Activity Is Not a Strategy: A Conversation with Emilia Korczynska

Everyone wants growth. Almost nobody has a strategy for it. Most SaaS companies, when they say they care about growth, are really just producing a lot of content and hoping something sticks.

Emilia Korczynska has spent years in the trenches of that problem. She's the VP of Marketing at Userpilot, where she built the content engine and helped make the company a major voice in product-led growth. She didn't start in SaaS. She was a journalist and a linguist, ran a translation agency, and built and exited a startup in China before finding her way into marketing. Which means she thinks about how products actually grow in the real world, not how the playbooks say they should. And her opening move is to separate two things most teams conflate: activity and strategy.

The six-pack problem

Strategy, Emilia says, is a repeatable set of activities with a clearly defined outcome. What most companies do instead is throw tactics at the wall, and worse, they often have no way to even measure what stuck. So they set out with no clear objective and then have no idea whether it worked, or whether they simply didn't do enough of it to move the needle.

Her analogy is the one I'll remember. Everyone understands that a six-pack takes going to the gym several times a week, specific exercises, and a certain diet. But there's a marketing equivalent of sitting on the sofa eating pizza, doing ten squats, and expecting a six-pack. People run a bit of SEO or cold outreach, expect the result to blow their mind, and don't notice that the effort they put in was wildly out of proportion to the outcome they wanted. The expectations aren't just high. They're unrealistic about what input produces them.

Who, not how

Emilia's own hardest lesson came moving from doing the work to leading it. What she had to unlearn was control. She'd started with the mindset that she needed to be the expert in everything and do it all better than her team, which she now describes as exactly the opposite of what lets you scale.

The book that flipped it for her was Who Not How. You don't actually need to know how to do everything. You need to know who can do it better than you. A team that looks up to you for every decision, that you have to micromanage and instruct, is a team you've capped. Her interview test for the opposite trait is sharp: hand a candidate a tool they've never used, ask them to accomplish a specific outcome with it, and watch two things. How fast they grasp how it works, and whether they'll dig for answers on their own instead of turning to you.

The straitjacket that makes good people quit

That flows into a culture point. Proactive people with a growth mindset, the ones who'll figure things out without running back to the SOP or their manager, will wither in an environment where they have to ask permission for every tiny decision and document everything to the letter. It's soul-destroying, Emilia says, and eventually they retract and quit. There's an individual side too: with the resources available now, there's genuinely no excuse left for "I don't know how to do something." It's both. Some people need the environment to let them run, and some just haven't learned yet that the answer is a search away.

Activation is a job, not a wizard

On the product side, the biggest mistake companies make between signup and activation is leaving users to their own devices. No onboarding, a confusing first-run experience, friction everywhere, and users who don't even know what outcome the product could get them.

Her fix starts with knowing your audience, which means asking just enough at signup and no more. Ask the role. Don't ask for company size, phone number, and shoe size, because every extra field lifts your drop-off rate. Everything else you can enrich from a work email. The one thing you can't enrich, and the single most important thing to know, is the job the person came to get done. At Userpilot, product managers, product marketers, and customer success folks all sign up with overlapping responsibilities but different objectives, and the activation point is different for each of them because it's defined by the outcome they want. So the real question isn't their company size. It's what they're trying to achieve, and then getting them to that value as fast as you can, ideally in the very first session.

The leaky bucket, and when churn is really a product problem

I pushed on retention, the right-hand side of the bow tie, the leaky bucket everyone underinvests in. Emilia agreed it matters, then complicated it in a useful way. If you only retain, you're not growing, you're staying level. And if you have a massive churn problem, the honest move is to ask whether it's actually a product, positioning, or market problem, because you cannot plug massive churn with a few tooltips, an onboarding flow, or an email sequence. If the only way you can grow is by stopping churn, something is fundamentally wrong with the business itself.

So she thinks companies are justified in weighting acquisition more heavily, up to a point. There's always a crossover where retention has to get real attention, and it comes sooner the larger your ticket size, because when you're landing big enterprise accounts, every customer who leaves is a serious chunk of ARR and a dedicated customer success function starts to pay for itself. On a $20-a-month self-serve product, it never will.

GEO is overrated, and the listicle penalty

I asked what tactic everyone's obsessed with that she thinks is overrated. Her answer was fast: GEO, optimizing for AI answers. Her reasoning is that it isn't the same game as SEO. There are so many possible queries and so many possible answers, shaped by each user's history, memory, and the exact phrasing and context of their prompt, that you simply can't control what an LLM retrieves in real time. You can raise your chances with genuinely good content and mentions in reputable sources in the right context, but anyone selling you "number one in ChatGPT for $500 a month" is running a scam.

She also flagged the backlash. Because AI systems lean toward recency and listicles, people started gaming it, taking an old post, changing "best marketing tactics in 2023" to "in 2026," refreshing the publish date, and blitzing out self-promotional lists. Google responded in late January with what SEOs are calling the listicle penalty, hitting sites that overdid exactly that. As content production gets cheaper, more companies overplay the hand.

Context beats the prompt

Since AI content is now nearly free to produce, I asked whether that makes marketing more or less important. Emilia refused the easy answer. Someone still has to be in the driving seat deciding what to produce, and the tools aren't autonomous, they need direction and judgment. But her sharper point was about who wins when everyone has the same tools. Counterintuitively, the field gets less level, not more, because the advantage flows to whoever has the most clout, authority, and real experience.

The mechanism is context. The quality of what an AI produces depends on the quality of what you feed it, and not just the prompt, the context. The same prompt with different context produces crap in one case and genuinely high-quality work in the other, and the difference is who can supply original information, real data, personal experience, and the weight of their own point of view. Which is also her answer to what a great marketer's durable skill is: knowing what to build, what's worth investing time in, and the most efficient way to orchestrate the tools. Anyone can vibe-code and deploy almost anything now, but knowing what's worth building, and keeping it from quietly becoming unscalable, still takes a professional.

What founders get wrong

The thing SaaS founders most consistently misunderstand about marketing, Emilia says with a laugh, is how much effort and budget it takes to hit the outcome they want. They'll expect a thousand signups, or assume a 10% positive reply rate on cold outreach is normal. She once shared her team's semi-automated outreach numbers, the best they'd done in years and a real achievement for their personas, and the outbound crowd jumped in to call the results weak. But willingness to pay in her market is low and product managers aren't especially responsive to outreach, so those were strong outcomes. The pattern underneath is that expectations exceed reality by such a wide margin that marketers get disappointed and abandon perfectly viable channels far too soon.

Key takeaways

A few things worth keeping.

Activity is not a strategy. A strategy is repeatable activities with a measurable outcome. If you can't tell what stuck, you're just making noise.

Match effort to outcome. Most disappointment in marketing is a ratio problem. You can't do ten squats and expect a six-pack.

Activation is defined by the job. Ask what the user wants to achieve, then get them there fast. It's about the outcome, not the company size.

Churn you can't fix with tooltips is a product problem. If growth depends entirely on stopping churn, look upstream at product, positioning, or market.

Context is the moat. When everyone has the same AI, the winner is whoever feeds it the best original data, experience, and point of view.

Frameworks worth stealing

The six-pack test

Before calling something a strategy, check two things: is it a repeatable set of activities with a clearly defined outcome, and can you measure whether it's working? Then sanity-check your effort against your expectation. If the input is ten squats, don't be surprised when the output isn't a six-pack.

Who, not how

As you move into leadership, stop trying to be the expert at everything. You don't need to know how to do the work, you need to know who can do it better than you. And screen for people who figure things out: hand them an unfamiliar tool, give them an outcome, and watch whether they dig or wait to be told.

Activation is the job, not the wizard

Don't leave new users to wander, and don't interrogate them at signup. Ask the one thing you can't enrich from a work email: what are they here to achieve. Then design the fastest path to that value, ideally inside the first session. Activation is reaching the outcome, and the outcome, not firmographics, defines it.

Context over prompt

When everyone has the same models, the prompt is commodity and the context is the edge. Feed the tool your original data, your real experience, and your actual point of view. The gap between generic output and genuinely good output is almost always a context gap, not a prompt-writing one.

Quotes worth keeping

The lines I wrote down.

Strategy is a repeatable set of activities with a clearly defined outcome.

There's a marketing equivalent of eating pizza on the sofa, doing ten squats, and expecting a six-pack.

If the only way you can grow is by stopping churn, something is fundamentally wrong with your business.

The quality of the output depends on the quality of the input. And not just the prompt, the context.

And the one that flips a comforting myth on its head.

If everyone has access to the same tools, the field becomes less level, not more.

Rapid fire round

Same questions every guest. No prep, no warning. Here's how Emilia handled it.

Best advice you've ever received? If you feel like quitting, learn to rest, not to quit. Her caveat: if it's toxic, a hand on a hot stove, then absolutely quit. But if you love the work and you're just exhausted from going hard at it, ask whether you actually want out or you just need rest. Founders, she says, don't ask themselves that enough. And you can always outsource the parts you hate for a little money and spend your time on what you're good at.

Advice you wish you'd listened to? That she stays too long in situations she's outgrown. She's been a bit risk-averse, too reluctant to spend money or make investments, and she catches herself wishing she'd started things sooner.

What would you tell your younger self? Just go for it. Money is a renewable resource, and it's a lot easier to make up lost money than lost time.

Ongoing challenge that keeps you up at night? Leveraging the full potential of the tools, AI and Claude Code specifically. Everyone in the industry feels they're falling behind on adoption. She doesn't want her team reinventing the wheel on repetitive work, and she's already automated landing-page generation end to end, but she calls that just the tip of the iceberg.

Favorite spot? Anywhere that serves laksa or authentic Chinese food. She lived in China for a couple of years and is a devoted fan of Asian food.

Tool you can't live without? Claude Code. She calls it incredible, though she still keeps ChatGPT around for smaller everyday tasks.

---

Emilia Korczynska is the VP of Marketing at Userpilot, where she built the content engine that helped define the product-led growth category. Before SaaS she was a journalist and linguist, and she built and exited a startup in China. Find her on LinkedIn.