is, it's coming, right? And those videos are getting very good.
- [17:40]
A
They're getting much better.
- [17:41]
B
They're getting very good. And you know, if you're running ads or you're doing TikTok shop, people are interested in, all right, I want to experiment with AI video. Can't you build me something that automates 200 AI videos? Like that is not something that I would do. And it's frankly I just tell people, like, the quality is going to be terrible. You can set up a system that will automate AI video, but is that output going to be good? Is it going to be a good reflection on your brand? Right now the answer to that I would say is no.
- [18:09]
A
So I think it's very hard without a human being deeply involved in doing some very clear instructioning and prompting for that video generation and really revving on it. Because I think what you and I probably also learned that it takes a bunch of revs. That initial first or second shot isn't that great, but you have to do a few revs. And then the other thing I would tell the audience and remind the audience it's like AI is really good at like getting you to the mean, getting you to like what average looks like. In marketing, it's about deviating from the mean and how you actually stand out. And that's where humans are so, so important. And experience and creativity really, really kick in.
- [18:47]
B
Yeah, especially as you said with like, with video, even if you're manually doing it in a tool, not automating it like manually at your desk, you're going to get bad output, like guaranteed. Right. And so if you're going to try to throw an automation layer on top of bad output, that's just a recipe for disaster. Like you're going to spend a lot of money in. I was going to say it's going to be expensive too. Right. Like we can build the system, we can build the engine. Like it's possible to automate AI video, but is it going to be good? 99% of the time, at least right now, you know, talk to us in six months. But that's something, as you said, you definitely need that human.
- [19:26]
A
And even if it was good, it's only going to be good if it's like this very, very, very focused use case.
- [19:31]
B
Yeah.
- [19:32]
A
Like you have all of the context humanly possible and even then you're going to get randomly bad output. And I think that's what's hard. And I'm with you. As we get more GPU capacity and everything online and the video models get better, this will all be easier. But the reality is the making of the videos and stuff, that's kind of the fun part. A lot of the part we're talking about automating today, that's actually the painful part.
- [19:54]
B
Right? Yeah, I mean the video creation, like it's, as you said, that is the fun part. Like iterating, learning. This is a good prompt. This is a bad prompt. Anytime you're going to try to automate that, as you said, it's going to be expensive. It's, it's. And you're just going to get terrible outputs. You'd be surprised that people who want this type of solution. But to me at least it's not something that I would build.
- [20:17]
A
Right. So now that we got you for the last couple minutes of the show here, Mike, as we're thinking about the buzzword of AI agents and like building kind of workflows and agents to do custom parts of marketing. First of all, what should people be thinking about maybe building for themselves to, to do marketing really well today? And then I'd love to hear a little bit more of like also like the way in which you build these things. Because people I'm sure in the comments are going to have a ton of questions around that.
- [20:42]
B
I think so there's huge AI agent hype going on right now. I don't know how closely you follow the whole law thing.
- [20:49]
A
Oh, moat book is crazy. We could talk. I have, I have some crazy open call ideas that I want to work on setting up. So have you Been a future show for everybody.
- [21:00]
B
I mean, listen, I consider myself pretty plugged into the. Maybe not as much as some people who go absolutely crazy. I don't know. I don't want to say bad things about it. I struggle where we stand today to see the actual value, like real business value out of these, you know, open claw type. Let's call it AI agents. From what I've seen and I've watched a lot of videos and I'm on X pretty frequently. It's like you send a telegram and maybe it can like book a dinner reservation, but it's your security. All of your API keys are leaked and it takes like six hours to do it. I don't know. So, you know, I think there is a big distinction between like an AI agent, which is a very hot thing right now.
- [21:45]
A
Yeah.
- [21:46]
B
And what I, you know, an AI automation. Right. Which is more predeterministic. Right. It's like we, we have certain steps that we want the AI to do and it will execute that in order. Right. So it's not an agent, meaning it's not making its own decisions, but it's following certain steps. So this was the initial TikTok scraper, right. Before I built it in Claude code and made it like a web app sort of thing. I built it in N8N, which is a workflow AI automation builder. And this is what that tool looks like under the hood when I originally built this in nn. But this is not what you would call an AI agent.
- [22:29]
A
No, this is a workflow tool. I would say it's a technical workflow tool. You're not just rolling in out of bed and being like, hey, I'm going to make a 30 node workflow flow and N8N. You gotta be fairly technical and proficient on that.
- [22:43]
B
Yeah.
- [22:44]
A
For the audience. Mike, how much time did you spend on N8N learning before you really thought, like, you could build something of meaning you think so?
- [22:52]
B
It's a great question because when I first started with let's call it AI Automation, I was terrified of nan, because I'm not a coder terrified of it. Right. And there's other versions of this that make N8N less scary. And I was using those tools because I didn't want to deal with API keys, I didn't want to deal with web hooks. Like, I do not come from a background of this. And so to answer your question, I avoided N8N. Even though I was building these things, I was using other tools. Avoided N8N for at least three to four months. But then N8N had a huge moment where it's like, listen, if you're going to do AI automation, you need to know how to use this tool. And it took me, I don't know, four months of like, truly building these types of flows. And this is with the help, by the way, of ChatGPT or Claude in my other browser, because I still don't know how to code. You're like, I just ripped that, like, Claude wrote this. And then, you know, we did some debugging and it finally worked. But you're right, N8N is not something that you're, you know, average marketer will learn or needs to know how to learn. But I do think the next step from this is like, vive coding in a tool like claude, which I think is actually easier.
- [24:12]
A
I think it's way easier. I'm with you that.
- [24:14]
B
Than N8N. Right. As I said, N8N had a huge moment. I think there's like a turning point going on right now where, you know, is this really the future of building AI, like these scraggly little nodes on a thing? Probably not, in my opinion. It probably is something like CLAUDE code where your average user can potentially learn to vibe code a solution that, you know, your team can use internally that's actually valuable and useful.
- [24:41]
A
Yeah. So your point is, like, if you're hardcore, N8N is powerful and can be valuable, but if you're just a marketer and you're like, hey, I'm just sick of doing a bunch of manual stuff that I have to do that I think there's a better way. And it's like, hey, let me vibe code a tool and claude. And then if it is helpful, maybe I work with somebody like you or somebody else to build a more permanent person, or maybe there's a great, like, off the shelf agent product out there that I can use and there's kind of like a hierarchy of how you want to solve that problem. It seems like, yeah, you're not going.
- [25:12]
B
To go messing around in this. Even if you start with something like a CLAUDE project or now Claude cowork, it's possible to keep up with all the cloud updates. You know, something that you're doing repeatedly has a chance to be automated. And, you know, especially if you're taking a lot of time and you're literally doing the same thing over and over and over again. Where is that line? Can it be solved with a CLAUDE project with a bunch of context files? Maybe? Right now we've got Claude Skills which are kind of just, you know, pre prompts that you can call or do you need to, you know, actually vibe code some kind of tool? It really depends on your use case.
- [25:48]
A
This is also Mike, where I think we'd be remiss if we didn't call out. Reminder that Gemini can consume videos up to an hour long. Google Gemini. So if you don't know, you can literally just record yourself doing something you hate. Like, you could literally just say, this is the thing I hate most about my job. I'm going to record myself doing it for an hour and ask Jim and I what you can do to not have to do that thing you hate anymore.
- [26:14]
B
Is it an hour? I didn't even know.
- [26:16]
A
Yeah, yeah, it's an hour. It can watch videos up to an hour long. So it'll just watch and break down everything and then reason through what you can do.
- [26:23]
B
That's insane. I didn't realize it was that long.
- [26:25]
A
I should do a whole show on that. I need to do like a workflow recording and do that whole.
- [26:29]
B
That sounds like a great use case. Because as I don't think many people even know that Gemini can literally feed it a video. It'll analyze the video, it'll output all of your steps and then give you an idea for, hey, this is what you can automate. And exactly, you know, here's how you can potentially build this if you want to think about building it. But using AI to help you build with AI totally is how I've gotten here.
- [26:50]
A
Well, yeah, And I think your story is an important story. It's like, you're not a web developer, you're not somebody who writes a ton of code, somebody who's curious and has taught themselves how to build to solve problems that you have like business domain and expertise in. Right. And I think one of the reasons we do this show is because we want to empower a lot of people to go and do that. And because it is now possible in a way that it never was before.
- [27:11]
B
Yeah, for sure. And I think the business use case is actually a very important point. And you know, I kind of started generally just sharing on Twitter, just broad AI stuff, broad AI automation. And then I don't know what, maybe it was like a TikTok thing that I did, but like the E commerce people started reaching out to me then that's kind of now my use case. But you're totally right. Like understanding the business case, like what pain points do they have? Will help you build these solutions a lot easier because you understand what they're actually trying to do.
- [27:42]
A
All right, as we're closing out the show, Mike, any last advice you give for somebody who might be, you know, you 12 months ago, you 18 months ago, what would you tell people to maybe save some time and compact that learning that you've done in a shorter period of time?
- [27:57]
B
So I think whether it's Claude code or Codex, I would actually start there.
- [28:02]
A
Right.
- [28:02]
B
I would not do n8n. I would not do make.com or zapier. I would start learning how to build something very simple, like just call an API, right? Like your first API call. And I literally have two monitors. I have Claude code in one and then just regular Claude where I can just copy and paste certain things. Claude can help you build in Claude, right? AI, in terms of this, at least for me, has been the best teacher. It's how I learned nn. It's how I'm learning Claude code and codex is just by having AI help me build with AI. But if I was starting today, I hate to say it, I would probably not learn N8N. I would learn how to vibe code in something like COD code and just start very small with a small use case, something that's taking up a little bit of your time. Can you build a small solution that will actually work? And I think with a little bit of practice with Claude, you can definitely do it.
- [29:00]
A
I love that. That's awesome advice. And Mike, this was great. I think you really transformed how I think about building these lightweight tools to solve problems that are just manual hard things that get in all of our way. And thanks for joining us on the show today. We really appreciate having you.
- [29:17]
B
Thanks so much, Kip. It was great to be here.
- [29:36]
A
Hey, everyone. You know, Kieran and I have been doing the podcast for a while now. We've been at this for a couple years. We love it. We could not be happier to be doing this, but we wanted to take things to the next level. We want to level up the impact we're having with Marketing against the Grain. So the next step of our journey is something we're really, really excited about. We're going to launch the Marketing against the Grain newsletter. And Marketing against the Grain newsletter is going to be amazing. If you are a marketing leader, practitioner, you're in the trenches doing marketing every day. This is for you. We're going to deliver right to your email inbox and you're going to get all the behind the scenes frameworks, practices, tutorials from us, from guests we have on the show, and from people even beyond the podcast that we think are going to be helpful and really have an impact on your day to day, week to week doing marketing. You're going to love it. It is something we've been talking about for a while. We're really excited to have it out in the world. We've already got a hundred thousand marketers who are on this newsletter. Please join. It's completely free. We'd love to have you as part of the Marketing against the Grain community. And it's easy. You can click the link in the description below, or you can head to marketingagainstagrain.com subscribe.