The Product Manager

Finding Your Unfair Advantage in Product’s Wild West Era (with Margaret-Ann Seger, Head of Product at Statsig)

• Hannah Clark - The Product Manager

Share episode

Product-market fit isn’t enough anymore. In an age where AI is making it easier for anyone to ship a half-decent product, the real battleground is distribution. Margaret-Ann Seger (Head of Product at Statsig) joins Hannah Clark to talk about how AI is reshaping go-to-market strategy, why speed and feedback loops are becoming product differentiators, and how her team at Statsig is baking collaboration and empathy into every part of their product dev cycle.

From lo-fi GTM experiments to the power of doing support as a PM, this episode is packed with practical ways to ship smarter—and more human—products, faster.

Resources from this episode:

Hannah Clark:

It's official. Vibe coding has entered the vernacular of the gen-pop, and we are entering a critical moment in which AI, digital products, and culture are coming together in one giant evolutionary leap. Democratizing access to digital product development, or in simpler terms, 'building on vibes', marks a permanent leveling of the playing field in which literally anyone can become a founder. And while eliminating barriers to entry will mean that more great ideas will have a pathway into the market, it also means that, well, more great ideas will have a pathway into the market. Like, a lot more. And at this scale, the implications of that are so much bigger than just having way more competition. It means that if we hope to be real contenders in this market, we need to hold our GTM and growth strategies to a much higher standard. And lucky for you, my guest today is Margaret-Ann Seger, Head of Product at Statsig. As you'll hear shortly, Margaret-Ann, or MA to those who know her, has an incredible sense for what breaks through the noise and gets growth engines running. But the absolute goal you'll hear in this episode are the tactics that she and the folks at  Statsig use to enable collaboration between their users and the product team. You are absolutely going to wanna take notes. Let's jump in. Oh, by the way, we hold conversations like this every week, so if this sounds interesting to you, why not subscribe? Okay, now let's jump in. Welcome back to The Product Manager podcast. Margaret-Ann, thank you so much for making some fire to chat with me today.

Margaret-Ann Seger:

Thanks for having me on the show, Hannah.

Hannah Clark:

So can you first tell us a little bit about your background and how you got to where you are today at  Statsig?

Margaret-Ann Seger:

Sure. So I lead product and design at  Statsig, which is the modern product intelligence platform. So from traditional AB testing through to, offline model testing, increasingly teams are building with models, smart feature flagging, product analytics. We help teams incorporate data at every part of their product building process in this age of AI. So pretty comprehensive platform. For me. It resonates a lot because I came from bigger tech companies. My started my career, was at Big Tech, started at Facebook, and then spent a little over six years at Uber. Both companies were made hypergrowth when I joined, but I had the luxury of having this suite of super awesome internal tools to use. And it wasn't until I left in 2020 and popped out in the real world that I realized not every company has access to those tools. And so what we're building at Sadig resonates because we're democratizing access to those same set of tools for every company.

Hannah Clark:

That's awesome. Today we're gonna be talking a little bit more on the growth side of things, how to nail growth and GTM strategy in speaking of democratizing in the age of AI when so many parts of the product development and launch process are democratized. We have to really figure out really smart, intuitive ways to stand out. So let's start out by referencing a line that you shared with me in a previous conversation, which I really liked, which is first time founders care about product, repeat founders care about distribution. I think that's a very succinct way to frame the conversation. So when you think about that, what's the most important mindset shift you think that product leaders need to make when we're transitioning from a product first to distribution first thinking.

Margaret-Ann Seger:

Good question. So this actually, my husband told me this line. He was telling it to me in the process of lining out his startup. So it was very timely to have that conversation. But I think, the spirit of it is that at the end of the day, you can build the most amazing product in the world, but if no one knows about it or you have no way to get it in front of the right people, you're not gonna be successful. And I think this is especially true in the age of AI, right? Because there's just so many more products, right? The bar for creating a product has gone down. Now everyone can create their own app, their own website, their own X, Y, Z. And so there's gonna be so much more noise. How do you stand out? How do you actually get distribution? And so I think, it's a couple examples that I often think of when I think of this done right, are things like, I don't know if you followed, this is not like a tech product example, but I. Haley Bieber's makeup line was recently acquired Road. And it was acquired for over a billion dollars and super successful outcome. And that was all built on just her brand. She's probably good at makeup, but I'm sure there are a million and a half people who know more about the nuances of creating makeup and all that stuff. But she has a phenomenal brand. She had distribution. You can light that up like this. I think back to my time at Facebook, we had amazing distribution and so any feature you launched could go from zero to hundreds of millions of users overnight because you just. Had a giant platform with already a huge adoption. And I think like even you're seeing it in the AI companies, right? Like cursor piggybacked on top of VS. Code. There's all these like just existing behavior and existing distribution channels that people can then build something incremental on top of and overnight, boom, light it up. I think that is going to differentiate the companies that are really successful is can they figure out a sustainable and creative distribution channel? And it's not just gonna be about do you have a good product?'cause that'll almost be table stakes in the age of AI.

Hannah Clark:

Yeah, and that's exactly, I think the concern that's on a lot of folks' mind is, you might have a great idea, but the functionality of a product that you're creating isn't necessarily, it's not without, there are many paths to that same outcome. So if we're thinking about strategies for differentiate great products from the crowd, knowing that anyone can create the same app that you can create, what are some of the ideas that you would suggest people start with when they're trying to. Make their place in the market.

Margaret-Ann Seger:

Yeah. I think an important one that doesn't get talked about as much is actually feedback loops. You might have an AI assistant writing your code. You might have, just assistance at every step of the product build and launch process. But then how are you getting feedback once it's actually out in the world? And both qualitative and quantitative feedback, right? You need to know quickly is what you're building actually working and just. Get those kind of feedback loops humming. So you're constantly getting signal and iterating accordingly. And then I think to that point, the differentiator for many people, and actually at stats, like we see this as a differentiator for us, is speed, right? So once you get those feedback loops up and running, are you able to incorporate that feedback and iterate faster than the next person, right? And so if your users are telling you something, can you capitalize on that immediately? And so I think that feedback loop plus speed are really gonna differentiate the winners. And I think one thing that we've seen is we actually will even put something out there that isn't fully built that's really rough around the edges, just to get signal so that then we can inform what we actually built and shortcut some of the like core product building process by doing that.

Hannah Clark:

So on that note, so really what this is about is getting to know the users and that like I think it's always been critical and now it's just like paramount. So what's one tactical approach that product leaders can use to really understand their ICP right now, given this democratizing landscape?

Margaret-Ann Seger:

Yeah, so this is gonna be controversial, but actually do support. I think PM's doing support and working with customers when they're getting problems is actually really helpful if you want to understand your user, which I think to your point is gonna be more important than ever. We take this to the extremist at Statsig and it blew my mind when I first joined. But the, we don't have a support team, like the entire team. The entire company is responsible for support and we've actually spun up a whole suite of complex tooling with Slack feedback groups that auto triage via a bot that like go into the on-call of the days queue and every team member becomes on call of the day on a rotating basis. I'm head of product, but I also do support. And so I hear when things are going wrong. I see when people are getting stuck on a certain flow and I think that actually lets me just stay in constant contact with the customer and their pain points and just like channel that in every conversation automatically. And I think this is controversial because there's so many companies, Sierra, and even just all these AI customer support bot companies that are basically their premise is you shouldn't be having to do these questions. We're gonna offload this to. A bot or the AI, and I just think that it helps you really stay in touch. And I also think that your customers appreciate it. It almost becomes a point of differentiation in this new world.

Hannah Clark:

Yeah, and this actually blowing my mind. I think it's actually a brilliant strategy because whether you're using a bot to answer customer support questions or you have a support team. If you're keeping that separate from the development process of the folks who are really closer to the product, you're telephoning that information or you're beholden to making sure that you're checking in with the right people. But I also think it's really brilliant because otherwise, what are the other ways that you're getting feedback? You're soliciting it from people who either are really eager to support a product because they love it already, or the people who really hate it. So being in the support area, it's not really where people expect it. You're going to take into account their feedback. They're just trying to get from point A to point B and that middle ground I think is otherwise very difficult to capture.

Margaret-Ann Seger:

Yeah, it's really cool and I think it, customers in their real world environment. You meet them here in the moment, you see the emotion. And it's cool too to see like our engineers building this empathy.'cause I think it just scales really nicely, right? It's not just the PMs who are doing this, it's the engineers. They get excited about, Hey, there's an opportunity to improve this flow. I'm just gonna go like ship a quick fix. And so you almost can achieve more by just doing this bottoms up empathy building.

Hannah Clark:

Yeah, absolutely. And I can really see the value too in being able to really understand the use cases that people have for the product in the moment and understand, what are they trying to achieve? Should there be an easier way that they shouldn't have to contact support? This is, my brain is buzzing, I love this idea. I'm really glad that it came up. Okay we'll move on. So when we think about the speed of build cycles, that was the other thing that you mentioned about differentiating in terms of being able to incorporate your feedback into your product roadmap very quickly. So you suggested, scrappy things, build them quickly, get them out there, get some feedback. So when you think about that, what's a framework that you'd use to decide what to build and test first? If there's just many options, many ways that you can take that.

Margaret-Ann Seger:

I alluded to this a little bit earlier, but we've started filming prototype videos. So doing the prototype and then actually filming a video as if it's a real product and saying it's open for beta request if you want beta access, and we don't build it. And so the customer reaches out, if they're interested or someone sees it on LinkedIn and says, oh, this is cool. We should try, try this. They'll ping us and if we get enough demand and excitement, we'll go build it. And so it's a very like low cost way to gut check ideas without having to go and build a full end-to-end product and all the edge cases and just like nuance that you have to think about there. We've tried this with a few things and it's worked out really well, so I think we're just gonna keep doing this and basically letting the market tell us what we should or shouldn't be building.

Hannah Clark:

This is such a peek behind the curtain. I feel like this is like a Wizard of Oz reveal moment.

Margaret-Ann Seger:

Maybe I shouldn't be saying this.

Hannah Clark:

No, I think it's brilliant because there's many things that you could do, but if you don't have the excitement to your kind of point about the distribution. Why put all the effort into the build cycle if you just don't have that momentum.

Margaret-Ann Seger:

That's the beauty of this is we get built in distribution on day one 'cause we know it's there. We've already validated.

Hannah Clark:

Fantastic. Okay, let's talk about the use of AI more in terms of getting through your GTM strategy.'cause I think that really the crux of this is how are you gonna launch a product and be successful when there's so many products just like yours that are launching every moment and a lot of people are thinking, I'll just ChatGPT, it. GTM strategy. I can just prompt that so you're laughing and I Yeah. To say no, I'm laughing because you're right, you're a hundred percent right. So if you think about, what is the missing piece? What are people missing by taking that strategy that we really need to think about that can't be automated, that we really need to do manually?

Margaret-Ann Seger:

So look, you can definitely use, I don't wanna rag on Chad g Bt, you can definitely use Chad g Bt for a lot of the concrete outputs. Knowing what those outputs should be and how to frame them is like an art in and of itself. And there's the whole meme about PMs are basically gonna become prompt engineers. Like I do think you can use these tools for your GTM strategy, but you need to be prompt engineering and really guiding that process. The reason there needs to be human doing that is it's about empathy. At the end of the day, your GTM strategy's only going to work. If you have deep empathy for your customer, you understand the pain they're going through, you're able to speak their language and you're able to position something that you know is gonna meet them where they're at. And I think like all sorts of inputs are needed there, right? You need to have quantitative inputs. So understanding what your users are or aren't doing, they might tell you something, they might do something else. Qualitative inputs, you should watch their sessions. You should be knowing not just what they say, but like actually what are they doing behind the scenes and where are they getting hung up. And I think that AI can actually help in both of these processes, right? There's a ton of startups right now exploring AI driven session replay synthesis, which I think is super clever, right?'cause it's like you record thousands, hundreds of thousands potentially of sessions. You can't watch all of those. You would like someone or a bot to go through and glean the key insights there and push those to you. But it still takes a human to take those insights. Know how humans behave and who their ICP is and what motivates them. And then put those two and two together to then figure out the framing. Once you have that framing, I think you can use ChatGPT to say, Hey, here's the raw inputs, here's what I'm thinking. Here's the customer. Can you help me position this a little bit better? But there's a lot of pre-work needed there.

Hannah Clark:

Let's get into the nuts and bolts of a strong GTM strategy.'cause I think that this is another area. If you've got an enormous amount more competition than you did before, everything has to be stronger. And in particular, your GTM strategy now versus five years ago has to be way more robust. So if we were to do a, like an anatomy lesson on the perfect GTM strategy now versus what would pass for a great strategy in 2020, what are the hallmarks would you say of a really good today facing strategy?

Margaret-Ann Seger:

This is a tough question because I don't know. It's tough to define like a great strategy today. It is very easy to say what we used to do that it no longer scales. So I'll start there. I think and we've seen this evolution even, I joined Statsig early 2022 and in the last three and change years, we've seen this happen. A couple things have completely shifted. One is, SEO is very different now as models and people just bypassing Google and going direct to the. ChatGPT or Claude or Gemini, it's, you're not gonna get much bang for your buck on SEO once critical mass moves there. The other thing is like a much higher bar for differentiation. So it used to be, for example, high quality content was enough of a differentiator, right? If you put out a great blog or if you had a great newsletter.

Hannah Clark:

Or podcast.

Margaret-Ann Seger:

Yeah. Yeah. Yeah, all these things were like enough, and now I think it's almost like table stakes in a sense. Especially, I think written content is becoming increasingly fraught because it's just so easy to generate, like AI slop essentially. And so how are you writing things that cut through that noise, especially when people are generating like hundreds and hundreds of AI blogs to just rank better, right? In search. The other thing is like actually, and this is shifting from just. What's no longer good into what I think the gold standard is. I actually think, and I've touched, I feel like I've hit on this point in a number of different ways, but I actually think like human touch is going to be the gold standard. Actually having a human in the loop is almost gonna be revolutionary and people are gonna naturally gravitate to the brands that feel more human. It's funny, someone was telling us, like A friend was telling us they thought Theas brand was really like different than your typical AI company. And I was like, why? Quantify that for me, and they were like, oh, there's just so many people you guys like have your employees out there. You have videos with your PMs launching things. You're like, have faces of individuals attached to your blog posts. You're very people first. And that's actually edgy in today's world, right? You look at the versel, the linears, like these super like. Clean, minimalist, almost sterile brands, and that's the gold standard. And it almost feels like messy and unconventional to have people have your employees out there actually. But we've found that a lot of people love it and I like it, right? It's like we have our dogs on our website, we have, the people behind the products, not just filming videos, but also answering your support questions. And I think that has actually been a differentiator for us. As everything, people joke about, what is it? The the gray floors, white wood aesthetic in houses and everything's just gone to that. I think this will be similar in tech where there's an aesthetic that everyone converges on, and if you actually are different than that or memorable in some way, I. That's going to be an edge for you.

Hannah Clark:

Yeah. I think this is really consistent with something that I've been kicking around in conversations a lot lately about how it just seems like the case is with everything from tech to fashion trends to just about everything, there's always a pendulum swing where there's a certain saturation point where people start to crave something in the other direction, and it's almost proportionate to where the saturation and how saturated we're looking at like a specific pendulum swing. So I'm seeing exactly what you're saying, it used to be that everyone really gravitated towards those really clean, minimalist, yeah. Design and websites. They're, yeah, and they're beautiful and very simplistic, but they're also the easiest to generate now. And so I think it's the transition from Instagram to TikTok where Instagram really popularized this very polished, very shiny content, and then people started to really crave this lo-fi aesthetic where it was very hands-on people first. And I think we're seeing a similar kind of thing where people are like, no, I really want evidence that someone, people are behind this. Yeah, I think there's something there. So if we're talking about growth teams using tools like at Statsig, what would you say is the biggest GTM mistake that you see teams make when they have access to a lot of different technology for measuring success? There's a lot of data to parse through and some easy bad correlations to make. How do we kinda make sense of that?

Margaret-Ann Seger:

Yeah, it's interesting. Increasingly teams are adopting tools to be more data driven and to log data and to, just. Have that input. The problem is a lot of the times the inputs don't line up. So if you have four different tools and they're all logging similar actions that a user's taking, you're then having to parse through four different data sets, four different data sources. They rarely agree. It gets super messy. Then people get frustrated because the data doesn't line up. You don't know what to trust. And so it ironically, like the more you add on these tools and try to be data driven, the harder it becomes to actually be properly data driven. A big trend in the industry to meet this problem has been consolidation and like these platforms being built that start to combine multiple tools and like we're one of them, right? So I'm not saying that this is the Statsig unique thing, but having one set of SDKs that are logging things. One source of truth data or a source of truth data in your warehouse, right? Like increasingly companies have a data warehouse that is their source of truth. That's great. Like all these tools should be building on top of that. They shouldn't be trying to like create new data sources. So you're seeing that across the board in our industry at least. And I think that actually helps a lot once teams can unify on this stack on a source of truth data set.'cause they start to trust data again and use it. But it's almost like a slow rebuilding process.'cause people have lost trust in data over time. So I think that's a big one. The other thing is, you'd be surprised how many folks say, yeah, we were data-driven. And you're like, okay, cool. How'd that feature perform? And they're like. I don't know, and that's because it's actually hard to launch new things as experiments. It becomes this big thing. We have to set up the experiment, what's our design doc? Does the DS team agree with how we're setting this up? Then we launch it and we do a big summary process. The tools that make this process just really lightweight, you're just launching a feature flag, you're training on and off, and it automatically creates a little ab test of the people who have the feature and the people who don't have the new feature. You can quickly gut check that this isn't taking your business metrics or your latency and infra metrics or spiking your cost metrics. If you can just quickly gut check that and go, and you just lower that barrier of entry for AV testing. Really lightweight AB testing. I think that's powerful. Is it going to be the proper end-to-end process and have all these kind of more advanced stats methodologies? Maybe not, but I think it's a really good entry for teams to just start building that muscle. So that's one thing we've been working on is just making that easier.

Hannah Clark:

Cool. So speaking of building that muscle, so a lot of what we're talking about here is like the combination of the mindset of development as well as the mindset of like, how is this, how are we gonna distribute this? And this is something I think everybody in the product team now has to internalize as we really have to have distribution and development kind of in mind at the same time. So how do we get confident in building that kind of a skillset and a mindset? Is there, I don't know if at Statsig, you guys have that kind of a mindset throughout your team now, or is there something that you've encouraged for folks who aren't necessarily directly involved with the product marketing process, but like still need to be thinking about how this is going to be marketed?

Margaret-Ann Seger:

It's tough to spread that DNA. One thing that I think is really cool, and this is more a like who you hire, DNA, yeah. But we try to hire really like product minded engineers who are passionate about the end user, who are passionate about the kind of performance of the product, and they wanna be involved with the marketing. They wanna be, helping, understanding who's actually adopting this product. Is it who we'd expect? Are they using it in the way we'd expect? Do the metrics look good? And so that, that has been cool to harness. Concrete example of that is our infra team, they've been working on a set of tooling that's similar to Datadog in a sense, because they wanna dog food our product, and they wanna use our product more. And I was like, that's great, that's awesome. And so they've gone really far down this rabbit hole, and now at this point they're debugging cev on Sig and they're doing the end-to-end, like monitoring and alerting and infra health day-to-day on Statsig. They came to me and they're like, Hey, we're using the product for this. Could we market to customers? Could we just sell this? Could this be a skew? I was like, yeah, let's do it. And obviously, it's rough around the edges because we're our main customer right now. But that's a really cool way to just come at the problem. You have a pain point, you solve it for yourself. You realize other people might benefit from this and you say, Hey, let's go actually market this. Now the engineering lead on this area is super involved with like myself, our PMM, our marketing team in kind of the positioning for the product.

Hannah Clark:

Yeah. Yeah. That kind of, I think, puts really a clear case on how you launch and how you think about distribution is so nuanced and specific to the story of the product where you're finding value in it. I would love if we could on another story.'cause I'm liking this idea, this kind of train of thought of how important it's going to be for storytelling and like the human aspect to be front and center in how we distribute our products in the future. Have you seen a really good example of that recently? Of a product that has really well leveraged their team? Who would you shine the spotlight on as a really good example?

Margaret-Ann Seger:

Yeah, so this is not recent. This was I think 2018. My gold standard that I always come back to when I think of just best in class GTM was actually at Uber. We did in 2018, there was like a six month period where they did driver forward. So for context, I think, Uber grew super fast. Drivers were a huge part of that, but they often felt, I think like a very unappreciated part of that. There was a lot of storytelling on like the rider use cases that were unlocked. How this was helping people get home safely from the bar or helping mothers get, their kids to soccer practice. But there was less storytelling about the drivers behind the wheel who were earning income in new ways, who were sending their kids to college, who were doing all these really cool things because they had this opportunity. And so there was a whole backlog of driver requests and features and quality of life things, and just like earnings visibility, like all these really core workflow things that drivers have been asking about for years. We, as a business felt like it was time to shine the spotlight on drivers. And so we said, why not combine these two things? Driver forward was basically this concept where every month for six months, the marketing and product teams combined did like a moment and the month had a theme and it would be one net new, feature release or kind of improvement to the driver app. Paired with a ton of storytelling and like even driver sessions where they would invite drivers to the launch, there'd be a launch party, it would launch in a different city at a different driver onboarding center. And it was super driver centric and it really worked like, it was cool. It was a big pivot in the trust dynamic between drivers and Uber. And it was a total masterclass in having product and marketing teams basically oriented entire product roadmap around these GTM moments. And so the woman who drove a lot of this was. Laura Jones, who's actually the CMO at Instacart, and she's just phenomenal. But I think that particular example is for me, the gold standard of how to take your user, take their pain, put it at the center of what you're gonna do, build a whole roadmap around it, and align marketing and product super tightly.

Hannah Clark:

Oh, okay. I love that. That's such a great example. And I'm very succinct way to frame like what that mindset's supposed to look like in action.

Margaret-Ann Seger:

It was really cool to see.

Hannah Clark:

Yeah. Thank you so much for joining us today. This was a great conversation. I think we really hit so many good points in such a like good tight amount of time. If people wanna continue to follow your work, where can they find you online?

Margaret-Ann Seger:

Yeah, so LinkedIn, I go by MA, but my full name is Margaret-Ann Seger and also the staffing blog is pretty cool. We talk about a lot of cool things and like I said, it's very people-centric, if you wanna get to know the team better, statsig.com/blog.

Hannah Clark:

Thanks so much.

Margaret-Ann Seger:

Thank you.

Hannah Clark:

Thanks for listening in. For more great insights, how-to guides and tool reviews, subscribe to our newsletter at theproductmanager.com/subscribe. You can hear more conversations like this by subscribing to The Product Manager wherever you get your podcasts.