Pipeline · Episode 9 · Nov 11, 2026 · 31 min
Don't Miss the One Sentence That Kills a Deal (with Alex Atkins from Sturdy)
Alex Atkins, VP of Growth and Operations at Sturdy, built a quarterly report called Churn Is Not a Customer Success Problem from first-party data drawn from tens of millions of customer conversations. It became Sturdy's single highest performing asset, distributed mainly on LinkedIn. Alex also uses AI to read every account conversation instead of trusting forecast opinions.
With Alex Atkins, VP of Growth and Operations at Sturdy
Key takeaways
- Mine your own first-party data for a story nobody else can tell, the way Sturdy analyzed what Alex Atkins put at "like 47 million" inbound and outbound business conversations.
- Title the piece with the claim your top persona already suspects, as Alex did with Churn Is Not a Customer Success Problem, so CS leaders read it to defend themselves and CROs read it because it said out loud what they already suspected.
- Go in ready to be proven wrong. Alex expected the data to blame CS for most churn, and it showed almost 75% of churn sitting outside CS and CX control.
- When a piece flops, keep the data and change the angle, the way Alex's first version, called something like "AI in Business," evolved into the churn report.
- Replace the weekly round of "who thinks what's going to close" with AI reading every account conversation, so one buried sentence in an email doesn't decide a deal without anyone noticing.
Sponsored by Podcast 2 Pipeline
Prefer audio? Listen here
Chapters8 chapters
TranscriptThe full conversation · 7,346 words · tap a time to play from there
If you're in GTM and B2B, you probably want to be generating top of the funnel content to be driving pipeline, and that can be pretty tough. And then how do you keep track of the deals and everything as it's coming in? That can also be pretty tough. Welcome to Pipeline. I'm your host, Joseph Lewin, and I'm super excited to be joined by Alex Atkins, the VP of Growth at Sturdy AI, uh
That's
who's
correct.
gonna be walking us through your content strategy and what you're doing and then how you keep track of deals. So I'd love to start high
Absolutely.
level with just how do you think about content and and how do you approach content to help drive top of the funnel interest?
That's a good question. I mean, I've also done this wrong so many times in the past. I think it just takes time to kind of test these things out, figure out what's going to resonate with your, you know, top personas in your ICP. But I figured there's there's generally one solution, and it's the tough road to take. And I think it's more difficult these days because of all the generative content out there, all the noise on things, you know, LinkedIn. I go on LinkedIn these days and I'm seeing like probably 10 to 15 posts that are clearly generated within 30 seconds, probably not read over by a human, and it's just creating so much noise. And so people are getting more and more used to that noise, and it's getting more and more difficult to find content that I think really resonates, um, or that teaches you, or that educates you and entertains you. And that's something I keep coming back to.
One of the big things I learned early in my career is I want to create really cool content, make sure it's educational and entertaining. So, you know, a little while back, just to kind of put this in into context and provide an example of what I might might mean by this, is that I created a report a little while ago called Churn Is Not a Customer Success Problem. Um, customer success, CX, being one of the top ICPs and personas here that, you know, at Sturdy.ai where I work. So I figured, okay, so that's a top persona, it's a top group. Let's do some digging into all the first-party data we have. So the title did most of the work, I think, but every CS leader read it to defend themselves, to scratch that itch, right? And every CRO, you know, um read it as well because, you know, it said out loud what they already kind of suspected, which is churn is a revenue problem that gets handled by the wrong team, um traditionally. But it's always stuck on that team, the team that actually doesn't have, you know, the tools to to fix.
Those problems because it's traditionally a reactive process. So the data was the unlock. And we had analyzed tens of millions of customer conversations and account conversations at this point. Uh and these are again, just to be very clear, um probably I think it was like 47 million conversations inbound and outbound between uh customers and uh our customers and their customers, really kind of understanding what kind of conversations we're talking about, all business conversations through email, chat, Uh, you know, uh, support tickets, uh, Zoom call transcripts, anything and everything you could ever think of, Slack, you know, um, anything that you've ever communicated with your customers on, any platform gets basically ingested into something like what we built. And then it goes through all of that unstructured conversational data, which is a massive amount of data at any company that you can't glean any insights from it 'cause humans weren't built for it. Um. To be able to basically kind of come up with this understanding like, okay, so when it comes to churn, you know, what is causing churn based on this conversational data that we have that nobody else has.
So I think that's a
Yep.
big piece right there. First-party data. What is a cool story we can tell with first-party data? So we showed where the churn actually originates. And it's product gaps. It's overpromising in the sales cycle, in the marketing and sales cycle. It's broken onboarding. And CS, you know, was cleaning up after all these decisions that were already made months earlier by other departments that they had absolutely no control over. So it became the single highest performing asset we've ever produced. And we're talking um millions and millions of impressions, which I'll get back to impressions and then you know a little bit later than that necessarily what I like to to to basically you know uh understand success based on, but a massive amount of impressions, uh hundreds and hundreds of downloads between the two that we produced. It was a quarterly report, you know, many hundreds of MQL downloads, um outperformed every other distribution, you know, I I think that we've ever
Wow.
made. It was free 'cause the distribution was primarily over LinkedIn, not necessarily through paid. Um, and then, you know, basically after all that was said, the volume of high-quality opportunities it generated with high intent, right? Because this is the right audience, you're scratching the itch. This is what they've been saying for all along. You're finally the first people to provide that first-party data on it, right? So they are so happy to see someone that knows that, that gets that, that provides the information. And then, oh yeah, that information is what they use to actually help these people do their jobs better, more proactively. Um And I think that was just a massive uh a massive opportunity for us to educate, entertain, and get the right people in the door and see us as thought leaders um and as trustworthy folks that they can come to to help them understand their customer intelligence as well.
So just an example on how to go and it's tough. It's not the easy route.
Yeah.
It's really difficult. But but guess what? If you do that, if you go that hard route and you don't go for what's easy, for what's quick, for you know, you you actually get the the return is so much. So much higher. So it's it's it's definitely the right
Yeah.
move.
And I mean, I totally agree with you that people are leaning further and further away from creating that kind of content because
Yeah.
it is
Sir.
incredibly hard. Like I used to be a product marketing manager for an engineering software company and going in, I knew nothing about engineering. And so we had to go in and start learning about, um, instead of just making all the content about our software and what it does and how awesome it is, like there's a place for that. You you do need some content. Where people understand
Further depth.
what
Yep.
you do, but
Yep.
for most of your content to be about what you do, most people don't care. they don't know that you're solving the problem that they have. And so then it took a ton of time being on customer calls or be and and then being on sales calls and starting to dig in. And then beyond that, just learning about engineering in general and the challenges that these engineers were facing that were uh next to what we were doing and you know second sources. And then I would end up spending 3 or 4 months working on usually one article that was super well-researched. Now it wouldn't be the same today, but it was an insane amount of time, but then it would
Oh yeah.
rank well because it was on a topic that people cared about. And then that would drive people into, you know, the next section of the funnel. And,
Right.
and then we'd end up being able to get them to the content that actually told them about what we did. And then
Right.
they'd
Yeah.
move forward because they understood the problem, right?
Yeah, let's
So,
think about that for a second. You're absolutely right
yeah.
because the high intent, that's why they came in. They wanted to learn from somebody who's a subject matter expert that they could trust in, right? So they actually want to learn more about your company because they learned that you previously, that you're somebody you can trust in. There's someone that has the information. You're somebody that can provide that information, obviously put in the work to do so, right? So that's why they got interested in the company because also, dude, people follow people. Typically, right? Traditionally, they don't really follow companies necessarily. In some cases they do, but if they saw you talking about it and they saw that you had, you know, worked this up, maybe provide some video content on it- I published this thing, I put it out on LinkedIn, people came to me. They didn't really come to Sturdy, but they got interested in Sturdy after they learned from, you know, from me and what I provided and what they sent them.
So I think you're absolutely right, that's how you get them down, you know, get them to the point where they want to learn about the company. You're not forcing it down, down their throat.
Yeah, yeah. And I mean, I think it's really interesting because you're hitting on something that I've thought about and worked on a lot, which is If you can find- and I'm gonna ask you a question about this in in a second-
Sure.
if you can find pain that people have
Yep.
in their job, that pain is what's going to ultimately motivate people. If they
Yep.
if they're thinking of, hey, it would be nice to do this in the future, it would make my job, you know, more effective or multiply what I'm doing,
Yeah.
you know, that's a hard sale. But if they're going, my job sucks because of XYZ reason, or, you you know, I love my job in every aspect, except for this one thing that I spend way too much of my time on, or, you know, it's the one place where I'm getting corrected from my managers, but I can't do anything about it. If you can uncover that pain and then what, what ends up happening a lot of times is that your, your prospect doesn't actually understand what's causing that pain. So they think it's coming from something
Totally.
totally different. And then your job as somebody who's in GTM, whether you're in marketing or sales, I mean, it really applies in in both. You have to be able to now take that pain and reframe the problem to the thing that you solve. And
Totally. Right.
then as soon as you do that, now they're really teed up to go, you understand my pain better
Yep.
than I do. Now
Yep.
you've helped me to understand where that pain is actually coming from. Now I want to understand how do I solve that? And then you can position your product to say, well, that's what we do.
This is what we solve for. Exactly. So so so the the pain point that you're talking about is absolutely true. And I think one thing that you pointed out is sometimes we do the best job explaining to them what that pain point is. Maybe it's something that's been uncovered recently, but you're right, they're pointing in the wrong direction or they haven't even really totally realized it. They they they know it in the back of their heads, but they don't know how to like basically explaining to somebody else what it is that's really bugging them, what's creating that really difficult job for them. I think being able to provide, you know, hard, you know, really quantifiable data and and first-party data to show to them that that's actually what's happening. It's not just like my opinion. Hey, I heard this or hey, I saw this. It's like, no, we have 47 million conversations that pointed out that this is your pain and this is how you solve for it.
And I think that also is just so much stronger, right? And then they're immediately like, yes, this is what I've been thinking. But thank you for saying that out loud because I've been thinking this for so long. And obviously with the data back, you know, source, I think just like, you know, 47 million conversations, it's a lot of conversations. There's a lot of data there. So it's it's it's
Yep. Well,
just
that's what I was
cool.
gonna ask you about. So I'll ask you, I mean, basically I was gonna say,
Yeah.
well, how do you go about finding that pain point? And so
Sure.
obviously you guys, you've got a ton of data and you still
Yeah.
have to mine that to find the right pain
Oh,
point.
dude.
So like, how did you go about figuring out
Absolutely.
what pain
Sure,
point to focus
sure.
on?
Yeah, so I came through analytics like originally was sort of my background, um, and a master's in in marketing analytics specifically from Bentley. And it was like, that was always my first instinct is let the data tell you the story. The amount of people that go in with this sort of preconceived notion or goal or motivation to prove is really, really, really, it's really difficult, right? And it's it's hard as human. You want to go in to prove yourself to be right about something. But I went in with the expectation that CS was in fact responsible for the vast majority of negative sentiment and or churn. And to be clear, the way we do this is actually pretty straightforward. Uh we can do it within Sturdy these days by just querying or prompting Sturdy because Sturdy has trained throughout the last 5 or 6 years out of all these millions of business conversations, specific, I don't wanna say keywords 'cause it's not keyword specific, it's phrase.
It's based on um the sentences and and the phrases that are being picked up by Sturdy. So it's not just a keyword, sad, mad, bad. It's like, Hey, so I saw that you were a little bit late to the call today, and that was a bit frustrating for me and my team. What you're gonna get out of that is, you know, response lag, you were late, frustration, negative sentiment, um, you know, probably a sense of urgency. So these are the things that Sturdy's picking up on. We're talking about dozens and dozens of signals, right? So we can see, okay, well, product confusion. I was confused by the product today. I didn't know where this call to action was, or I didn't know where the specific, you know, tool was. I couldn't find it. That's set in an email. Maybe one email says that, or maybe a call transcript picks up. Yeah, dude, we were
Hmm. Yep.
We were a little confused about this um the sheet. We weren't sure where to look, where the you know the averages come out, you know, how you go about calculating those averages, whatever it
Or
might
yeah,
be.
another one, just to kind of add on to what you're saying is
Please.
you've got your sales team that promised something that actually can't be delivered on. So then they you know you have a call where they go, Wait a minute, we were told that you guys could do XYZ, and then the delivery team is like, well-
I have to go back and look this up. Do you have- just a quick throw this in your court because that happens all the time- do you have a pulse on how many uh how much churn is related based on our data, how much churn is related to marketing and sales over promises? Out of 100%, what do you think that is? It's higher than you think, but it's not insane. let's say there's 6 different groupings that we created: sales and marketing over promises, It's a it's it's what do you think it was that that led to churn ultimately?
I mean, I've been in the shoes both of like driving and on the delivery side. So I would
Sure.
say
Sure,
it's
sure,
pretty high,
sure.
probably like
It
30% or something. 30.
was wasn't 30%. That's that's that's I mean it's a good guess. It was 17%. So I'm pretty sure I'll have to double check this. So that means 17% of customers churned because of those exact promises that were not real. They were completely fake, brought on by sales. Purely to get that number hit, right? And what does that do? That then reflects poorly on the CS or CX team because they can't retain those customers. That's not on them at all. That conversation happened long before they came into the mix. So that's one big one. You nailed that one. That happens- the amount of times
Yep. Wait, I
I've
I want
seen
to say
that
one
happen
more thing on
is-
that, because I
go for
I
it.
would say that
Please,
churn
please.
number just having been in both seats, I would say that churn number is only as low as it is because you have a good CX team who's able to recover after the fact and go, hey, you know,
Sure.
let's talk this through and figure out exactly
Sure.
what you're looking for. And then they get really creative on finding a solution, but then
Yeah.
it makes you way less profitable because now your CX team has to spend tons of time to figure out how to custom do something that If
Yep. Maybe
the sales
not
team
just
would
CX,
have- yeah.
maybe the engineering team who has to now go double down on something that, you know, was in the pipeline that hadn't been actually created or worked on
Yeah.
yet that actually now needs to
That's
be
huge.
prioritized.
Sorry,
You're absolutely
I keep interrupting
right.
you. I'll let you finish your thought.
No, no, it's good, man. this is good conversation. No, but you're absolutely right. So that happens all the time. And then you have this, you know, situation with like product. We I was mentioning this earlier. Product was the highest contributor to customer churn. And it was about 25% of all churn was contributing to product. I think CS and CX was around 27% as well. But I mean, just to be clear, that means almost 75% of all churn is completely out of the realm of of control of CS and CX, right? 'Cause it's sales overpromising, it's billing issues, it's product confusion. So if we look at product confusion, I mean like or or just product issues together, I mean bugs, you know, elements of the product that just don't work the way they're supposed to, right? And so there's all these things that got picked up that then led to customers churning. I mean, yes, you had some CF CS as well. Response lag, they didn't get back to me in time.
I've been waiting on my response for XYZ. That leads to then frustration, negative sentiment and ultimately churn, right? So those things exist too. But I think it's just so interesting to see. I mean, this data interested me. That's why I was so excited about it. I did not have an expectation that nearly 73, 3 out of 4 churn, you know, uh companies that churn, churn completely, you know, for for some reason completely aside from CX and CS. it's because of overpromising from the sales team and the marketing teams, from all these other, it's from billing. It's from all these things that are completely out of CS control. It created an amazing story. And I think that story, ultimately, with that data, is what made it such a big, you know, um, sort of viral piece. And created the interest, created the excitement, created this source of truth and trust, and then said, okay, but how did they know this? How did they figure this out?
Fun fact, it's the product that we actually built that you can come check out yourself if you want to learn a little bit more about how it works. So it actually worked. It was teed up very well. And am not in any way saying I did that all like on purpose. I got lucky, you know, things
Yeah,
worked
that's
out well and
part
I got
of
very
it, right,
lucky.
man? It's like
That is part of it. have to be totally truthful. The data actually worked out to be a really good story because if it wasn't, I wouldn't run with it, right? And then the way it was teed up to actually leverage the product to figure out the data points that we're trying to show and prove that worked out the way they worked out, that's also so lucky to be able to kind of tee folks up for that. So I tried a lot of content before I found one that worked. And I think that's also was sort of a learning here is like I experimented with multiple pieces of content and storylines before I found something that really clicked and scratched a niche with a specific audience.
Yeah, and this is going back a ways, but I remember when I was first getting into marketing, I did a bunch of courses from Digital Marketer and
I know
uh
Digital
you know,
Marketer, sure. Yeah,
yeah,
yeah,
Ryan Deiss,
very familiar.
his
Yeah,
team.
of course.
So they were saying of everything that they've ever run, 1 out of 10 of their campaigns would actually truly be
Sure.
like out of the park. And then most of them would be like, break even or maybe just slightly positive. And then
Sure.
a good portion, I don't remember the percentage, but a good portion were complete failures. And
What?
they're- they had tons of data going into all of it. They had,
Sure.
I mean, they were a huge marketing team. They know all about, you know, they were teaching people how to do this stuff. And for them, 1 out of 10 would actually be a home run, which means, you know, that's a 10% ultra success rate. That sounds horrible. Um, but at the
Yep.
same
yeah, for a team that's teaching digital marketing, like, that's like- those are some of the best we have. Um, like, I mean, that's that's as good as it gets, right? So I would just say this to folks, and that happens all the time: if you are working on a really kick-ass piece of content and you have some cool educational, entertaining content that's potentially first party, you think it's gonna really hit home and it doesn't land, that's fine. Maybe just iterate, evolve, figure out a different angle with the same data points. 'Cause that's what I had to do to find something finally that worked incredibly well, but it didn't work the first time. It definitely didn't work. I think my first iteration, like my first piece that eventually evolved into this was called like AI in Business or something completely lame like that. Like, how do you use AI in business? Dude, it was awful. And I literally I had to force it down, just like try to push it. People were just not resonating.
So we had to take it back to the drawing board. That's totally okay. That's fine. Didn't mean we had to change everything. We just had to figure out how to leverage the data from a different angle, look at it from a different angle, and then how to ensure that we put it in the report, include it in the report in a way that would resonate with folks. Um, the ICP, the right people, not just people, right? I think that's another Pretty critical piece there too. So.
I love it. Um, okay, so let's switch gears, um, because you had talked about That's how you've been generating pipeline is,
Sure.
you know, that first-party content. And then, uh, I mean, you guys have a leg up on having all of that data. Um,
Sure.
though there
Yeah.
are other ways to to do that. I mean, I used to just interview
Yeah.
customers and then I would manually go back through back in the day, or now you have
Dude,
all kinds
totally.
of AI tools that make it
Totally.
so much easier. And you're looking for those like across your your ideal customers. What are the words that they use to describe what you're doing and worked well? What are they using to describe what didn't work well? And then it sounds
Yeah.
like your tool basically
Yeah.
can automate a lot of that process and make that a lot easier.
Yeah.
Um, but then now
Yeah.
you've got people in your pipeline. Um, getting
Mhm.
people into the pipeline is one thing, but then what do you do to help serve clients through the pipeline to actually get them across the finish
It's so
line?
difficult. It's so difficult. I think the one thing that everybody like who's listening to this will resonate is that like no pipeline, no two pipelines are the same. It's so funny to be like, well, this is what I do in my pipeline. And it's like, you know, 90 days average, this is what we're looking for ACV. Like this is how we we take these folks from one
Mhm.
step to the next step in the opportunity. And you know, we went from 8 to 9, 9 to 10, and then to close one. It's just it's so hard to map those things out. So I've always kind of come back to like, instead of these sort of static stages and these these calls we have week over week with our AEs to kind of see where things are, get that pulse. You know, how how are we doing with XYZ? Oh, they're great. Like definitely gonna close, probably like 90% close. You know, probably in the next quarter. I'll get them before next month, right? Like these are the conversations we have all the time. But I I I tend to not listen to uh opinions. And that's just, I think, something that, I mean, everybody has an opinion. I like to go back to the data, right? The objective data. And so I I I'm using AI for this as well. Um but what I do is like truthfully, I use AI to listen to every account conversation at once, inbound and outbound.
So we're talking about like pre-sale and post-sale. So sales call transcripts, email threads, like I was mentioning before, you know, while the deal is live, support tickets like, Renewal call, uh, renewals, emails, CM notes, CSM notes, and AE notes, all these things. So this is for pre-sale and post-sale. No human can read all of that, as I mentioned before. So it's like, I need to know what deals are slipping for real before the red updates in the stage.
Yeah.
And suddenly it's like, oh, we went from green to red really quickly. Why? The contextual element is so important. So I know like which accounts like just asked, you know, a question that sounds, you know,
Right.
like an expansion potential or like a contraction or like they're slipping. I didn't know which customer said fine in a tone that means we're evaluating your competitors, right? I would not be able to pick up on that as a human being. So that's not a dashboard refresh, man. Like that's what we used to have. It's a signal that hits the right person at the right time in the conversation that happened in real time. And that's so, so, so important. So I talked about how that could apply to marketing and it applies to sales in the same way, right? So it's it's about understanding what conversations are happening right now with all the different folks in the pipeline. And if certain things are being raised or picked up by Sturdy, having those things forwarded to the right person at the right time or right automation to ensure the slipping doesn't occur, that we stay proactive about it. If there's a potential for expansion within that deal or more seats or more usage, great.
If they're not saying that explicitly, now we know that they're thinking it, let's go in and get that deal expanded. So to be clear, humans weren't built to do this alone. This is something that AI does incredibly well. Because by ingesting
Yeah.
every single account conversation and prospective customer conversation, we can understand exactly what they're thinking and maybe what they're inferring that we would've never otherwise seen. And we're not about opinions anymore. I don't have to sit in a meeting every week and go through who thinks what's going to close and when. We literally know that information already, and it's a real-time. You know, dash if we want it to be a dash, and we are getting real-time updates in Slack, in Teams, in email, wherever you want it, right? You know, let's just say for example, like, you know, um, a specific massive account, you know, at HubSpot is coming through the pipe. They've been sitting for a little while. They mentioned some language within one email that goes to the AE only that they've been thinking about other tools. Now they don't even have to mention other tools. You don't have to mention the competitors.
But just the language suggests that they potentially are looking for other, you know, competitors to you and what you're providing. That AE may read that email. That AE may talk about that email and might bring that email to their boss or they might bring it sort of when it comes to projections on whether or not that that HubSpot account's going to close. They may not, right? But to to miss that critical piece of information, that one sentence in thousands and thousands of sentences that have gone back and forth forward an email and calls would be a deal killer. So don't miss that sentence. Never miss a critical risk ever when it comes to those deals in your pipeline. And that is how I'm ensuring that those things get across the finish line proactively.
So it sounds like your guys' tool uses sentiment analysis in from the calls, and then then you're also ingesting all the emails and communications back and forth as well. So then it's kind of pulling all of that together, tying it to a specific customer or a specific opportunity and then being able to- you could do whatever you want to do with that data
Yeah, yeah,
at that
absolutely.
point, put it into a dashboard
Yeah.
or notify somebody specific or update
Yeah,
the pipeline or whatever.
you you mentioned something I just want to double-click on real quick is the the sentiment piece. I think sentiment analysis is baseline at this point, right? Like we should be able to pick up a conversation. Whether or not somebody's upset or happy or stoked and ready to provide a case study because they're the best customer and they love you so much. So that's that's a big piece. And a lot of people say, well, dude, it's cool. I got Gong. Gong provides some sentiment analysis. Um and and Gong is a great tool. I think Gong is a really cool revenue intelligence tool for this purpose. But one thing about like a lot of these tools, these point solutions, right, is they're coming in, they're focusing on one thing,
Yeah.
calls. Nobody does emails. I think that's a huge thing. I think The biggest storage of unstructured conversational data is through email, and people don't even think about that. So email is huge, but it's not about these one or two or three. I mean, then you got Jira, you got all these like things, you know, you can use for like tickets. Nobody is able to take all of those sort of disparate siloed conversational data points and bring them into one place, match it with a record that you're able to then say Alex from this email is the same Alex from that call, same Alex from the Salesforce contact, right? Pull it into one place and then run everything on it. And that is, I think, incredibly powerful stuff. And to do it in a governed way too, so that everything is safe, right? I think that's the big question about AI these
Hmm. Right.
days. But that but to be able to do that and to be able to then have your finger on the pulse, it's like literally like you are, you have this new vision that nobody else really has
Yeah, it's like
and you
it's
know
magic.
it's grounded. It's magic and it's grounded in data all the time. And that's really what I come back to every time. I hate opinions. That's just something that I feel very strongly about because everybody has one. But if you have a data-backed source that says this is what's happening right now, we can get ahead of it. We can get ahead of churn by 2 months. We see this happening because a lot of times by churn, you know, it's already happened. They're already gone. They've made the decision. We see this deal slipping before it's even started slipping. Let's get in there. Let's make sure we create some sort of scenario in which we can get this thing across the finish line for whatever we have to do to get that across the finish line. That kind of intelligence, that level of intelligence is just starting to kind of arise, I think. Um, in certain industries and verticals, but it's it's like a cheat code. It's like, wow, we suddenly know exactly what's happening with every single account and we know how to get them across the finish line.
So, you know, opinions are out the door, I think, on this one.
Yeah, I guess, you know, uh, opinions are like Buttholes, everybody has one and they
Everybody
usually
has
stink.
one. I almost said it. I'm glad you did because it's your show. No, but it's
Yeah,
true. Everybody everybody has one and a lot of times. And I'm not trying to say like people are trying to look out for their own skin, but like if somebody has a HubSpot closing 80%, you know, chance of closing, but then that sentence comes in through an email and they didn't catch it and suddenly now HubSpot's not closing anymore. What happened? I don't
or
know.
there's pressure
Or like-
from their leadership to be adding deals in that probably shouldn't really be there and they feel the pressure to do that because
100%
You know, if they
yes.
don't, they're gonna
100%
get fired. So then they put stuff in that even
yes.
they know isn't actually great, but they gotta have something. So this way,
Yeah, yeah.
everybody can
Have that
actually
source of
be
record,
honest about
man.
it.
100%. Have that honesty. Have that exact 360 view on exactly what's happening so anybody can be held accountable. That's really the truth. That's what we're looking to do, right? So I think honestly that's how we figure out our pipeline. Not just getting folks in the pipeline with cool, unique content that's educational and entertaining. Once you get them in there, which is a great high intent way to get them in the pipeline. Great. We got them there. Now let's measure exactly what's going on in every single conversation, the highest volume of data ever to be produced. And then get these insights proactively to figure out the best means of getting them across the finish line. I mean, it sounds easy because it's becoming easier every day, but I mean, that's exactly how you go do it, right? We're moving into this new era, um, and we should all be closing far more than we've closed in the past.
We should be getting far more people into the pipeline than we've been getting in the past, but it's just harder work,
And probably
right?
wasting less time with people who were never going to close in the first place, which is,
Exactly.
I mean, that's
Exactly. Think about
where
that.
it's at.
Think about how much time you're wasting on those, absolutely, that we're never gonna close ever. I think that's a really good point. I didn't think about that. Yeah, yeah, you're not wasting any time anymore. You're focusing on the ones that make the most sense, that you have the highest propensity or or or opportunity to close, and then you get that money across the finish line. I mean, it's it's the kind of intelligence I think most leaders want, um, but people
Love it.
are still just kind of learning about. So hopefully it catches on soon.
All right. Well, I got the most important question for you, Alex.
Sure, what's up?
So you've got a Beatles poster behind you. It looks like a synth.
Yeah. I got
Who
a couple of Beatles posters.
is, who's, who's, who's your favorite band or or music artist?
Oh, dude, I don't know if you can tell. Also, fun fact. So this whole room is decked out in in vintage Beatles posters. These are all original
Okay,
posters
I could only see
from
the
the
one.
'60s. Yeah. So we got one. These are all from the '60s. This is the Beatles' last concert. So these are all from the '60s, original posters. Uh, my favorite band of all time. And my wife ensures that the only place they're hanging in the entire house is this one room. So they're all-
So they're
I
all there.
got one behind this monitor, I got one over here, I got two over here, I got another one actually over there in the corner. Yeah, another. And the synth is mine. I, I grew up early, early, like, you know, teenager, learning how to play the Beatles. I, I remember the first vinyl I ever put on was my parents' house, um, and it was a Beatles album. And from there on, I was hooked. And so I I I learned how to play all those songs and and now that's my creative outlet. So that's that's my little uh MIDI keyboard that I pull over to this station to hook up to play music here on the keyboard. Um but yeah, no, huge huge fan of music and that's that's
Love it, man.
probably my biggest nerd nerd moment is is my Beatles nerding and and and how how unfortunately how much I know about the Beatles. So
I love
Beatles
it.
trivia
Yeah.
next time we talk.
A bunch of my uh you know most of my friends growing up, their parents were listening to like '80s bands, AC/DC and you know
Hey, I
'70s,
love that too, man.
late '70s, '80s. Uh, but my dad definitely listened to Beatles and like jazz and stuff like
For
that.
sure.
So I grew up
Let's
way more
go.
listening to the Beatles and stuff growing
Let's
up.
go. I love that. I love that. No, it's it's true.
Uh, but.
you can always find a good person, a good conversation over a Beatles, you know, any Beatles track or any Beatles album. That's usually how I've made most of my friends. So whether or
Awesome.
not that's a good way to make friends, it seems to work. So yeah, man, no, I'm glad you picked up on it. That's great. Nobody's
Love
ever
it.
asked about the posters.
So Alex, how can people
Sure.
find out more about you and uh
Yeah,
and about Sturdy?
no, it's just sturdy.ai. It's pretty, pretty straightforward URL, thankfully. Um, but yeah, I'm just Alex dash uh and dash Atkins um on LinkedIn. So yeah, um. We're we're working really hard to continue putting out fun, unique content that's hard to build but hopefully stands out. I mean, that's the one thing I would always recommend. It's so easy to generate and post content these days. Do everything you can to not do that. You know,
Yeah.
video content like this is one of my favorites, by the way. Cool, snackable, um, educational stories. I really love how that is applied to what you're doing. So I think that kind of content's a really cool way to do it. But yeah, uh, you can find me on LinkedIn, find me, uh, you know, on uh sturdy.ai. So that's probably the best way
Love
to
it.
do it. Yeah.
Awesome.
Thanks.
Well, with that, thanks everybody for tuning in. Don't forget to subscribe wherever you listen to your favorite podcasts, and we'll see you on the next episode.
Clips from this episode
You've got a content calendar to fill and a pipeline that isn't moving, and the quick route is right there: generate a post, publish it, move on. Your buyers are scrolling past the same posts you are. Then the weekly forecast call starts, a rep says a deal is definitely going to close, and you have nothing to check it against. I talked with Alex Atkins on this episode of Pipeline. Alex took the hard route on content, a report built on first-party data, and stopped taking forecast opinions at face value.
How do you create content from first-party data that generates pipeline?
Alex Atkins, VP of Growth and Operations at Sturdy, built a report from the company's own conversation data instead of taking the quick route. The report, Churn Is Not a Customer Success Problem, used tens of millions of customer conversations to show where churn originates. Alex calls it the single highest performing asset Sturdy has ever produced.
From persona to published
- 1Pick a top persona in your ICP
- 2Let the data tell the story
- 3Title it with the claim they suspect
- 4Publish on LinkedIn under a person's name
Alex started from the audience. Customer success and CX are among Sturdy's top personas, so Alex went digging into the first-party data Sturdy already had on that group. The data came from business conversations between Sturdy's customers and their own customers:
- Email and chat
- Support tickets
- Zoom
- Slack
Alex's question was what causes churn, based on conversational data "that nobody else has." The answer: product gaps, overpromising in the marketing and sales cycle, and broken onboarding. CS was left with the cleanup.
"First-party data. What is a cool story we can tell with first-party data?"
Alex Atkins, VP of Growth and Operations at Sturdy
"And CS, you know, was cleaning up after all these decisions that were already made months earlier by other departments that they had absolutely no control over."
Alex Atkins, VP of Growth and Operations at Sturdy
Distribution cost nothing. Alex says it went out primarily over LinkedIn, not through paid, and brought in high-quality opportunities with high intent.
Why did a churn report get read by both CS leaders and CROs?
Alex Atkins says the title of Sturdy's report did most of the work. Every CS leader read Churn Is Not a Customer Success Problem to defend themselves, and every CRO read it because it said out loud what they already suspected: churn is a revenue problem that gets handled by the wrong team.
The data backed the title. In Alex's account, most churn comes from sales overpromising, billing issues and product confusion, none of which CS controls.
"that means almost 75% of all churn is completely out of the realm of of control of CS and CX, right?"
Alex Atkins, VP of Growth and Operations at Sturdy
The report also came from a person, not a logo. Alex says people follow people, and the report brought buyers to Alex first and to Sturdy second.
"I published this thing, I put it out on LinkedIn, people came to me. They didn't really come to Sturdy, but they got interested in Sturdy after they learned from, you know, from me and what I provided..."
Alex Atkins, VP of Growth and Operations at Sturdy
Readers then asked how Sturdy knew all this, and the answer was the product Sturdy built. Alex says some of it was luck. The data happened to make a good story, and Alex says that if it hadn't, the report wouldn't have run.
How do you find the right pain point for first-party data content?
Alex Atkins, who has a master's in marketing analytics from Bentley, says to let the data tell you the story. Alex went in expecting CS to be responsible for the vast majority of negative sentiment and churn, and the conversation data showed otherwise. Alex says that result "created an amazing story."
At Sturdy, finding that story means querying or prompting Sturdy itself. Alex says the product has trained over the last 5 or 6 years on millions of business conversations. It reads phrases, not keywords like "sad, mad, bad." A customer writing that you were late to the call and it frustrated their team reads as response lag, frustration, negative sentiment and probably urgency. Alex says Sturdy picks up "dozens and dozens of signals," including product confusion, and sorted churn into 6 different groupings.
Joseph Lewin said a prospect often thinks their pain is coming from something totally different. His view is that your job in GTM, whether you're in marketing or sales, is to take that pain and reframe the problem to the thing you solve.
"what ends up happening a lot of times is that your, your prospect doesn't actually understand what's causing that pain."
Joseph Lewin, Host of Pipeline
Alex says first-party data makes that reframe land, because "It's not just like my opinion." It's what the conversations show.
What should you do when first-party data content doesn't land?
Alex Atkins says to iterate: keep the same data points and look at them from a different angle. Alex's first piece, the one that eventually evolved into the churn report, was called something like "AI in Business," and people "were just not resonating." Alex says that didn't mean changing everything, only figuring out how to look at the data from a different angle and put it in the report in a way that would resonate.
Before you scrap a piece
- Is it not resonating?YesSame data, a different angle
- Wrong people reading it?YesAim at your ICP
Alex calls the audience another critical piece: the ICP, "the right people, not just people."
Joseph Lewin remembers the team behind Digital Marketer saying 1 out of 10 of their campaigns would be truly out of the park, and they were a huge marketing team with tons of data. Alex's response: "those are some of the best we have."
"I tried a lot of content before I found one that worked."
Alex Atkins, VP of Growth and Operations at Sturdy
What to Try on Monday
- Pick a top persona in your ICP, the way Alex started with CS and CX leaders, and dig into the first-party data you already have on that group.
- Go into the data to let it tell the story, not to prove what you already believe.
- Title the piece with the claim that persona has been suspecting but hasn't heard anyone say out loud.
- Publish it on LinkedIn under a person's name, because Alex says people follow people more than companies.
- If it doesn't land, keep the data points and try a different angle before you scrap it.
How do you track deals in your pipeline without relying on rep opinions?
Alex Atkins uses AI to listen to every account conversation at once, inbound and outbound. That replaces the weekly calls where AEs say a deal is "probably like 90% close." Alex wants to know which deals are slipping for real before the stage turns red, from what buyers actually said.
Don't miss the one sentence
- The weekly forecast callAn AE says it's probably 90% closeEvery account conversationA buyer's own words on the deal
- The weekly forecast callSlips show when the stage turns redEvery account conversationSlips show before the stage turns red
- The weekly forecast callOpinions about what's going to closeEvery account conversationWhat was said in every conversation
“That one sentence in thousands and thousands of sentences that have gone back and forward in email and calls would be a deal killer.”Alex on the one sentence
Alex's coverage runs pre-sale and post-sale: sales call transcripts, email threads, support tickets, renewal calls, renewal emails, CSM notes and AE notes. No human can read all of that. When a deal shows risk or expansion potential, the signal goes to the right person in Slack, Teams or email.
"I don't have to sit in a meeting every week and go through who thinks what's going to close and when."
Alex Atkins, VP of Growth and Operations at Sturdy
The risk Alex is guarding against is small and easy to miss: a single line in an email that hints the buyer is looking at other tools.
"that one sentence in thousands and thousands of sentences that have gone back and forth forward an email and calls would be a deal killer. So don't miss that sentence."
Alex Atkins, VP of Growth and Operations at Sturdy
Joseph Lewin said reps feel pressure from leadership to add deals that probably shouldn't be in the pipeline, and that this way everybody can actually be honest about it. He added that it means wasting less time with people who were never going to close in the first place.
If you build a report from your own conversations this quarter, watch for the second audience. Alex's churn report was written for CS leaders and pulled in CROs too. If your first version gets the "AI in Business" reaction, look at the same data from a different angle before you go hunting for new data.
The whole episode on one page
Animated version for LinkedIn (GIF) · Want it printable? Get the field sheet
Frequently asked questions
What is first-party data content?
Alex Atkins describes it as content built on data only your company has. For Sturdy, that meant the business conversations its product ingests, across email, chat, support tickets, call transcripts and Slack. Alex says buyers were happy to finally see someone provide first-party data on a problem they'd been talking about all along. That made Sturdy a source they could trust, and it brought in high-intent opportunities.
Is AI-generated content still worth posting on LinkedIn?
Alex Atkins says generated content is creating noise. On LinkedIn these days, Alex sees probably 10 to 15 posts that were clearly generated within 30 seconds and probably not read over by a human. Alex's advice: "It's so easy to generate and post content these days. Do everything you can to not do that." Video content and snackable, educational stories are among Alex's favorites.
Why do customers churn if it isn't customer success's fault?
Alex Atkins says Sturdy's data traced churn to product gaps, overpromising in the marketing and sales cycle, broken onboarding, billing issues and product confusion. By Alex's account, almost 75% of churn sits outside CS and CX control. CS still owns part of it, Alex adds, through response lag that leads to frustration, negative sentiment and eventually churn.
How is reading every account conversation different from call recording tools like Gong?
Alex Atkins calls Gong a great tool and says sentiment analysis is baseline now. The difference Alex draws is coverage. Point solutions focus on calls, while Alex sees email as the biggest store of unstructured conversational data. Alex describes pulling every source into one place and matching the Alex from an email with the Alex from a call and from the Salesforce contact.
This article is based on Episode 9 of Pipeline featuring Alex Atkins, VP of Growth and Operations at Sturdy.
Show notes & linksGuest links, resources and the full episode description
💡 Summary
A report called Churn Is Not a Customer Success Problem became the single highest performing asset Sturdy has ever produced.
On this episode of Pipeline, Joseph Lewin talks with Alex Atkins, VP of Growth and Operations at Sturdy, about how Sturdy brings people into the pipeline and how it tracks deals once they're there. Alex built a quarterly report from Sturdy's own first-party data, tens of millions of customer conversations. It showed that churn starts with product gaps, overpromising in the sales cycle and broken onboarding, and that CS was cleaning up after decisions other departments made months earlier. Every CS leader read it to defend themselves, and every CRO read it because it said out loud what they already suspected. For deals, Alex doesn't rely on what AEs think will close. Alex uses AI to listen to every account conversation and catch the sentence that says a deal is slipping.
Sponsored by Podcast 2 Pipeline
✅ What you can use by Monday
- Pick a top persona in your ICP and look in your own first-party data for a story nobody else can tell them.
- Let the data tell the story, even when it shows you were wrong going in, as Alex was about CS causing most churn.
- If a piece doesn't land, keep the data and change the angle: Alex's first version was called "AI in Business," and people weren't responding to it.
- Stop asking AEs who they think will close, and read what accounts are saying across calls, email threads, support tickets and CSM notes.
- Include email when you track deals, because Alex calls it the biggest store of unstructured conversational data.
🕐 Chapters
Intro: top of funnel content and tracking deals
The churn report built on first-party data
Why hard content wins and people follow people
Finding the pain and reframing it
What really causes churn: 17%, 25% and 27%
Most content misses: iterate on the angle
Using AI to read every account conversation
Why call-only tools miss the email sentence
🎙️ About Pipeline
Pipeline is short conversations with GTM peers who are actually filling pipeline right now, so you leave with something you can use by Monday. New episodes weekly on Apple Podcasts, Spotify, and wherever you listen. Produced by Podcast 2 Pipeline.
🏷️ Tags
first-party data, content marketing, customer churn, customer success, pipeline generation, deal forecasting, revenue intelligence, AI for sales, LinkedIn content, Sturdy, Gong
Enjoying Pipeline?
Could you help us out? Following the show and leaving a five-star rating helps more people find it. It takes a minute, and we really appreciate the support.
On Spotify, tap Follow on the show page, then the star rating under the description.



