The vast majority of our customers are building ambitious SaaS businesses by way of smart, opinionated products. They don’t have massive amounts of capital, rooms full of Enterprise AEs, or six-figure contracts. In the many years I’ve been at ChartMogul, my conversations with founders and operators often revolve around the same thing:the weeds of PLG complexity.
I wanted to create an event that didn’t shy away from the messy, important stuff.
Broken n8n workflows. Activation paths that look obvious internally, then fall apart in the real world. Compensation plans for hybrid GTM motions that drive the right incentives. This year, we’re running a series of intimate, city-based meetups for PLG founders and operators. We’re calling it The Product-Led Growth Lab series.
The first meetup took place in San Francisco during the week of Stripe Sessions. Brenna Loury, CRO of Doist spoke candidly about the challenges of moving up market in a PLG business. The group was still deep in discussion when it was time to move to the rooftop for refreshments and be joined by other conference attendees. This was a great sign that the topics were valuable, and it was much more enjoyable to finish the debates over some charcuterie and beer in the sunshine!
From there, we traveled on to the Croatian coast. In partnership with SaaStanak, we pulled together two afternoons of PLG content overlooking the Mediterranean sea. Each session was standing room-only during a heat wave and yet people were engaged: asking sharp questions, sharing what they were seeing in their own companies, and making new connections. I’m really proud of how it turned out.
There was so much useful, operator-level knowledge shared, we wanted to make it available for anyone who didn’t make it to this year’s SaaStanak. So in true ChartMogul fashion, we’re recapping each session, highlighting the practical takeaways, and sharing the decks so you can come back to them whenever you need.
Jump to a presentation
- The Expansion Engine: Architecting a Scalable PLG Stack for Automated Upsells, Tom Andrews
- When Automations Guess Wrong: Behavior Shows What Happened. Users Tell You Why, Aleksandra Korczyńska
- From Traditional SaaS to AI-Native in 5 Days: 3 Companies That Did It, Wes Bush
- LTV Has a Blind Spot: Findings from 3,700 SaaS Businesses Over 6 Years, Thomas Anastaselos
- People-Led Growth: Fixing Your Product’s First Impression and Path to Activation, Paulina Spaccarotella
- Invisible PLG, Ana Zrno
- [Workshop] From $1M–$20M ARR: Growth Levers and New Signals for Success, Sara Archer and Kyle Poyar
- Where PLG Meets Sales: Building a Hybrid Growth Engine in Practice, Marko Kebe
- The Launch to Adoption Gap: Why Fast Shipping Doesn’t Equal Product Growth, Natália Kimličková
The Expansion Engine: Architecting a Scalable PLG Stack for Automated Upsells
Key takeaways
- Expansion needs connected data, not just better timing. Tom showed how expansion becomes a larger share of revenue as SaaS companies scale, but most teams are still trying to act on signals trapped across product, CRM, billing, and automation tools. The fix starts with building a lightweight “Data Rome”: a central layer where usage, customer, and account data can be routed into workflows.
- Automated upsells depend on signal maturity. Tom broke expansion signals into three levels: thresholds, trends, and propensity. Thresholds are simple and reliable, but reactive. Trends use velocity and frequency to spot momentum earlier. Propensity models are more predictive, but require deeper historical data, data density, and technical maintenance.
- AI works best inside the automation layer. The idea here isn’t to “AI”ify the upsell motion. Rather, it’s about embedding AI into key workflows: triggered by usage data, enriched with account context, routed by business logic, and used to draft outreach or create tasks for the team.
Why it matters
Expansion is often talked about like a sales execution problem: find the right moment, write the right email, make the right offer.
Tom’s session showed that, for many SaaS businesses, it’s actually an infrastructure problem. When the right data flows into the right workflow, expansion becomes more timely, more relevant, and much easier to scale.
When Automations Guess Wrong: Behavior Shows What Happened. Users Tell You Why
Key takeaways
- PLG needs to rebuild the listening loop that sales-led teams never lost. Sales-led conversion is higher because reps don’t just react to behavior. They ask what’s getting in the way, then adapt the message, timing, proof point, or offer. Aleksandra’s point was that PLG teams can replicate that without making everything manual: ask targeted questions at key moments, turn those answers into structured data, and use them to route users into more relevant automated journeys.
- Feedback is the missing segmentation layer. Behavioral data can tell you that a high-intent user didn’t upgrade, or that a once-active account has gone quiet. But it can’t tell you whether the blocker is budget, security review, missing functionality, timing, low job volume, or a bad experience. Aleksandra showed how asking one well-timed question turns a vague lifecycle trigger into a specific next step.
- AI makes open-text feedback operational at PLG scale. You can now use AI to categorize open-text data, identify patterns, and route users into the right workflow. That turns qualitative feedback into structured data teams can actually use in lifecycle automations.
Why it matters
Read this deck if your lifecycle automations are technically “working,” but still feel too generic. Aleksandra’s session is a good reminder that behavior tells you what happened, but user feedback tells you what to do next.
From Traditional SaaS to AI-Native in 5 Days: 3 Companies That Did It
Key takeaways
- AI-native changes the order of value. Traditional SaaS makes users earn the outcome: learn the interface, configure the settings, import the data, build the workflow, then eventually see value. Wes’s point was that AI-native products should flip that sequence. Start with the smallest meaningful outcome a user can get in 60 seconds, then design the product experience around delivering that result as quickly as possible.
- The product has to have an opinion. A lot of traditional SaaS puts the work on the user: blank states, setup decisions, configuration, “choose your own adventure” onboarding. Wes argued that AI-native products should do more of the deciding for the user. Let the user approve instead of making them build from scratch.
- Don’t automate away the user’s sense of ownership. Wes used the Betty Crocker cake mix story to show why “easier” isn’t always better. The original mix only required water, but people rejected it because it felt too artificial, like they hadn’t really made anything. When Betty Crocker added the egg back in, people felt involved again. For AI-native products, the lesson is to remove the tedious work, but keep the one small action that makes the result feel trusted, personal, and earned.
Why it matters
This session turns “AI-native” from a vague strategy into a product exercise: pick the first valuable outcome, remove the work around it, keep the moment that creates trust, and build the experience backwards from there.
The deck is especially useful because it gives teams concrete prompts and frameworks they can use in conversation: What could the product do before the user has to learn the product? Which decisions should AI make by default? And what small action still needs to stay so the user feels ownership of the result?
LTV Has a Blind Spot: Findings from 3,700 SaaS Businesses Over 6 Years
Key takeaways
- LTV is useful, but it is not as trustworthy as it looks. Thomas tested 39,000 LTV predictions against what actually happened across 3,688 SaaS businesses. The median cohort generated 9% less revenue than LTV predicted over 12 months, which is not catastrophic. But the median hides the real problem: 34.5% of cohorts were wrong by more than 50% in either direction.
- LTV breaks when ARPA and churn assumptions break at the same time. The standard formula uses blended account-level ARPA and recent account-wide churn. That means it can overestimate new customers who have not expanded yet, or misread cohorts that churn differently from the broader customer base. Sometimes those errors cancel out. Sometimes they compound, and the number gets very wrong.
- Some businesses should discount LTV more than others. The data showed that reliability varies by segment. E-commerce, infra/dev tools, and workplace/productivity had higher rates of badly wrong LTV predictions. Larger accounts also had more systematic overprediction. Thomas’s practical recommendation was not to throw LTV out, but to treat it as a compass and use it alongside cohort revenue curves, net MRR movements, and observed payback period.
Why it matters
This presentation showcases brand new data from our insights team. LTV can still help teams make decisions about CAC, payback, and segment investment, but only if they understand where the formula breaks and check it against what customers are actually doing.
People-Led Growth: Fixing Your Product’s First Impression and Path to Activation
Key takeaways
- Users experience onboarding emotionally. Paulina reframed activation through a more human lens: users are deciding very quickly whether they’re in the right place, whether the product understands their problem, and whether they feel capable of moving forward.
- Great onboarding reduces the distance between hope and first win. Every user arrives with a problem, but also with hope: that this product might finally help them do the thing they came to do. Paulina’s human activation loop captures that journey: Problem → Hope → First Win → Momentum → Trust. The faster a product can help someone feel real progress, the more trust it earns.
- Audit the experience through one specific human. Paulina’s workshop pushed teams to stop auditing onboarding as themselves and start auditing through the lens of their primary user. What did this person hope would happen? What do they expect next? Does this screen create momentum, or does it introduce confusion, friction, or a trust break?
Why it matters
This content is a useful reset for teams that treat onboarding like something you “set and forget.” Paulina’s framework gives teams a simple way to revisit it every month: look at the experience through your user’s eyes, find where trust breaks or momentum disappears, and fix one moment that gets them closer to a first win.
Invisible PLG
Key takeaways
- The first evaluator may no longer be human. Ana challenged one of the default assumptions behind PLG: that there is always a person navigating the product, clicking through the flow, and emotionally responding to the UX. In an agent-first world, the “buyer” might be an operations agent tasked with reducing support backlog, comparing vendors, testing a sandbox, and measuring cost.
- Invisible PLG is built for systems, not just people. The shift is from persuading humans through UX to enabling systems through docs, schemas, APIs, machine-readable pricing, autonomous sandboxes, and deterministic onboarding. The product still has to be useful and usable, but it also has to communicate value programmatically.
- Documentation becomes GTM infrastructure. Ana’s uncomfortable point was that the things many teams treat as support or engineering concerns may become acquisition infrastructure. If agents are part of the evaluation path, documentation, schemas, pricing logic, API reliability, and outcome instrumentation become part of how the product gets discovered, evaluated, and bought.
Why it matters
This session gives teams a useful diagnostic for the next era of PLG: could an AI agent discover, evaluate, test, price, and integrate with your product on its own?
If the answer is no, your product may still be visible to humans, but increasingly invisible to the systems shaping how software gets evaluated.
[Workshop] From $1M–$20M ARR: Growth Levers and New Signals for Success
Key takeaways
- The growth machine has to change after $1M ARR. ChartMogul data shows that getting to $20M ARR is not just about doing more of what got you to $1M. As companies scale, the strongest performers improve the quality of the revenue machine: higher ARPA, more expansion, better retention, more annual contracts, and stronger reactivation.
- Revenue quality beats revenue speed. Growth rate matters, but speed alone is not the full story. Sara and Kyle made the case that durable growth comes from getting more out of the customer base you already have: retaining better, expanding more, pricing smarter, and protecting the revenue you worked so hard to win.
- The founder’s edge has to become team language. Remarkable companies are usually built by remarkable people, but the real unlock is making that restlessness scalable. The founder’s taste, standards, and refusal to accept “good enough” need to become a way of working the whole team can understand, repeat, and raise themselves toward.
Why it matters
This workshop turns “how do we get to $20M ARR?” into a sharper operating question: what has to get better as we scale?
The answer is rarely one silver bullet. It’s usually a set of compounding improvements across ARPA, retention, expansion, focus, and execution, paired with the leadership standards that make those improvements stick.
Where PLG Meets Sales: Building a Hybrid Growth Engine in Practice
Key takeaways
- Hybrid growth needs flexible swimlanes. Marko showed how BetrSign moved from a relationship-led, enterprise-heavy motion into a hybrid model with self-serve, trial-to-buy, digital sales, and account-based sales. The trick is not assuming every household-name logo needs an enterprise sales motion, or that every smaller account wants to self-serve. You need clear lanes, but enough flexibility to move buyers into the journey that matches how they actually want to buy.
- “Where are all the self-serve conversions?” is a very real moment in many PLG journeys. Before declaring yourself product-led, it helps to ask whether the product can actually be evaluated, adopted, and purchased without seven meetings, three lawyers, and one heroic sales engineer.
- Compensation and KPIs have to change with the motion. A hybrid GTM model gets weird fast if the team is still rewarded like a traditional sales org. BetrSign moved away from booked ACV as the main measure and toward realized revenue, MRR growth on existing accounts, product feedback loops, demos, and content inputs.
Why it matters
Hybrid GTM is not just a self-serve funnel bolted onto sales. It’s a way to solve the system and process problems that show up when different buyers need different paths: how you route accounts, define ownership, design incentives, set product boundaries, and meet customers where they actually are.
The Launch to Adoption Gap: Why Fast Shipping Doesn’t Equal Product Growth
Key takeaways
- Fast shipping does not equal product adoption or revenue growth. The core tension in Natalija’s talk is that teams have optimized for release velocity, but not for user attention. Shipping more features can create the illusion of progress, while adoption stays flat or even gets worse. The real question is not “did we ship it?” It’s “did the right users notice it, understand it, try it, and make it part of their workflow?”
- A release is a technical event. A launch is a behavior-change system. Natalija drew a useful line between releasing code and actually creating adoption. Too many launches still rely on passive discovery, changelogs, one-off announcements, and broad campaigns. Modern launches need to be scored, targeted, landed, and measured with intention.
- Targeting should be behavioral, not just demographic. The deck makes a strong case that demographics are proxies, but behavior is signal. The best launch moments come from what users are already doing: hitting friction, repeating a manual action, reaching a milestone, or entering a lifecycle moment where habits are still forming.
Why it matters
This is one of the freshest takes on SaaS product marketing I’ve seen in a while because it names the problem most teams are living inside: the product is shipping faster than users can absorb.
Natalija’s framework reframes product marketing from “announce what shipped” to “design for adoption.” In a world where every team is launching constantly, the advantage goes to the companies that can score what deserves attention, target the right users at the right moment, and turn a release into actual behavior change.
Closing thoughts
Taken together, the sessions made one thing pretty clear: PLG is no longer just about removing sales from the buying journey.
It’s data architecture. Feedback loops. Onboarding psychology. Hybrid GTM design. Launch systems. Agent-readable infrastructure. And yes, still, the very human work of helping someone understand why your product matters to them.
PLG is getting more technical, more operational, and more human all at once.
