Lauren Stein, chief of staff, on orchestrating GrowthLoop’s Compound Marketing revolution

The origin story really highlights the core problem we solve. Our two founders were marketers within Google. One worked on Android, the other on Google Ads – and Google is now a partner, which brings it full circle. The founder on the ads side had a clear quarterly goal: build 30 different customer journeys for various campaigns. Simple enough, right? But in practice, he could only manage to build one journey that entire quarter.

Why? The bottleneck was data access and the operational friction involved. Getting the necessary audience list required constant back-and-forth with the data team, writing SQL queries, building the audience segment, creating marketing briefs – it was incredibly slow and cumbersome. Then, getting results back to analyze performance was a whole separate challenge. He experienced firsthand how difficult it was to access and activate customer data for marketing, even within a technologically advanced company like Google. They realized this wasn’t just a Google problem; it was an industry problem. Marketers needed a way to work directly with their data without these dependencies and delays.

GrowthLoop’s founders saw that existing solutions weren’t solving the root cause. Customer Data Platforms (CDPs) existed, as did big marketing clouds from Salesforce and Adobe, but they operated by creating another data silo. You had to pipe your data out of your central warehouse into the CDP, where it was stored and managed separately. This wasn’t ideal, especially for large, security-conscious enterprises that had already invested heavily in consolidating their data into secure cloud warehouses like BigQuery, Snowflake, or Databricks. These companies, particularly in sectors like finance or healthcare, are very hesitant to duplicate sensitive customer data and send it to yet another third-party platform.

The core insight was: why move the data? Why not build a platform that operates directly where the data already lives? That’s the cloud-native approach. GrowthLoop sits on top of the customer’s data warehouse. We don’t store their customer data at rest; we provide an interface and an engine to query, segment, orchestrate journeys, and activate that data directly from their own secure environment out to their marketing and sales channels. This makes it inherently more secure, eliminates data duplication costs, avoids vendor lock-in around data storage, and allows marketers to work with their freshest, most complete data source. We started with BigQuery support and then built connectors for Snowflake, Databricks, Redshift, SQL Server – essentially meeting customers wherever their data foundation is built.

Bootstrapping was fundamental to our early DNA. It forced discipline, efficiency, and a relentless focus on building a product that delivered real value profitably. It allowed us to build on our own terms, guided solely by customer needs. I was personally attracted to that smart, efficient approach when I joined. However, bootstrapping inevitably means slower growth compared to venture-backed competitors.

As we gained traction and saw the market validating our cloud-native, composable approach, we knew we had an opportunity to scale much faster. We decided to seek external capital, but we were very selective. We wanted a partner who shared our long-term vision and ethos, who wouldn’t just push for growth at all costs but would trust us to run the business effectively.  

TJC was the perfect fit. Their philosophy isn’t about dictating operations; it’s about backing strong teams with the capital to execute their vision bigger, better, and faster. They trust our expertise. Furthermore, they were actively looking to invest more in AI, specifically agentic AI, and saw GrowthLoop as their strategic entry point into that space. So, it became a strong partnership: they provide capital and counsel to build GrowthLoop into an enduring business. This creates optionality to increase velocity to achieving our product vision. Which, in turn, helps accelerate our growth both organically and inorganically, while staying true to the principles that made us successful in the first place.

The Compound Marketing Engine really crystallized from numerous conversations about our core mission and where the market was heading. We were tracking the rise of using AI agents for various marketing tasks — people were getting excited about AI agents building campaigns or writing copy.

But as we explored it, this kind of “agentic marketing” felt incomplete for us. It described how marketing could be done – using agents – but it didn’t capture the why or the outcome. It answered “what is it?” – marketing with agents – but not the critical questions CMOs and their finance teams grapple with daily: how does this make my job easier? How does it let my team focus on strategy? How does it demonstrably improve results and help us grow faster, especially when budgets are tight and we need to do more with less?

The breakthrough came when someone drew the analogy to compound interest – the idea of growth building exponentially on itself over time. That resonated deeply. The word “compound” perfectly captured our vision: to move marketing beyond discrete campaigns into a continuous system where every action informs the next, accelerating growth. It’s about transforming marketing from a cost center focused on consumption to a value creator focused on compounding returns. So, we defined this new category, compound marketing, powered by the Compound Marketing Engine, to shift the focus towards measurable, accelerating growth – aiming for step-changes like 1% weekly improvement versus 1% quarterly.

The process starts with your single source of truth in the data cloud, which can include first-party customer data as well as business intelligence that informs the Engine and the campaigns you generate. Because GrowthLoop operates directly on the customer’s data warehouse, the Compound Marketing Engine has secure access to the freshest, most complete data foundation without needing data duplication.

Once the context is set, our specialized AI agents begin collaborating. We have agents focused on data, audience building, journey orchestration, research, and performance insights. When a marketer initiates a task or asks a question – say, “Launch a campaign targeting users who abandoned their cart with products over $100” – the relevant agents work together. You might see the Audience Agent proposing segments based on warehouse data and the Journey Agent mapping out the touchpoints. It’s designed to feel like interacting with a highly capable team.

The Supervisor Agent oversees this multi-agent collaboration, ensuring coherence. Crucially, human creativity and approval remain central to the workflow. The agents make suggestions, build drafts, and surface insights, but the marketer makes the final call before anything goes live. We believe strongly in keeping the human in the loop, especially for strategic oversight and creative problem-solving. The agents are also still learning and improving – think of them like the first mate to the marketer’s captain. Marketers can and should review, edit, or refine any element suggested by the agents within the campaign canvas before exporting.

That concept resonated strongly with our vision. David Edelman’s book discusses how the future isn’t just about having smart tools, but about integrating them intelligently. The “smart” piece is the human strategy – the marketer defining the goals and optimizing the overall approach. The “integrator” piece is having the right tech ecosystem working together seamlessly, allowing you to mix and match point solutions effectively to achieve those goals.

That’s precisely what the Compound Marketing Engine enables. It acts as that intelligent integration layer, sitting on top of the central data warehouse and connecting to various execution channels (paid, owned, product, even support tools like Intercom). It orchestrates the flow of data and insights, allowing marketers to leverage specialized tools (like Braze for email, Google Ads for paid advertising) more effectively because the underlying audience and journey logic is unified and intelligent. It’s about making the entire stack smarter and more cohesive, driven by the organization’s data cloud and the marketer’s strategy but accelerated by AI agents handling the complex coordination and execution tasks.

While it’s still early days since the official launch, the results from customers using the underlying agentic capabilities have been very promising. We have clients – including Allegro, which is one of the largest e-commerce companies in Europe – who have doubled their return on ad spend (ROAS) by leveraging these automated insights and orchestration features. The primary driver is the acceleration of the marketing cycle. Tasks that previously took days or weeks – pulling data, segmenting, briefing, building, launching, analyzing – can now be compressed significantly.

This speed allows teams to dramatically increase their testing velocity – essentially taking more “shots on goal”. They can experiment with more creative variations, audience segments, and journey flows than were manually feasible before. The continuous feedback loop provided by the Insights Agent means they learn much faster what’s working and what isn’t, allowing for rapid iteration and budget reallocation towards winning strategies. It’s not just about doing things faster; it’s about learning faster and compounding those learnings into better results over time.

What I truly love about being a chief of staff, especially here, is that there’s always a new problem to solve, and my role is different week to week, which keeps things incredibly dynamic and interesting. The martech ecosystem is tight-knit, so you always have people to collaborate with to solve problems.  

A typical week usually starts with kickoff calls on Mondays across our teams – customer success, sales, and partnerships. We align on the week’s goals: how is the pipeline looking? What’s our forecast? What deals are we trying to close? Are there opportunities needing executive sponsorship? This feeds into our leadership meeting, where we review all business units, align on decisions, and prep for things like board presentations. For example, leading up to our recent Compound Marketing Engine launch, leadership was deeply involved in every detail – naming, messaging, and video assets. It was a real company-wide push.  

After Monday, my week often involves many external calls: managing vendor relationships, talking to customers and partners, joining sales conversations. Internally, I’m playing project manager across different areas, ensuring our executive team’s time is optimized and that everything on the ground floor gets executed effectively. The specific focus changes constantly depending on the priorities, which is what makes the role so exciting.  

Definitely. We actively avoid solving the same problem over and over. If we see an issue surface multiple times, maybe three times, we pause and ask if we need to fundamentally fix the underlying infrastructure. That’s often where I step in. Maybe we need a new system, a different process, or even a dedicated cross-functional team.

A concrete example happened with our engineering team. We saw an uptick in support tickets and noticed some data pipelines weren’t running as smoothly as they should. So, we initiated what we called a “Dyson Sprint” – named after the vacuum – with the simple goal of cleaning things up. For one week, the team dedicated itself entirely to clearing the ticket backlog, optimizing those pipelines, and refining user stories. It was about getting our house in order so that future innovation wasn’t slowed down by existing infrastructure debt. It ensured we had a clean slate before major pushes, like the Compound Marketing Engine launch. We schedule these focused cleanup sprints periodically whenever we identify a recurring bottleneck that needs dedicated attention.  

During the pandemic, I was contracting with a few different companies, exploring various industries to figure out where I wanted to land next. My background is in digital marketing and advertising operations. When I connected with GrowthLoop’s founder for a potential contracting role, two things immediately clicked. First, the industry itself. I have a minor in computer science, and GrowthLoop offered the chance to merge my operational knowledge with the technology side, which really excited me.  

Second, the way the business was being run was impressive. They were bootstrapped for the first five years – profitable, efficient, growing smartly without outside funding initially. That’s rare and signaled they were doing something fundamentally right. As I learned more through contracting, their cloud-native approach stood out. Instead of creating another data silo like traditional CDPs, they built a solution that sits directly on the customer’s data warehouse. This focus on security, efficiency, and meeting customers where their data already lives felt forward-thinking. I saw the massive opportunity. When they offered me a full-time role, it was an easy yes. It felt like the right place to be to build something different and impactful over the next five to seven years.

Partnerships are fundamental to our model. On the cloud side, we need to build with platforms like Google Cloud, Snowflake, Databricks, AWS, etc. Our solution is only viable if customers have their data consolidated in a warehouse, so it would be a disservice not to align our roadmap closely with theirs. We’re constantly talking to their product teams, understanding their direction (like Google’s Agent2Agent Protocol or Snowflake’s Data Cloud initiatives), and ensuring GrowthLoop complements their ecosystem. It’s about creating synergy where our platform enhances their value, and vice versa.

With the end-channel tools – the Brazes, Salesforce Marketing Clouds, Klaviyos of the world – our approach is about making them smarter, not replacing them. Customers already use these platforms. Our goal is to help them get more value by activating richer, more timely data from their warehouse through GrowthLoop. We had a customer, in partnership with Braze, who optimized their usage significantly by managing segmentation and journeys in GrowthLoop, reducing storage costs and complexity within Braze itself, allowing them to get more out of that investment. It’s about enabling a more intelligent, integrated ecosystem where our platform acts as the activation layer on top of the warehouse, feeding smarter audiences and triggers into the tools marketers already use.

I don’t subscribe to the idea that humans will lose their jobs to AI. I think humans will always have a crucial place. Evolution shows species adapt, and humans are incredibly adaptable. The key is adopting technology in the right way, creating a symbiotic relationship where AI enhances human capabilities rather than replacing them.

Within the Compound Marketing Engine, AI agents handle much of the complex, manual work – data queries, audience building, suggesting campaign structures, and analyzing performance. This frees up marketers to focus on higher-level strategy, creative thinking, understanding brand nuances, and making the final judgment calls. The human remains firmly in the loop, providing the essential context and oversight that AI lacks. AI can optimize and accelerate, but the strategic direction and the understanding of complex business goals still require human intelligence and intuition.

The one I keep coming back to is “don’t sweat the small stuff.” It’s incredibly easy, especially in a fast-paced role like mine, where things change constantly, to get bogged down in minor details or frustrations – a meeting rescheduled, a small process hiccup, wearing the wrong shirt colour, whatever it might be.

I use a mental framework of “above the line” and “below the line.” Above the line items are the core functions critical to business success: building a product, marketing it, selling it, and delivering it. Everything else – the small daily frictions, minor disagreements, unavoidable changes – falls below the line. When I start feeling stressed about something minor, I ask myself: is this truly impacting one of those core functions? If it’s below the line, I consciously decide not to waste energy on it and move on. If it’s above the line, then it deserves focus and attention. We’re human, mistakes happen, things evolve. Adapting to the changing tide without getting overwhelmed by the small stuff is key.

I’m always reading multiple books at once! Right now, Personalized by David Edelman and Mark Abraham is one – it’s fantastic for thinking about customer strategy in the age of AI. I’m also working on my writing, so I picked up A Few Short Sentences on Writing after hearing Lenny Rachitsky mention it. On the business side, I’m reading Good to Great by Jim Collins, a strategy book by Seth Godin, and Mark Benioff’s Trailblazer – he’s a huge innovator. For podcasts, Lenny’s Podcast is a weekly listen, and Humans of Martech is great, too.

That’s tough because it depends so much on the specific business problem! But if I had to generalize, I’d start with the data warehouse as the foundation. We use ours at GrowthLoop for everything – support tickets, customer health, product integrations, our own data management. It’s fundamental. Beyond that… it gets tricky. Maybe Lovable, the tool for building internal apps quickly, because it offers so much cross-team flexibility. But honestly, for the third tool, I’d need to know what the specific business goal was. Is it sales optimization? Then maybe a call recording tool. Is it team collaboration? Maybe Fellow. Is it coding efficiency? Maybe Cursor or Copilot. You have to start with the business problem and pick the tools that solve that, rather than picking tools in a vacuum.