
Digital Coffee: Marketing Brew · 2025-10-08 · 24 min
Key moments - from our scoring
Substance score
45 / 100
Five dimensions, 20 points each
Amy Klein breaks down a practical methodology for extracting actionable insights from social media data. Rather than treating metrics as a report card, she advocates diving deeper to understand why posts perform - examining content formats (video vs. stills vs. carousels), messaging strategies, hooks, and CTAs across high and low performers. She emphasizes moving beyond engagement rate as the primary metric, noting the rise of "lurkers" who consume but don't engage publicly, and instead aligning metrics to specific business goals: awareness (views and reach), consideration (engagement), or conversion (clicks and UTMs). Klein also addresses the growing role of AI in content workflows. Her key advice: start with human analysis and critical thinking before supplementing with AI. She illustrates this with a ChatGPT reporting error where AI flagged St. Patrick's Day posts as high performers when the actual drivers were UGC content and a specific cocktail-making video format that performed well across contexts. She positions AI as a "competent intern" useful for heavy lifting and editing, not primary content creation. Looking ahead, Klein predicts that as AI becomes more prevalent, audiences will crave authentic human connection - live videos, organic stories, and employee-driven content will outperform polished AI-generated material. She highlights Buffer's employee engagement strategy as exemplary, with multiple company voices building relationships rather than a single brand face.
Start by defining your business goal - awareness, consideration, or conversion. For awareness, track views and reach; for consideration, look at engagement and comments; for conversion, monitor clicks and use UTMs to track traffic to your website. Different posts in your calendar can optimize for different goals depending on your strategy.
People increasingly consume content without engaging publicly - they're "lurkers" - and often engage in private DMs instead of on the post itself. This means high-performing content may show lower visible engagement, so you need to choose metrics aligned with your actual business outcome rather than just engagement numbers.
Treat AI as a competent intern: do your own human analysis first, provide AI with specific high-performing examples and tactical patterns, let it generate drafts, then edit heavily to remove AI fingerprints and verify it hasn't hallucinated. Research shows starting with human thinking then editing AI output produces better results than starting with AI.
As AI-generated content increases, audiences will seek more authentic human connection - live videos, organic stories, unpolished behind-the-scenes content, and employee-driven narratives will outperform polished or AI-generated videos. Brands with multiple employee voices and genuine human faces will build more loyal followings.
Jobs will shift rather than disappear. Marketers who can critically evaluate AI output, provide strong inputs, catch hallucinations, and think strategically will become more valuable. Those who simply rely on AI without critical thinking will be replaced by marketers who know how to use AI effectively.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode contains some useful tactical advice about analyzing social media metrics and using AI as a tool, but much of the substance is recycled wisdom (AI as an intern, start with human work first) wrapped in extended explanations. The core insights - look for patterns in high/low performing posts, align metrics to business goals, use AI critically - are solid but not novel, and the episode pads heavily with conversational filler and tangential anecdotes that dilute the per-minute insight rate.
pull the findings these are the tactics that are really working. And I'm going to start with copy.
it's not AI that's going to take a marketer's job, but a marketer who knows how to use AI will end up taking that job
The frameworks presented are conventional: the AIDA-like funnel of awareness/consideration/conversion, the 'AI as a competent intern' metaphor, and the emphasis on human authenticity in an AI era are all widely circulated in marketing discourse. The St. Patrick's Day ChatGPT hallucination example is instructive but anecdotal rather than a fresh conceptual contribution. No counterintuitive claims or first-principles thinking challenges the status quo.
with AI becoming more and more prominent, people are going to seek human connection even more
I like to think of AI as like a really competent intern
Amy Klein is a self-described social media and content strategist with some credible experience (SaaS and consumer brands, digital events reaching 10,000+ people), but the transcript provides no evidence of operating at significant scale or in a senior leadership capacity. She appears to be a freelance consultant/coach rather than a practitioner who has built or scaled a major revenue-generating function. Her insights are grounded in hands-on work, which is valuable, but lack the weight of someone who has operated a P&L or driven transformative business outcomes.
she's led standout campaigns for SaaS and consumer brands, coached execs on LinkedIn Thought Leadership
I've been doing a lot of testing with AI to see if it can help me get faster with the reporting
The episode relies heavily on abstract frameworks and generalizations with minimal concrete data. The St. Patrick's Day ChatGPT example is the main specific case study, but it illustrates a failure rather than a success metric. References to 'Buffer' and an unnamed marketer's interviews with '20 different marketers' provide some grounding, but no actual metrics, ROI figures, engagement rates, or replicable numbers are shared. The advice remains at the tactical level without quantification.
This was for a craft cider company. So it was the process of creating a cocktail using the cider
There is, there's a brand who I really admire on LinkedIn while they're all over, but I, I follow them on LinkedIn most closely and it is, it's actually a social media scheduling software, it's called Buffer
The host asks reasonable open-ended questions and shows genuine curiosity, but rarely pushes back, challenges claims, or dig into contradictions. When Amy makes broad assertions (e.g., 'people are going to seek human connection even more'), the host accepts them without pressure for evidence or counterexamples. The conversation meanders through pleasant anecdotes (coffee vs. tea, the MIT study with cheeky researchers) rather than pursuing sharp follow-ups that would deepen understanding. The interviewer is friendly but not analytically rigorous.
Got you. Yeah. I mean, how do you actually do that?
And then I mean looking for trends ahead, like what are you seeing shifts in
Computed from the transcript - who did the talking, and the words that came up most.
Diving deep into the realm of social media, the podcast features a dynamic duo - Brett and his guest, Amy, a bona fide whiz in content strategy. They kick off with some light-hearted banter over their beverage preferences (coffee wins over tea, obviously), but soon enough, they plunge into the nitty-gritty of social media content strategy. Amy emphasizes the importance of data-driven insights, advocating for a curious mindset when analyzing metrics. She reveals her systematic approach to dissecting high and low-performing posts, encouraging listeners to identify patterns that can inform future strategies. Curious minds will find her breakdown of what truly drives engagement - hint: it’s not just about likes and comments, but rather the deeper, often unquantifiable connections that content fosters. As they navigate the intricate web of social media marketing, Amy highlights the shift towards authentic, human-centric content, especially in an age where AI is making waves. The episode wraps up with some sage advice on leveraging AI without losing the personal touch, reminding marketers to remain vigilant and engaged with their audience.
Transcribed and scored by The B2B Podcast Index.
Foreign. That's good. And welcome to a new episode of Digital Coffee Marketing Brew. And I'm your host, Brett Dyster.
If you please subscribe to this podcast and all your favorite podcasting apps, leave. A five star review does help with the rankings and let me know how I am doing. But this week we'll be talking about social media content strategy. The things that we all know we should know and sometimes we forget that we know or don't know a lot about it anyways.
But anyways. With me is Amy, and she is a social media and content strategist who's turned data into stories and audiences into communities. She's led standout campaigns for SaaS and consumer brands, coached execs on LinkedIn Thought Leadership, and hosted digital events that reached 10,000 plus people. Her work helps brands connect more clearly, creatively and strategically.
So content drives real results. But welcome to the show, Amy. Thanks for having me. I'm really excited for our chat today.
Yes, the first question, that's all my guest is, are you a coffee or tea drinker? Coffee. But when I was pregnant, I switched to tea, so I didn't have to do all the math of the caffeine. Back to coffee now.
So I have a healthy appreciation for tea. Yes. Because tea you can get not. Well, you can technically get decaf coffee, which I think it's more like it has low caffeine.
It doesn't have any caffeine, but it's very low. That's true. Sometimes if I want the flavor, then I'll go for the decaf. Got you.
And I gave a brief summary of your expertise. Can you give your listeners a little bit more about what you do? Sure. So I work in content strategy and I really focus on social media.
But what I found very much moves the needle, is getting very curious about your posts and specifically your metrics and your data. Not just using your data like a report card to figure out how you did, but diving in a layer deeper, figuring out why the numbers are where they are and how you can leverage those insights in order to strengthen your strategy overall. Got you. Yeah.
I mean, how do you actually do that? Because I'm pretty sure a lot of PR people and marketers are like, yeah, the data says we're doing well, but I don't know why we're doing well. Honestly, a lot of it is just curiosity and asking questions. So I have a system and if you don't mind, I can just jump right in with talking about it.
That's quite all right. I have A system I implement with all of my clients. And we start with high level of looking at the posts, the metrics overall. But then we'll look at the posts and we'll group them into high performing posts and then we'll group them into the lower performing posts, starting with the high performers.
The biggest thing that you can do that will benefit you with these high performers is to look for different patterns or trends. And what I mean by patterns or trends is look at the content styles, look at maybe video versus still images, graphics, carousels, what is performing best? Are there any that really stand out? Are there any that consistently stand out?
And then you can look at things like your messaging, look at your hooks, look at your content pillars and it's the same what is standing out and is there anything that is consistently standing out? When you do that, you can figure out what tactics are higher performing and then do the same thing with the lower performing tactics and then you kind of have a roadmap. Okay, next time I plan out my content, I am going to do these tactics more often and these tactics less often. You might need to use one of the ones that are lower performing for a different strategic, business focused reason.
But if you lean heavier into the higher performing, then you can start to strengthen things and that's kind of where it gets fun and you can get a little bit more nuanced. So I mean, how do you determine the high performing, low performing? Is it more like engagement and comments? Is it more towards like getting them to your website?
Is that specifically to the business? Because I mean higher performing, low performing can be subjective, it seems like. So how do you figure that part out? That's a really good question.
I used to live and die by engagement rate. So if you asked me this a couple years ago, I would say, oh, just look at the engagement rate and you'll figure out how the content is making impact. When people see it, what is actually impacting or moving them to engage. I'm moving away from that lately because we're kind of in what some people call the age of the lurker.
So people are consuming the content and they might be interested in it, but they're not engaging as much. And when they are engaging, oftentimes they're engaging in the DMs so they might see something that really resonates or that reminds them of somebody and then they'll share it with that person and they could have a friend full on conversation with that person in the dms, but it's not showing up as engagement on your post. So kind of like you mentioned earlier, it will depend on what your overall business goals are.
So if you're looking to gain more awareness, you're going to lean heavier on views, you're going to lean heavier on reach. If you're looking more in the conversion, you'll look at different engagements and then getting into the. Oh, sorry, conversions, you'll look at different engagements and then getting into the clicks. And I know on certain platforms like Instagram that could be a little bit more difficult.
That's where we start getting into UTMs and, and things of that nature. But these, these metrics will depend on your business goals and then when you're doing that, you can start to optimize towards those individual goals. And a really important caveat there is you don't have to focus on just one goal for your overall content planner. You can actually take that strategy I laid out and optimize your posts throughout your content planner.
So some posts may be your posts to get more people, some posts may be your posts to get them to know you a little bit more and to consider you, and some posts may be aimed at converting them. When you're looking at your data, you can sort by the different metrics depending on what you're looking for. And when you're developing those posts, that's when you start to implement the tactics that are working and you can thread them all out throughout your calendar. Got you.
And then I mean for content wise, AI is always the hot topic for everything. I mean we see like AI videos and VO3 is really taking over LinkedIn and everything else. So how do you like, like create those high engaging or high performing things utilizing AI? Because not everybody has the know how to use video editors or Photoshop or whatever.
So how, how would you tell like those that are maybe overstressed or one man teams, how do you can create those high performing content that will actually work? All right, I'm going to give you some advice and then I'm going to give you a warning. So my advice is to look at your data, but look at it very critically. And I would say start from a personal human point of view.
Eventually you can get into the AI when it comes to the actual reporting. But if you don't take the time up front to put in your personal point of view and actually dig into it with your hands on, you won't know when AI is hallucinating. And I can give you a great example of this. I've been doing a lot of testing with AI to see if it can help me get faster with the reporting and then with the content creation, but specifically with the reporting.
I put in a month's worth of data into ChatGPT and I, you know, gave it a prompt to sort by a specific metric and tell me what the high performers and the low performers were and which commonalities it spotted. It did give me some good insights. It did, but. But it also flagged that this was March reporting.
So going back a little bit, that St. Patrick's Day posts were doing great and that I should post more St. Patrick's Day posts. Now, obviously I'm not going to post St.
Patrick's Day posts not in March, but even if you translated it into like holiday posts are doing great, it's a little bit higher level or a blanket statement. I knew that that was false because I had previously looked at holiday posts specifically the things that were making the post perform. I knew after looking at them because I went back and I was like, hang on, hang on, what's going on? So the things that were making them perform weren't the fact that it was St.
Patrick's Day, it was the actual content in the video. So one of the really high performing posts was ugc and it was a carousel of UGC images that is going to do great. The other one was a specific video style that we had been developing that kind of showed the process. This was for a craft cider company.
So it was the process of creating a cocktail using the cider, and it was a St. Patrick's Day video. But that type of video had been performing really well in different contexts. So I knew even though Those were both St.
Patrick's Day posts, the reason why they were doing well wasn't because of St. Patrick's Day. So if I hadn't done all of the reporting ahead of time, if I didn't start by doing my own analysis and then supplement with AI and then also double check AI's work, then I would have thought, oh my goodness, St. Patrick's Day posts are great.
Holiday posts must be great. I should post about all of the holidays, when in truth, holiday posts tend to be some of our lower performing posts. So we actually kind of stay away from them when possible. So I guess that was my warning wrapped up into that.
Use AI, but make sure you start with your own work when it comes to using AI to create content. So taking that a step further and applying the findings to the content, what you can do is in the process that I just laid out, you can pull the findings these are the tactics that are really working. And I'm going to start with copy. Copy is, I would say, the easier thing to create, but you can pull different tactics.
These are the types of hooks that are really working. These are the topics, the content pillars that are really working. These types of CTAs are getting me a lot of engagements and comments. And if you pull those out, you can also pull out some examples so that you could say, you don't just say like, I don't know, question hooks are really working.
I'm pulling this out of thin air. Or hooks that have a list of three different things that are really working. You could say here's what it is and here are two examples that were high performers that can help you train your AI to implement those tactics more. AI does have fingerprints, so if you're not then editing it, then the AI fingerprints will stay.
And one good example is if you do that like list of three, it'll have, it really likes the alliteration. So it'll have two things that make sense and alliterate and one that's just kind of random and they just tacked it on there because it had the same alliteration and it thought it sounded good. So once you do that, like it can do some of the heavy lifting for you and it really can help you speed up your workload. But once you do that, make sure that you are editing it afterwards.
Got you. So like for example, for me, like for podcasting, I let it do like the show notes because I've already created it. I'm just letting it reinterpret or jot it. Be my little secretary and jot down what I'm.
What people are saying gets my last name wrong all the time. I have to fix it. But it's just one of those things where you do the work or you create the actual like primary content and you do let it do the secondary content. Is that what I'm hearing?
Yeah. And then you just edit it afterwards. So I like to think of AI as like a really competent intern. If you can give it the right input, then it's going to give you some solid work.
But you should always check its work because it's just a really competent intern and it's not a senior level employee. There was actually a study done by mit, I believe it was, that talked about AI and the AI use. And one of the things that it found was that if people started with AI and did the primary work with AI, their brain wasn't firing when it was doing the editing, it didn't recall. But the people who started as a human then got the AI input and then went back for the editing, their brain was lighten up.
So having that first stage is actually really going to strengthen not just your own understanding of the work or your own intelligence, but it's also going to strengthen the work itself because you're putting out things to represent you or your business. Yeah, I think I saw that study. I think it also said that if you over rely on AI too much, you actually get dumber because you aren't, you're relying on something and you're not using your brain or critical thinking as much. Yes, and, and I could go on a little tangent about this study because the people who wrote the study were very, very cheeky.
And they wrote it knowing that people would put the study into ChatGPT to get summaries. So they actually put in tags like if you a LLM, like just read this part and it gave like notes so that they could shape what the output was so they could see which news outlets actually read the study and which ones just threw it into chat GPT and got the answers. So it only told half the story. But yeah, that was, that was part of the story.
So yeah, I mean, for another example, my other podcast, I re news articles, but I let chat, I don't know, I let my RSS feed that has AI like bullet point the objective things, but I still read it so I understand what I'm talking about. So is that, is that kind of like the competent way of doing it? Like you read it, but then you let chat, chat CPT or whoever, Gronk or whatever you decide you like to like bullet point it for you. So at least you get kind of like the bullet points.
But you still read it and you understand it, you digest it. Just kind of like your content, you create it. But then you let chat GPT or whatever help you maybe with some grammatical errors or whatever. Because we're human, we, we do it all the time.
No one's perfect about writing. Absolutely. Absolutely. No, I think that that is spot on what you should do because then you have the knowledge, but you're also streamlining.
It's much more efficient to have someone else bullet those out as long as you're, you're reading it and you couldn't speak to it. Got you. And then, I mean, looking for trends ahead, like what are you seeing shifts in, are you seeing more shifts in more video usage, AI, video usage? I mean, I think I just read something about an ASMR video that was popping off on social media because VO3 made these like gelted looking fruit and they were cutting through it.
It was weird. I was like, I don't do ASMR. I just saw it on LinkedIn. I was like, this is so weird.
Like, are we getting more into that? Because I, like, I'm not. I'm kind of like rejecting or doing my own cognitive dissonance on some of that stuff. Are we?
Because I'm rejecting it a little bit. I'm like, I don't really like this that much. I think it's cool. But I don't know what are the trends coming up that marketers need to at least be aware of so they just don't avoid the pitfalls or at least understand where like editing tactics are coming from.
I think that's a really good question because I think a lot of people are rejecting it now. I spoke with a marketer who I met through LinkedIn and she had been holding interviews with upwards of 20 different marketers and she was asking people the same question. And I was towards the end. And she told me that by the time she got to me, my answer was actually the common answer that everybody's kind of been talking about, and that is with AI becoming more and more prominent, people are going to seek human connection even more.
And human connection that they can be sure is actual human and not really, really, really good AI pretending to be human. So if you think of things like live videos or even like live webinars, the type of content that you might put on your Instagram, Facebook stories or a Snapchat story, the content that's a little bit grittier, a little bit more organic, less polished, that is going to be the content that people are looking for because they want something real and people go on social media for human connection.
So the less human things become, the more humans people are going to seek out. And I do think that there's going to be an influx of AI and I'm sure some of the AI videos will see stick. But in all of that, people are going to gravitate even further towards the human element. So this is a great opportunity for a brand to have a face of the company or multiple faces of the company.
There is, there's a brand who I really admire on LinkedIn while they're all over, but I, I follow them on LinkedIn most closely and it is, it's actually a social media scheduling software, it's called Buffer. But they have an amazing employee engagement strategy. Their employees are kind of their content creators in and of themselves and they talk about their own expertise, they talk about the challenges that they're facing at the company and overcoming, how they're, how they're fixing it, how they're working it.
And in that they're creating multiple touch points with different people online. So it's not just having one face of the brand, it's having many. So maybe I don't connect with, I'm going to make up names. Maybe I don't really connect with sue because she has a very overly technical way of speaking, but I might really connect with Jen who thinks a little bit more creatively and then maybe there's somebody else who likes that other kind of speaking that more technical and connects with that even more.
That kind of approach is really valuable. And not only do I think that's going to happen, I hope that's going to happen a little bit more because that is going to really build relationships between a company and its followers. This, this brand in particular has a very, very loyal following. They do a lot.
Right. But one of the big things is that outreach. Gotcha. So what you're saying is that the scripts, AI avatars aren't going to be, aren't going to be popular?
Because they do. They just released it, I think like a couple weeks ago, where you can actually like make yourself look like an AI avatar. And I'm like, I have the script. I'm like, I'm good, I'm good.
I'm going to be myself. I think it's going to like bounce over this way a little and then swing back. So people will start to do that and they might do that for the novelty of it, but some might stick. I don't think that that's going to be the way everything trends, at least at the time of my speaking right now in June of 2025.
I don't see it going that way. I mean, do you, do you see like a lot of like, like low level, like social media stuff jobs, like going away because of AI just because it's easier just to have AI interpret things or maybe do some low level content stuff so you don't need to hire somebody else? Do you see that happening in the future a little bit more often? Because I mean, they're all going to eventually take some jobs away.
We're just not really quite sure which ones right now. I think it's less of the jobs going away and more of the jobs shifting. I have to give credit to my sister on this one. She told me, when electricity was invented, lamplighters became obsolete, but all of a sudden we needed electricians.
So maybe there will be a subset of jobs that will go away because of AI. I think that it's going to be even more important to learn the skills that are going to make you successful with AI. Many people, I've heard many people say this, it's not AI that's going to take a marketer's job, but a marketer who knows how to use AI will end up taking that job because it will just be a person who can work more cross functionally, they can do more general activities, they can do it more efficiently.
But as I mentioned earlier with the reporting, you still do need to have the critical thinking skills to give solid input. You know, garbage in, garbage out. So if you can give that solid input and then and check on, at this time, at least AI is still hallucinating. So check and make sure that the work is solid, that you're able to edit out the AI fingerprints so it doesn't turn into being just another robot.
Got you. And people listening to this episode, they're wondering where can they find you online? To learn more about you, LinkedIn is the best place to find me. My name is Amy Klein and I have red hair.
So if you see there, Amy Klein is not an uncommon name. So Amy Klein with red hair, that's me. All right, any final thoughts for listeners? None that I can think of right now.
Thanks so much for having me. Yes, thank you Amy for joining our my podcast and just speaking on AI content strategies, content creation and social media strategies, of course. Have a good one and thank you for listening. As always, please subscribe to this podcast and all your favorite podcasting apps.
Leave. A five star review really does help with the rankings. Let me know how how I am doing and join me next week as I talk to another great thought leader in the PR marketing industry. All right guys, stay safe.
Get to understanding how AI can help you and how you can fix AI's errors and see you next week later.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.