Don't Assume You Are the Customer: A Conversation with Piotr Kacala

Piotr Kacala has helped shape some of Poland's most iconic product and engineering cultures, and he's blunt that it came at a cost: "lots of gray hair and scar tissue, metaphorically."

Piotr is a product and tech leader with nearly two decades of experience, a board advisor and interim CTPO. He spent seven years leading engineering at GOG.com inside the CD Projekt ecosystem (the Witcher studio), was CTO and board member at Displate, and founded Product Cards, through which he coaches companies and runs AI Product Heroes, a cohort course on building products with AI. He's seen hyper-growth, complex stacks, and the messy reality of building real products with real teams.

Don't assume you are the customer

The one product principle that held true across CD Projekt, GOG, and Displate, three completely different businesses, is this: don't assume you are the customer. Teams build for themselves, optimize for their own taste, and assume their intuition represents the user, and that's exactly when you start making expensive mistakes. His example is GOG Galaxy 2.0, which had the right problem but not for the right people. The promise was to replace all your other gaming clients, Steam, PlayStation, Origin, in one single interface: novel, hard, and, it turned out, wrong. When they finally started listening to customers, too late, they discovered GOG users didn't have many clients to manage in the first place. A costly revelation.

Telling smart people what to do destroys them

Asked how founders and CTOs accidentally destroy good teams, Piotr lists his "sins." First and worst: telling people what to do. It makes no sense to hire smart people, pay them a ton of money, and then treat them like keyboard monkeys who just execute tasks. It hits both of the two reasons he hears when people quit, "I'm not developing" and "I don't believe in what we're building," because they can't grow and they had no say. The second sin is creating a process for everything, drowning the company in ritual so no one has to think, which breeds mediocrity, not innovation: death by a thousand coordination needles. The third is chasing every shiny object, AI now, crypto before, instead of maintaining focus.

The Dunbar breakpoints

Piotr has led teams from startup to scale multiple times, and the numbers where things break are consistent, following Dunbar. Around 10 to 15 people you assemble a second team, and suddenly you need explicit process, because the coordination overhead rises and what worked implicitly with one team breaks with two. Around 50 you stop knowing everyone, and culture no longer transmits by osmosis, you have to transmit it explicitly. Around 150 you're running an organization, not a team, entirely different skills, with a mid-management and leadership layer that's a big step up.

The middle-management layer is what breaks at scale

The thing that broke unexpectedly at Displate was the middle-management layer. Startups hand out senior titles like candy, deliberately, to attract and retain talent, but some of those people wouldn't be hired at that level in a bigger company. When the bar rose, some couldn't keep up, and it wasn't their fault: thrown into very senior roles with lots of responsibility, not enough support, and no prior experience, they needed time to grow that they didn't have. Worse, it compounds, a mediocre manager makes everyone below them mediocre or drives the good people out, creating a layer that blocks information upward and distorts execution downward. It hid in plain sight for months.

His approach to the hard conversation: state where the bar is, name the pattern you're seeing, and work closely for three months to see the trajectory, with scenarios like an individual-contributor path if it doesn't work out. What he stopped doing was wishful thinking, keeping his fingers crossed, because that's harmful to the person, the team who can see it, the company, and to you, since you become seen as the weakest link.

Delivery is a symptom

Founders come to him saying "speed up our delivery, it's costing us money." What they usually need is a good strategy, not a deck that says "our goal is to make money", and discovery muscles: understanding the customer and competition through continuous probing. Delivery problems are typically symptoms of unclear priorities, no customer connection, and building the wrong things. Optimizing delivery in that state is like optimizing a cart that gives you stones instead of gold, you ship things nobody needs, however pretty or fast they are.

AI is a sprinkle, not the forefront

Piotr's whole thesis with AI Product Heroes is that AI should be a sprinkle on top of the product development lifecycle, not the forefront. Every new technology runs the same cycle, innovation, big valuations, founders treating everything as a nail because they're holding a hammer, pushing the circle into every shape in the toy. Customers are tired of AI everywhere, and not every product needs it ("you don't need AI to make your sandwiches"). The dangerous fantasy is "in six months I'll fire half my engineers and everything will be great." The reality is a disaster. So he teaches the fundamentals: talk to customers, do discovery, build cheap prototypes you can show next week, and learn what AI can and can't do.

The loneliness at the top

Piotr is unusually candid about the human cost. It was a lonely road, and he made a decision to wear masks to be there for his team even when he didn't have the capacity, which took a real toll. His advice is to build your own unofficial "board of directors", peers at similar-size companies, mentors, coaches, for confidence and outside perspective, and to invest in physical and mental health. He points to a founders' community session on suicide rates (roughly three times higher) that turned into a group therapy moment, and warns that the expectation for executives to hold a tough facade is harmful, because you grow into the archetype.

Key takeaways

Don't assume you are the customer. Building for your own taste produces expensive, confident mistakes. GOG Galaxy was novel, hard, and wrong.

Telling smart people what to do destroys them. They stop growing and stop believing. You paid for talent and get a keyboard monkey.

Teams break at Dunbar numbers. ~15 (a second team needs process), ~50 (culture stops spreading by osmosis), ~150 (you run an organization).

Delivery is a symptom. Slow shipping usually means unclear strategy and no customer connection. Fix discovery, not just the cart.

AI is a sprinkle. Build the product muscle first, then use AI on top. Not every product needs it, and firing your engineers won't end well.

Frameworks worth stealing

The trajectory conversation

When a manager or hire is underperforming, don't cross your fingers. State plainly where the bar is and the pattern you're seeing, then work closely for a defined window (he uses three months) to see the trajectory. Agree in advance on the fallback, an individual-contributor path or another role, so it's honest for the person, the team, and the company, instead of dragging out wishful thinking.

Discovery before delivery

When someone asks you to speed up delivery, check the layer underneath first: is there a real strategy, and a genuine connection to customers and competition? Delivery pain is usually a symptom of unclear priorities and building the wrong thing. A faster cart full of stones is still worthless, so fix what you're shipping before how fast you ship it.

Quotes worth keeping

The lines I wrote down.

GOG Galaxy was a novel thing to do, a hard thing to do, and a wrong thing to do.

We hire smart people, pay them a ton of money, then treat them like task executors. It doesn't add up.

If you focus on delivery, it's like optimizing a cart that gives you stones instead of gold.

Rapid fire round

Same questions every guest. Quick questions, quick answers.

Best advice you've ever received? Listen to what you want, not what other people want. Life got so much easier once that clicked.

Advice you ignored and wish you'd listened to? Stop being a perfectionist. He pours too much time into making things perfect, and even products can't be, so he wishes he'd learned it earlier.

What would you tell your younger self? Freedom is achievable quite easily, and you don't need as much money in the bank as you think. He knows plenty of wealthy people who don't feel free, and he now finds joy in small things, a coffee, McDonald's fries, a walk in the pouring rain.

Ongoing challenge that keeps you up at night? Maintaining focus across five to ten client accounts a day, and removing the tax of context-switching, which is why he leans on pen-and-paper planning.

Favorite spot? Coner Grill, a steakhouse in Warsaw's Praga district, one of the best and most consistent steaks he's had.

Tool you can't live without? Pen and paper first. Digitally, Tana, which he calls "Notion on steroids," a graph database of atomic notes that became his notes, calendar, and tasks.

---

Piotr Kacala is a product and engineering leader with nearly two decades of experience, a board advisor and interim CTPO, and the founder of Product Cards. He spent seven years leading engineering at GOG.com within the CD Projekt ecosystem, was CTO at Displate, and runs the AI Product Heroes cohort course. Find him on LinkedIn.