[00:00:00] Speaker A: Most brands don't actually have a messaging problem. What they have is a misunderstanding of.
[00:00:05] Speaker B: What messaging actually is.
[00:00:06] Speaker A: In this episode, conversion strategist Chris Silvestri breaks down why positioning and messaging are two totally different disciplines. How teams can get audience research completely wrong, and how AI can support but never replace the strategic role of a human copywriter. If you've ever wondered why your message isn't landing, this conversation might be the missing piece.
Hello everyone, I am Paxton Gray, CEO of 97th Floor. We are a digital marketing agency built to deliver world class organic and paid channel strategies for mid level and enterprise organizations. Thank you so much for joining us today for another episode of the Campaign. The campaign is a marketing podcast about better understanding your customer, innovating beyond best practice and converting visitors into customers. You can find past episodes on YouTube, iTunes, Spotify
[email protected] Today's guest is Chris Silvestri, founder of Conversion Alchemy and one of the sharpest minds out there when it comes to conversion focused messaging.
Chris is a SaaS message market fit specialist and conversion copywriter. Previously he worked for 10 years as a software engineer in an industrial automation before taking a sharp turn to enter the digital marketing world as UX lead at the usability testing startup Conversion Crimes, Chris has been working as a messaging strategist and copywriter for B2B SaaS brands like Moz since 2016. In this episode we get into stuff most marketers skip. One, why confusing messaging and positioning can tank your strategy from the start. Two, the right and wrong way to approach competitive research. Number three, what good audience research actually looks like and how to turn open text survey answers into high converting copy. Number four, how AI can help systematize messaging research without replacing the critical human intuition that makes great copy work.
Chris also shares his Path framework for AI assisted audience research and Persona development. Whether you're a founder trying to find message market fit or a copywriter trying to stay ahead of AI, there's a ton in here to steal.
[00:02:26] Speaker B: Chris, thank you so much for joining us today. Pleasure to have you on.
[00:02:29] Speaker C: Hey Paxton. So thankful for being here.
[00:02:32] Speaker B: Yeah, I, I think it would be a great way to start our conversation today.
Asking you in in your consulting work and the work that you've done, what do you see are some of the most common or biggest mistakes companies are making when it comes to their messaging and their positioning?
[00:02:50] Speaker C: Yeah, so I think it's an interesting question because I think the probably the first mistake that I see is that they don't really understand the difference between Positioning and messaging. So a lot of the work is actually while we work on those is to like make it clear, like clearly separate the two steps in the workflow so that they have a good understanding of, okay, we're doing this positioning now because we need to know what we need to know at the step and then the next step is the messaging. So the way that I define positioning is basically understanding what you do, who you do it for, and how you do it better uniquely or in a different way. Right. So that's, I typically use like a simple one pager positioning canvas, but most of the work there happens in a live workshop.
Next is the messaging work, which is actually starting to define how you say all of that. Right. And that's through different lenses, which could be your pitch, your strategic narrative, different example of value proposition from a gain, loss, logic or benefit perspective, for example, and your brand voice as well. So first I would probably say, yeah, a lot of companies don't really have a clear idea of what positioning and messaging are and how you kind of move from one to the other.
The other one, I would probably say a lot of companies when they, when it comes to actually writing the copy. So once you've done positioning and messaging and you need to write the actual words on a page, the first thing that they jump to is to. They want to say what they do. Right. And which is a bit counterintuitive, but that's probably not the first step you, you want to on the page. The first thing that you want to do is actually to match what people are. They are thinking when they land on your page or when they see your email or when they look at your sales deck. So you want to kind of match their motivation, their intent, their awareness. And so when you first start by telling them this is what we do, they might stumble on some friction right away, which is probably something that you don't want to do.
And the other one, it's probably.
I see a lot of people mistakenly consider their messaging more of like the actual creative assets that you deliver or that you create.
Rather than thinking of messaging as I think of it as like the architecture and like a system that you install in your company that you can refer to at any point, at any touch point with your customers and, and then you can constantly iterate on, optimize and yeah, being the actual architecture infrastructure inside your company that drives all your marketing initiatives.
[00:05:35] Speaker B: Yeah, yeah. When I was studying advertising, we would talk about an idea having legs, meaning like it's not the one off execution of that Idea. But that idea can be expressed in many different formats and media. And that, to me, seems like that's what true messaging is, not the actual asset. Like, the asset is the execution of the messaging, basically. Right, Exactly.
[00:06:04] Speaker C: Yeah.
[00:06:04] Speaker B: You said when a company says straight out, this is what we do, that it introduces the opportunity for the customer to encounter some friction. What do you mean by that? Like what? What kind of friction?
[00:06:17] Speaker C: Yeah, so what I mean is that as soon as someone lands on your page and the first thing that you do is start talking about you yourself rather than themselves, they immediately experience some kind of friction, as I said, which in practical terms means they might bounce on the page or they might be instantly skeptical about your message. And so it already starts getting a bit harder to convert them when on the in. On the other hand, if the first thing they do is actually writing and speaking to their motivations, pain points, desired outcomes, the first decision that they need to take, which is the decision to scroll down or to keep reading, it's already done. Right. For them. It's okay. You're basically already telling me what I want to hear. Let's keep scrolling and let's keep going down the page.
[00:07:11] Speaker B: Yeah, that's a group. Like, what's. If you're saying stuff about me and what I care about, what's to disagree with? Really?
[00:07:17] Speaker C: Yeah.
[00:07:17] Speaker B: There's no, not scroll. I love that.
So let's get into audience and message.
I do want to touch on audience research. That is a huge point.
I think a point of friction for a lot of companies.
Before we get to that point, though, I do want to ask, how much does message change depending on audience?
And is there a limit, like a rule of thumb that you would say a brand should spread their message amongst different audiences? You know, I've seen companies where it's like, hey, we're for everybody. And so I have these 20 different.
I'm for accountants, I'm for HR and for mortgage loan officers. And it's like, man, how can you possibly have messaging effective for all these people? Is there a rule of thumb or a cap or how do you think about that?
[00:08:12] Speaker C: Yeah, so I think that's a totally fair question and something that I see a lot of people get confused on. And there's totally the need to have all of those different assets, different pages. Those are very useful when you need to match intent, especially now with AI, like AI search, and that needs to see real intent. Uh, but I think that the cap, as you called it, it's part of the messaging strategy, is defining Your strategic narrative. And what is your point of view? Right, the insight, the reason why you built the product so something that you saw in the market that was being done wrong or some mistakes that only you have seen customers make, and that led you to building your solution. I think that is the actual overarching narrative or lens that you need to have. And then the specific pain points, those can change based on the different Personas and should change, especially in B2B when you have a lot of different stakeholders and decision making unit, all of those dynamics.
But it's important to maintain that strategic narrative and that point of view. So then with that top of mind, then you can change the minutiae, the pain points, the motivations, and reframe the benefits in terms of those specific Personas.
[00:09:38] Speaker B: So let's get into those Personas.
There's a trait that a lot of marketers have that I've noticed, which is like, we love to sit behind the screen where it's safe and kind of execute all of our stuff. And getting out in front of that and talking to customers and talking to, learning more about the audience is kind of where a lot of people tend to get stuck, you know, and, and when we don't know a lot about an audience, one of the hallmarks of a brand that doesn't know a lot about their audiences, they go to competitors and they say, well, what are competitors saying? Let's just say that, or let's just say that, but better versus just saying like, go, go to the audience, allowing you to, you know, say something new potentially.
Um, so I, I think the reason people get stuck on it is the concept of audience research is, can be viewed as like super fluffy, you know, and it's hard to say there's an exact formula to it because when you've, you know, it's like saying, where is the gold in the hill? You know, it's like, I don't know, but you gotta dig until you find it. So it's not so I think people get stuck on that. What, what, what's your take? Like?
[00:10:56] Speaker A: First off, is there a place for.
[00:10:57] Speaker B: Competitive research when you're looking at building out your messaging? And then number two, how should brands approach the market research portion of building out their messaging strategy?
[00:11:10] Speaker C: Yeah, so tackling the first point, competitive research, I think that's totally the place for competitive research, but not in the sense, okay, these are our competitors, let's copy them. So what I mean by actual competitive research and how I see and think about competitive, the research is looking at the other players in the market, in the industry, and trying to deconstruct what they're doing from a lens of we know what our SCPs are, we know what the pain points, benefits that we need to address are, and also we need to know how other companies are solving for all of those.
This way you kind of establish the benchmarks. Right. So it's important to understand what other players are doing so that you can differentiate yourself.
But also there's a bit more of a meta deconstruction that I try to do, which is looking at competitors, messaging and trying to deconstruct what kind of Personas they are speaking to and in what way. Right. So that we can differentiate both maybe on the Personas or the roles that we can address if our product is better for them, but also with the language that we use. So a lot of if you look at any competitor's website or messaging, a very simple framework that I like to use to kind of deconstruct what they're doing.
I got it from the Mac Labs Institute. So it's a very good resource. If you're into conversion optimization, you can check that out. And they have a very good heuristic, but in simple terms, it's divided in four areas. So at the top of the page you typically have the motivation section, which basically it's anything that you write to match the user's desired outcomes, pain points, or purchase prompts. So any copy that you see at the top, that's your motivation section. And you can start, I don't know, categorizing all the copy that you see, dividing it into themes, organizing it, and create your different competitive research assets. Second section is typically the value section. So what are they doing? What are the features that they promote? What are the benefits associated with those features?
Then we have the proof section, which is how do they back up all of those claims or how do they address objections after that? The anxiety sections, which is all about social proof and kind of making sure that people are not encountering any of that friction that we were talking about.
So with these simple kind of sections, you can already start getting an idea of how competitors are writing their copy, using their messaging to address visitors or even in an email. It's pretty much the same formula.
And through that you start already understanding, okay, these guys are addressing, I don't know how familiar listeners are with the level of awareness or from unaware. There's problem aware, solution awareness, product aware, and most aware. Right?
By looking at the copy in this way, categorizing it you already start putting what they're doing into those kind of buckets. You can say, okay, these competitors is addressing maybe problem aware customers or product aware. So maybe we have room to adjust the solution aware. Right. So trying to find the gaps that you can cover or maybe just to give to get insights on what other players in the market are doing. So this is probably one way that I would try to deconstruct competitors. The other one is probably looking at reviews. So if you look, if, especially if you don't have a lot of reviews for your own platform, you can go on websites like G2 or Capterra and look at your competitors reviews. Those give you a goldmine of insights on the language that their customers use, the features that maybe they complain about, and all sorts of good insights and voice of customer that you can actually reuse and mirror in your copy. Right. So this is probably the way that I would look at your competitors, but not copy their messaging.
[00:15:19] Speaker B: Yeah, yeah, I like that.
Something that we've talked about too is it's important to know what the competitors are saying. Main like, also because it's the messages that your audience is hearing when they're doing research.
So like the sites that they're looking at, they're not just looking at your site, they're looking at the competitor sites too, as they're evaluating options.
And it's helpful to know what messages are they being fed, what context are they being given, what does the landscape look like?
Not so that we can copy the landscape, but so that we know how to better stand out within that landscape. Right?
[00:16:00] Speaker C: Yeah, yeah, totally. And you can see it in, in sales meetings, like you can ask people what other solutions they were considering and you can see like, are they bringing up some of those, some of the copy that competitors use, like verbatim and see, okay, is this sticking with them or not? Yeah.
[00:16:18] Speaker B: You've talked on some other podcasts about using the language that your audience uses in your messaging. How do you, you've mentioned reviews. What are some other ways that you kind of harvest that the language that they're using?
[00:16:34] Speaker C: Yeah, so 70% of my work is research, I would say. So if you think about conversion copy, like think about, think less about writing and more about research, I would say. But in that research, there's a lot of different ways that you could gather language. The most important, like the 8020 way, if you need to kind of optimize, it's obviously speaking with customers, so conducting interviews. And that's what we said before. Right. A lot of people are kind of scared of going out. I was scared myself. Like the first two years I wasn't conducting interviews, I was just doing service because I was basically terrified. And so you kind of have to break through that fear. But like interviews give you probably the best and most vivid voice of customer language that you can actually use through your copy other ways. Website surveys.
So basically those pop ups they appear on website, you can ask simple.
It's a mix of I would say of qualitative and quantitative data especially if you have a lot of traffic. So you can start understanding who your ICPs are with qualifying questions like what brought you to the site today or what best describes you and give them a couple of options.
Or you can ask questions on their frictions objections that they have maybe on pricing page you ask them what's preventing you from signing up today and you can already start getting some some of that voice of customer.
Another way that I typically like to do, especially if we struggle to get interviews is to send email surveys. So just an email invite with a link sending them to a survey and the survey contrary to what a lot of people think, it's a lot of open ended questions and you would be surprised. But if you have product market fit, if your customers are happy, a lot of them are going to like fill those out gladly. Right. And those give you a lot of great, great insights as well. Because when you leave like the open field and you have like a large text field, people especially if you're, I don't know, maybe it's the fact that they're not on camera, not recording, they might feel like more more like able to express themselves in writing. Some people especially so you get a lot of good insights that you can actually categorize in some spreadsheets. So it's also convenient on the front.
[00:19:03] Speaker B: Yeah.
[00:19:03] Speaker C: And the other probably another good way it's to maybe run some message testing. So platforms like Winter are pretty good.
Or the other other way which is a bit of of a controversy lately, it's using synthetic research.
So in that sense use platforms that simulate your customer Personas and and could give you some voice of customers. Right. In quotes, voice of customers. But that all depends on how good your real human research is in the first place. I would never just do synthetic research. I would always use it as a complement to real human customer research.
[00:19:47] Speaker B: Yeah.
So I have a couple questions on this and I do want to get to the synthetic research and some AI stuff before that when you talk about using their language. So let's say you send out this survey and you get these open responses back.
What level of using their language? Like what are we talking about here? Are you saying I'm taking their exact words and I'm just flipping them back into the copy? Are you mimicking their tone, their level of formality? Like what, what is it in there that you're like, how would somebody approach that?
[00:20:24] Speaker C: Yeah, it varies a lot, I would say depending on the type of voice of customer that we have.
Some of those snippets are very good to just literally take and use as headlines, but it all depends on how vivid they are, how specifically they represent the pain points and motivations that you uncovered in research.
And especially if you see a lot of people kind of repeating the same words, that's a clear sign that you could actually use them like verbatim on the page. Just because it's again, it's part of the conversation that a lot of people are having.
But in general, like even if it's just one single word, maybe I'm writing copy and I'm kind of debating should I, should I use this word or that other word. Then I jump into my voice of customer bank and I see, and I literally like do a common F search and I see what kind of words did they use for this specific use case or to define this specific thing that I'm trying to describe. Right. And I literally go in and use the most used word that I see in my voice or customer. So I think, yeah, it depends on first your goal for the type of copy, but also on how vivid and how frequently mentioned that specific words or sentence is in your voice of customer.
[00:21:44] Speaker B: Do you have a story you'd be willing to share where that was particularly effective or you found some kind of unique verbiage that you ended up using really worked.
[00:21:54] Speaker C: I don't, I don't remember the specific verbiage, but there was. So this client worked with a couple of years ago was a very unique, very unique type of client because it was a B2B SaaS but in a field which basically their customers hated software. And so this company was trying to sell them software, right? And the company was selling software for portable toilet and septic container management.
So super ultra specific and very like blue collar. Almost like they, the, the people they were trying to sell the software to or like very like experienced 50 year old, 60 year old owners who never really actually touched any software. Maybe they use spreadsheets at best.
And so we use that, we did some customer interviews and we used a lot of this, the specific words that they using, some of them were like very in the U.S. and they were in the U.S. california, I think. And so like very characteristic voice of customer that you might only see these guys use. And that's, that was super important because on the, on the how it works page, rather than just doing, I don't know, these are the features that we offer, maybe divided into groups by specific use cases. What we did was quite unique because we, we kind of turned the how it works page into like a diary of a typical day of one of these guys at work because we, we knew from research that they wanted to see their kind of daily activity reflected in their copy. Right. So we, we literally said when you get to work, this is the first thing that you do with the software. You jump in, you log in your route for the container and blah, blah, blah. Then your driver takes on the software and then this is what happens. Right.
Completely matching and setting expectations at every step, using some of their voice of customer as well.
[00:23:52] Speaker B: And it was successful?
[00:23:53] Speaker C: Yeah, I think increased conversion 20% from the previous version across the site.
[00:24:00] Speaker B: So yeah, that's awesome.
[00:24:02] Speaker D: Recent data shows that brands are seeing a 30 to 50% decrease in traffic because of Google's AI overviews. That's an enormous traffic loss. If this is something you're thinking about, either because you need to recover traffic or stay ahead of all the changes in AI search, I want to offer you a completely free AI overview audit. I'm Brittany Dillard, a senior SEO strategist at 97th Floor, and our teams use this exact audit for our clients to discover where brands sit in AI Overview results compared to competitors.
This AIO audit informs our strategy for how to drive more conversions as AI continues to develop in search. And we've been able to help brands maintain conversions and therefore revenue even while search traffic drops. If you're interested in getting this comprehensive AI overview audit done for your brand completely free, head on over to 97th floor.com AI audit and fill out the form on that page so our team can get started. A link to that page will also be in the show notes. And trust me, you want to take advantage of this one.
[00:24:59] Speaker E: Hello, this is Emma Lammy. I produce the campaign and today I'm bringing you the latest in AI news. What if your browser could do more than just show you web pages? What if it could actually help you get things done?
That's the idea behind a new wave of agentic browsers which go beyond traditional navigation to act more like AI powered partners. These tools are Built to understand context, take action, and automate workflows, all while adapting to how you use the Internet. They're the center of what's quickly becoming the next big battleground in tech the browser wars. After retiring its experimental browser Ark, the browser company is back with Diagnosis, a fresh start built entirely around AI, still based on Chromium. DM might look familiar at first, but it includes a built in assistant that can read all your tabs, pull in content from files, and help you plan, summarize or code, all in a single interface. Users can personalize how the AI interacts with them and even create modular tools called skills to automate specific tasks.
Perplexity, best known for its AI search engine, recently launched Comet, a browser designed to anticipate your needs as you browse. It acts less like a search tool and more like a research partner, learning from the sites you visit, connecting information across different pages, and helping you take action without constantly switching tools.
Meanwhile, Opera's Neon is also gaining attention. Unlike the AI enhancements in Opera's main browser, Neon is designed from the ground up to respond to user intent. Its AI engine can perform local tasks for privacy and generate content like reports or websites, and manage workflows without needing third party plugins or constant Internet access.
These new browsers reflect a broader shift happening in how we interact with the web, from searching and clicking to instructing and collaborating. Whether it's chatting with open tabs in dia, cross site automation and comment or task execution in Neon, the browser is becoming less of a passive tool and more of an active digital partner. You can join an invite only waitlist for any of these platforms. Now, let's get back to the episode.
[00:26:53] Speaker B: So let's, let's talk AI first. Let's touch on this like synthetic research, or so I've always felt. And I haven't done this myself, so I'll say that.
But from the research that I've done on it, it's always felt like it was some. It was like a layer of unnecessary redundancy.
The idea being like, I do all this research to understand my audience, I gather all this information and then I now feed this information into AI. And then I say, pretend to be my audience and I'm going to give you messaging and you tell me whether or not you respond to it. By the point where I have enough information to feed the AI accurate representation of my audience, I should now know my audience and so I should already kind of know the answer to the question, you know. And so it's always struck me as Just kind of being this weird, unnecessary, like a tool looking for a use rather than actually useful. How's your experience been with that?
[00:28:00] Speaker C: Yeah, that's a totally fair point. And I see like, especially from like the point of view of a company who's got lots of customers, lots of data as well, very good system to collect the data and use it.
It's probably like overkill in the sense or maybe not even useful. Right. But maybe the differentiators or the better use cases for synthetic research I think are in case when you may be just starting out and you maybe don't have access to customers or very little data. In that case, using synthetic research, if you don't get to 100% accuracy for a company, for an early stage startup that does starting out, Even getting to 60%, 65, 70% accuracy, it's already gold that you can actually use to get started, get your website up, get some email series done rather than procrastinating or wasting a lot of time, maybe in communities. So that's probably one good use case. And the other use case, it's when you have some research, but not a lot still. So maybe you have. And that's probably the case for a lot of my clients.
Some of my clients, maybe we are able to only conduct five customer interviews compared to maybe the 10 or 15 that we normally run. Right? Yeah. And there might be a lot of different in times in terms of the percentage of the data, the statistical data you cover between five and even 10 interviews. Right. So with synthetic research, that's probably a very good use case. You can take those five interview transcripts, feed them to the AI and kind of expand your data and get more specific as well. If you don't have a lot of very specific detail on your Personas or even expand into asking them questions that you haven't asked them in the interviews.
Yeah. And also probably the final use case is to scale, but also systematize your research.
Because if you imagine one goal that I'm trying to kind of build for my client as well is to kind of install a continuous research system as well. How do you install research that runs in the background continuously looking at market trends or how competition changes, or maybe you launch new products now you need to research again. How can you install some kind of research that runs in the background and always gives you fresh data that you can actually readily use for writing your messaging and develop your marketing? Right. So synthetic research is a very good use case for that, especially when you start plugging it with APIs that link to your systems.
Or maybe even if you just want to test creative in your ads before even launching your ads, you could test it with synthetic research, a couple of variants, and only launch the variant that wins in your ads. Right. So those could be useful ways to use synthetic research because of the speed, the cost, and the ability to kind of install it inside your systems.
[00:31:11] Speaker B: Yeah. Okay, so what platforms are you like, are you using for this?
[00:31:19] Speaker C: Yeah, so the ones that I tried. So I did it first myself. So I literally, at the beginning, I was literally taking all the research that I did for clients and built my own kind of repositories. When I started doing it, we still didn't have projects in ChatGPT or Cloud. Now you can build a project with all the research that you have and kind of start prompting AI to kind of wear the shoes of this Persona. But now there's platform like Synthetic Users is a very good one, especially if you want to conduct usability product development research.
Another one, it's called ask rally.com rally. So that's a very good one if you just want to basically build your Personas at scale. Like you could build 25, 50, 100 Personas, all, all different Personas wearing the shoes of your roles. Right.
And that's very good if you want to test marketing messaging because you can basically chat with your Personas, ask them questions, and it gives you like agglomerate response and then you can dig into the individual response.
And a lot of the things that these platforms, I mean, the platform that do synthetic research. Right. Are doing that you should actually look for. When you look for synthetic research platforms, it's to look how they are addressing bias. For example, are they trying to make sure that the AI is not super positive and wants to praise you all the time? Are they trying to kind of coordinate all of those little aspects and nuances and also how are they building those Personas based on your role? So if you tell them, okay, we have a CMO Persona, head of growth Persona and a founder, then they need to kind of replicate and distribute your 100 or 200 Personas across those three roles. What's the split? Right. How are they doing it?
All of those kind of help you get more realistic data and that's typically embedded in the algorithms that these platforms built.
[00:33:32] Speaker B: Yeah. Okay, cool.
So, so I think that, that, that makes sense.
It's a great, great use for if you don't have enough data. And then the ongoing research in the background I think is really interesting. Did you say that's Something that you are doing or that's something you're trying to figure out. How to.
[00:33:52] Speaker C: I'm trying to figure out, but there's some of these platforms already, so rally. I think it's launching their API soon and then you can start working with tools like N8N to make that automated.
So that's very super interesting and I'm excited to try it out.
[00:34:09] Speaker B: Yeah, I mean you could theoretically just build that into any workflow. Hey, here's my newsletter before it goes out. It automatically goes.
There's an NN trigger goes and validates and then makes a. Yeah, that's very cool.
So let's talk about the content production and AI. We've talked about content validation and AI.
I know that you have a framework called Path.
Walk us through that framework.
[00:34:36] Speaker C: Yeah, so this is kind of the framework that I came up with when I was building my own Personas. Right. Just because of need and necessity. So the first one, it's the P, the prepare. So in that sense it's basically conducting real human research or trying to collect as much as possible.
Typically I try to cover three areas in my research which are the internal research, which it's anything about speaking with the team, my client's team, learning about the product, looking at support, chat transcripts, all of that.
Then the external side of research is divided into the prospect, customer and also non buyer. Right. So trying to understand all of those three areas.
And finally the market research is looking at competitors, like we said, competitors, reviews. So trying to cover as much as possible of those three areas in the preparation, preparation.
Then the second stage is the articulation and this is where we basically generate our AI Personas. So in this case we feed all of the research to my Personas. I typically if you do it manually, which is not going to be as accurate as if you use some of those platforms because of those algorithms and all the things that we mentioned. But I mean if you want to get some quick testing done, I tested a couple of email sequences, I tested headlines and it works pretty well. Just to give you confirmation maybe if you have a lot of different variants that you want to pick from.
And in the articulation I try to use LLMs that have a big context window. So for this purpose I used to use Gemini 2 point. Now there's 2.5. That's because it had a biggest context window. Now I think most LLMs, even chatgpt with projects have pretty big context windows, so you could probably use any of those. But yeah, I would keep that in Mind as a factor. What's the context window? So that you can feed as much as information as possible, because those transcripts tend to get quite long when you feed them.
And then I basically prompt the AI in a separate chat. So I have one chat, which is typically my marketing chat is kind of the strategist or assistant. Right. Helps me come up with ideas or variations. And then we have the actual Persona chat or different Personas chats.
And I use a simple prompt that basically instructs the AI to wear the shoes of this Persona, Never break character. And also I like to give it like another tag that's called thoughts, so that at the beginning of every response, it gives me their thoughts, like what's going on in their head? And then it expresses those thoughts in kind of like what they are actually saying virtually. Right.
And so, so this is kind of the Persona. This. The next stage is the test stage. So in this case, we want to run different scenarios, probe assumption, simulate different responses, probe for objections. All of the. It's pretty cool because everything that you can come up with, you can do it. So it's like having a conversation with any of your customers. You can ask them any questions. So at the beginning, it's quite spooky and.
But pretty cool.
And the final stage, which is the H Harmonize, it's basically where we combine all of the findings that we have and then we relaunch another round of research. So maybe some of the findings pointed out that this Persona resonates more with specific messaging angle. We might test that in our sales conversations and see how those perform. So what the reaction of real human Personas are right.
Refeed everything into the cycle. It's basically like a flywheel, Like a cycle.
So, yeah, this is pretty much the path framework that I came up with.
[00:38:25] Speaker B: Yeah, I love that.
And with this, you're saying one of those platforms that you talked about, like you would operate this entire thing in a single platform, or are you kind of mixing any of these for different purposes?
[00:38:41] Speaker C: For those platforms can actually handle everything for you, so you don't really need to switch. If I did that internally with different platforms, it's a bit more. Obviously, it's a bit more disjointed, but you also have some control as well, some fine tuning and control. And so for when I did it internally for myself and my clients, I started with Gemini for the context window, then shifted to Claude, mostly for the writing of different variants.
And then after Claude, especially because Claude, if you have like, even with the first paid plans you start stumbling on limits for Claude. So I started to use another platform which is a third party platform that plugs into Claude's API. And it's nice because I can use it with my team as well. So you can collaborate on the same chats and especially once the research, the messaging strategy is done, then we basically use the same client project and, and each of us writes with the same exact voice, hitting the same exact messaging pillars using the same research. And so it's pretty cool to see how also faster it is if you have maybe a new team member come in. How to have them write copy. That's almost 90% good first draft, even if they have never jumped into any of the research that we did for clients.
[00:40:04] Speaker B: Wow, that's, that's really cool.
I, I feel like I already know your answer to this question, but I'll ask it anyway, which is like, I think a lot of people are worried about AI and copywriting, if that's their skill set.
Do you think that fear is warranted? And I, I'm pretty certain your answer is going to be no, but to drive a little deeper down into that. So like, what do you think is the, will become the role of the copywriter in the future as AI continues to develop?
[00:40:43] Speaker C: Yeah, yeah. Obviously I'm a bit biased, but I would say no, but in the sense that like I'm, what I'm seeing, it's all the, all the copywriters that are anti AI, then you ask them how, like what tools did you actually use? Like what's your experience? And then they might say, oh, I use the free version of ChatGPT and that's it. And that's probably 80% of them. Right. And so like it makes me wonder like how deep did you actually go into actually trying these tools? Because that's not been my experience.
And I say if you rely on formulas, templates when you're writing copy, then probably AI can replace you. But the thing that it can can't really replace you now, but I would say probably in the future as well. It's if you have a strategic vision. So if you know that the copy comes from the research work, then there's the strategy in between and then you can't really write any word without all of those foundations, then I would say you can still be the effective copywriter using AI. Right. So, and I think the, the role of copywriter moving forward is going to be more and more of the architect or kind of orchestrator of these kind of AI systems just because a lot of the writing is basically done for you. Right. So you need to kind of know when to jump in, when to be the human in the loop inside these systems, how to set them up, and of course, have that copywriter's intuition or taste for what works. That's super valuable when you go in and edit the final copy, because there's still a lot of editing, and I suspect that's probably never going to change. You still want to be kind of the human to kind of empathize, especially if you immersed yourself in a lot of research to start with, you know, kind of what works and what feels human. So that's probably the last step that it will still maintain.
[00:42:44] Speaker B: Yeah. And I, I think, tell me if you agree with this, but it seems like the human, like you mentioned, immersing yourself in the research, it's still going to be important. And I think for some people, the temptation is going to be like, great. Well, if I've got these AI tools, I don't even really understand or look at this information. I could just feed it in here, get the output, and it's like, done. I was like, nah, actually you kind of still do need to do that part of the job. Would you agree with that?
[00:43:12] Speaker C: Yeah, yeah, totally. I still, I mean, now I love doing customer interviews. So, like, I wouldn't give them up for like, any. Any reason.
And it's still important because I think, like a lot of these platforms, synthetic research, can actually do, like, interviews for you, like written interviews. I interview your AI Personas. But I think one of the most valuable aspects of doing real customer interviews, when you are in front of the person, either virtually or in person, it's actually to be able to follow the conversation, even if you have a script of questions, to be able to follow the conversation intuitively and see where it goes. Like, follow wherever it leads and kind of being able to ask more nuanced questions based on where the conversation goes.
So it's not just as scripted as just having like a. Responding to a survey. Right. So that's super important, I think, to still be the human immersing yourself. And also same for looking at a lot of reviews. Like, you want to kind of jump in and absorb some of the language. So when it comes time to actually doing the writing or editing, if that's most of what you're doing, you will still have that taste and that sense of what works because you absorbed it.
[00:44:25] Speaker B: Yeah. I think that's some really great parting advice for our audience.
Thank you so much for joining us today. I feel like what you've talked about is people often say AI is about augmenting and making you more powerful, but at the end of the day, how many people are actually doing that? And I think what you've laid out here is a really great framework plan to actually augment and not just replace or in some cases have some sort of subpar output. But like this is actually allowing the human to do what the human does best and allowing the computer to do what the computer does best, you know.
So thank, thank you for highlighting that and that framework path. We're going to write something up about that on which you can
[email protected] Also Chris, where would be the best place for our listeners to connect with you?
[00:45:24] Speaker C: Yes, so you can find me on my
[email protected] where I have a scorecard that you can try free and kind of gives you an idea for where you stand in terms of copy strategy. And also I'm typically on LinkedIn, so you can find me there, connect and we can have a chat about anything.
[00:45:41] Speaker B: Okay. All right. Well, thank you so much for being on the show.
It was a really great conversation. Appreciate it.
[00:45:47] Speaker C: Thanks so much Maxson.
[00:45:51] Speaker A: That's all for today. Thank you so much for listening. If you enjoyed this episode, please consider leaving us a five star rating and review and subscribing so that you don't miss future episodes. Huge thank you to Chris Silvestri for joining us today. You can Find Chris on LinkedIn if you want to talk with him more about conversion copywriting, you can also see his
[email protected] Some major takeaways from this episode are number one, messaging isn't positioning and mixing them up derails your strategy. Positioning is what you stand for. Messaging is how you say it. You need to say both separately and intentionally.
Number two, the best copy doesn't start with writing. It starts with research. Chris said that 70% of what he does is research. Customer interviews, surveys, reviews. Great messaging is built by listening first. Number three, AI can scale your workflow but not replace your thinking. You can use AI to test, validate and extend your insights. But at least for now, human strategy and taste still matter most.
You can find past episodes of the campaign and marketing tools and templates@nicenf floor.com also you can learn more about the agency and get in touch with the marketing specialist if you want support for your own marketing campaigns.
Please join us next week for a conversation with 97 Floor's own head of content marketing. Rachel Bascom. We're talking about what it means to be audience first and why it is so critical in 2025. Till then, see you.
