Culture and leadership

Sana ran this principle years before generative AI, which meant AI tools slotted straight into an existing discipline. Elf adds a step most skip: before asking whether to automate a task, ask whether it can simply be eliminated.

Going hybrid first, embedding AI across all teams, fails for lack of momentum. Keeping the acquired Uizard team semi-separate proved harder to unwind later; a tight public launch deadline (June acquisition, October conference) forced fast integration.

Shiftmove kept tools separate in deal one, merged on day one in deal two, and everything ran smoother. Visible signals help people live the new direction instead of second-guessing for six months.

Goeldi says you can only apply AI effectively to your own company when you have this deep, tangible understanding from doing it yourself, not from reading about it. Even his light bulbs: Claude Code read the API docs, scanned his network, and told him which button to press.

The test is whether time moved the needle on quarterly goals or just filled the day with dopamine-hit admin. Items sitting at the bottom of the list for two weeks usually mean someone in the company is waiting on you.

The two things a growing company needs are a technical team that understands the industry and accessible potential customers to test and sell to; optimize location for those two, and remember happy people are harder workers.

Over the holidays Modiano tested Lovable, Cursor, and ChatGPT canvas with identical projects, comparing raw-prompt output against reasoned-through PRD output. The same discipline, formalized as applied evals, is what she urges companies to run before betting on a tool or architecture.

Chong looks for experimental cultures and systems thinkers: how product gets built, how team input reaches decisions, whether the decision system is clear. Decisions made behind closed doors by an exclusive group are her red flag; founders can run the same audit on their own org.

The lesson Melanie heard most consistently across a year of guests. Founders get invited everywhere; the test is whether the activity serves company growth, and the same filter applies to product scope decided in customer conversations.

Startups are chaotic and cannot be prepared for directly, but leadership can be trained where the worst outcome is a failed outing or a burned hand. Practice leading unexpected situations long before real capital and careers are at stake.

Knecht estimates 95 percent of what teams do is objectively suboptimal, yet people refuse to revisit decisions because they overestimate the cost of change. Swift fine-tuned adjustments are cheap, and customers respect a team that takes a day to make a better decision.

Founders who hide problems until the cash is gone get the question: why didn't you tell me six months ago when the damage was smaller? Honest, regular updates including challenges are the core of a working angel collaboration in the first two years.

At a prior scale-up she bought the automation tool herself, implemented with the teams' resources under an EBITDA-positive rule, and went use-case shopping. First wins were mundane (someone in Italy stopped downloading and uploading documents and went home an hour earlier), but an HR employee who started there now runs process automation at a listed company. Contrast case: a top-down SAP and IFRS rollout failed because she underestimated change management for local accountants.

Instead of preaching speed, leadership should redesign default rules. Examples: justify not sharing data rather than sharing it, and answer why AI cannot do a job before requesting an additional hire. System shifts beat individual exhortations.

The move is from 'this is your fault' to 'you have goals, I have goals, we have limited resources, let's prioritize'. Founders must lead this culturally and repeatedly; scoping scary changes as reversible three-month experiments (two-way doors) helps teams accept focus and its cost of closed options.

Wildfire revealed the Google deal via a silent slideshow of the company's journey ending with the two logos merging, to a standing ovation (and a missed quarterly sales goal from the champagne). Google's hands-off integration then gave the team breathing room to define its own place, which Victoria in retrospect valued.

The trap is binary: pretending the pitch-deck vision is certain, or flailing and hoping a senior hire knows everything. Instead list what you know, list the specific unknowns (e.g. how to stack-rank client pain points), and give the team freedom to experiment within that frame.

Scale-ups have already prioritized to the top 10%, which makes cutting further feel impossible, but Waelchli says you almost always get rewarded for it. Accept deliberate debt: build the hacked version now, fix the rest in 24 months.

Gunzenhauser deposited the balance sheet at court only after communications were handled. He prepared the team speech diligently because in such moments people look to the leader for assurance that they are not left behind. Partners heard it from him directly before the press.

In Frontify's Series B, funds presented in front of about 100 people on site plus remote staff. Employees noticed culture-relevant subtleties (a partner sidelining his associate) that the leadership factored into the decision, and investors must answer honestly because promises are made in front of the whole company.

Advice relayed via Myke Naef, originally from Marissa Mayer. Applied to hiring ties, it changes the process afterwards: the person who chose invests heavily in making the hire succeed. Most decisions fail not because they were wrong but because nobody invested enough in making them right.

Engage board members between meetings too, sharing highlights, challenges, and strategic decisions to keep alignment.

Oviva's product roadmap committees were very broken for months until the principle re-empowered people with ownership instead of escalating every call upward. He counts this as culture, not process, and it let the company scale much further.

Framing the meeting changes what people bring to it. A physics PhD and self-described strong thinker type, he deliberately evolved toward feeling the room as scale shifted his role from doing sales and ordering toilet paper to leadership.

Quarterly planning runs jointly with product managers as the voice of the customer. Precise phrasing builds the clarity about why work is happening, and keeps teams from disappearing into proud scaling projects that serve no business goal.

She wanted maximum certainty the original idea was dead, having still believed the problem was their sales execution. Parallelism meant that when the old model was pronounced dead, three vetted new directions were ready the same day, converting existential doubt into energy.

Decide upfront and announce whether everyone keeps their job, then give employees the same free choice about the new direction. Pile lost some people who wanted crypto careers, which is fair; transparency about the process is what protects trust.

Unlike older generations used to executing orders, Gen Z team members want to see the impact they can have on the mission.

Her stated goal: it is acceptable to lose people who joined for the old mission, unacceptable to lose anyone because they did not understand why the change was happening. Naming the personal loss takes the steam out and turns the meeting constructive; leadership alignment took three to four weeks of frequent long conversations.

With a million small problems on the desk, the temptation is to handle the big pivot when there is time. Start jotting notes and diving into the complexity right away; in hindsight she could have taken the decision earlier, and coming into leadership conversations grounded in the why was the precondition for success.

Read the comments honestly: sometimes the verdict is we really did not nail that, sometimes it is sorry, we are a scale-up and this will not change. Distance between the management team and employees grows with every level added.

Also ask the strategic question before doing any deep operational work: does this task need to exist, and is teaching or empowering the team higher leverage than doing it yourself.

Most companies add controls as they grow; On's management deliberately did the opposite at around 1,000 people to over-index on trust and responsibility, betting that the culture of autonomy was worth more than control.

Putting values on the intranet or naming meeting rooms after them changes nothing. On relies on employees sharing how a spirit came to life years ago, and with roughly 100 new joiners a month, the founders and co-CEOs stay personally involved in onboarding so new people become culture carriers.

Be a helping hand but not the fallback option when the road gets bumpy; patience and tolerated failure are part of the design.

Telling people what to do while keeping the responsibility trains them to return at every barrier. Delegating ownership means they are entitled and trained to take decisions, connect dots, and find resources themselves, which frees your headspace and grows them into independent small-scale leaders.

Jumping in communicates 'you don't know how to do this, I'm better' and makes the person less likely to take ownership next time. Explain why you did it and what it means for them; also start small when building delegation, chopping projects into low-visibility pieces like building a muscle.

Never give feedback on what you think or heard happened; only observable behavior gives you a basis for conversation. Then anchor the change to a driver that matters to them, like a promotion in six months, so they leave motivated rather than lectured.

Nobody walks into the boss's office with a feedback list unprompted; make it a prerequisite with specific questions. When they finally speak up, that is your money time; if you argue and justify, they will never come again.

Saanya Ojha contrasts two board meeting styles: founders trying to ace the test with exhaustive decks versus founders handing out the test by directing the board to where help is needed.

Critics argue it only works with near-infinite capital; defenders call it a decision-making framework for the AI age.

Enzo Waelchli works from a co-working space on those days; they streamline priorities and protect critical thinking when everything else is reactive.

Sustaining OKRs is a culture, not a task.

The 'make the boat go faster' concept: focus on one thing and everything else must power that push.

Following Charlie Munger: get incentives wrong and everything else breaks.

Find coaches through founder referrals or your VC, distrust miracle-promising marketing, and in the intro session bring your hardest material: a great coach leaves you deeply seen and rocked out of your comfort zone.

Victoria Ransom names this her most operational leadership mistake, subtle enough that it rarely makes it into interviews.

Excuses and apologies are useless; the goal is a shared view of where breakdowns occurred and unchecked assumptions surfaced up front. Trusting this process over solving it yourself is the core mindset shift.

Also: speak up when you do not understand something, challenge assumptions, and work with people smarter than you.

Progress often comes from subtraction: fewer products, fewer metrics, fewer narratives. Clarity compounds faster than cleverness.

Hanan's rule from running multiple World Cup campaigns: always be ready for chaos, because nobody prepares you for a shock like that. Structured processes bring efficiency in a perfect world, but geopolitical and client uncertainty mean flexibility in staffing and problem-solving is the real prerequisite.

Hanan's test before any AI spend: define what you actually need the technology for, and check whether the organization is ready to adapt processes. AI in marketing predates ChatGPT and Gemini by years; the philosophy of using technology to get smarter and more efficient has not changed, only accelerated.

Elf reaches for it beyond work too: life will not do it for you. The principle frames excellence as active daily resistance to decay rather than a state you achieve once. She also names the moment a startup stops feeling like one: the day you walk into an office and no longer know everybody's name.

Leaving Google on Friday and joining Exa on Monday felt like moving several years into the future: onboarding questions, database queries, report writing and bug fixes all run through one AI brain wired into internal systems.

Chong stresses the uncanny ability to simplify what feels complex and undefined into a direction, and names storytelling charisma as the most underestimated trait, since you cannot recruit people to a journey no one else sees yet without it.

Company values, ways of approaching problems and vocabulary stay common, while local variation is embraced. Rituals reinforce the common layer: the Friday demo (30 minutes, under one minute per demo, no slides, show real work) and the Grande Roue, flying the whole company to a secret European location for a week of team building and on-stage commitments.

Perry Reifer's test for any culture doc or list of behaviors: could a different company adopt it unchanged? If yes, it fails. Making culture stick requires talking about it every day in every situation, not once or twice, and that takes deliberate intentionality from leaders.

Wildfire ran on an intuitive culture for years and only later wrote down its values: scrappy, humble, roll-up-your-sleeves, speak up ('the next best idea might come from the newest person in customer service'). Victoria: values are a direct embodiment of the founders.

Jake has watched companies go from a billion on paper to selling for cents on the dollar via acquisitions made to paper over growth promises. Outcomes like 'a billion-dollar exit' are not intents; without knowing what you actually want to build, founders spread thin across options out of fear of failure, copy org charts, install OKR rituals nobody reads, and hire big-company saviors.

In doing mode everything lives as messy intuition in the founder's head. Clarity means answering: what are we building, what does this era require, what is the environment, and what do I need from each function and person. Once that exists, saying it is trivial.

A team of three to fifteen in one room is aligned by default, so OKRs add little. Pre-PMF the only objective is product-market fit itself, with key results drawn from your traction metrics. The bigger and faster-scaling the company, the more OKRs pay off.

Values on a poster do not change behavior. If a large company rebrands toward values its employees do not hold, the identity only becomes true by exchanging the people, which is a different beast from documenting who you already are.

Stat cited in the episode's framing; Nicole Herzog has sat on all sides of the table as founder, investor, and board chair at Sherpany.

Oviva scaled to over 800,000 patients treated across Europe on 120M dollars raised while working to keep culture, trust, and speed intact. Processes should stay minimal and adaptable, empowering teams rather than creating bureaucracy.

She had optimized herself for being an effective leader of a 100-plus-person company, not for executing fast or doing her own accounting. The other honest bootstrapping hurdle is personal: her agency-first model was a calculated path so she could pay herself a salary.

At that size you must invest in documentation and order while lacking the manpower to keep it from taxing productivity; a venture-backed company must keep showing growth throughout. Past that point the burden becomes manageable again.

Regular updates, feedback collection, and open communication are what make employees carry the new brand; change management is a long-term process, not a one-time project.

The team was simultaneously absorbing 1,400-plus new customers a day and building a new warehouse in Pratteln within 100 days. Feedback stayed open until the last moment via Confluence and Jira, but the deadline prevented endless opinion loops; with more time, a million more questions would have come.

Le Shop's self-image was a French-speaking logistics startup; overnight it became a technology subsidiary managed from Zurich, with people fearing lost autonomy. Her two do-differently lessons: invest much more in face-to-face change management and start planning the whole project far earlier.

It underpins both internal alignment and external comms, and technical founders can learn it.

Confronting these honestly and early replaces any formal vulnerability-assessment exercise.

Communications is part of a founder's job every day and forever, going far beyond press releases into talent, investment, and customers. Founders must constantly flex between their own conviction and the perspectives of customers, investors, and press, and few feel natural at it.

On grew from 100 to over 3,000 employees through an IPO while deliberately working to keep its entrepreneurial spirit.

Founders play a distinct stewardship role in preserving culture, and starting with why (mission) before what (actions) keeps decisions aligned.

The company hired the person, resourced the idea, and a robot-armed one-step spray-on shoe made it to the Olympics. It shows what a culture that embraces failure enables: if people fear mistakes, they never challenge the status quo or innovate.

The 63 percent figure comes from a Sifted survey; the disengagement figure was cited by leadership coach Stanislaw Grunstein. The symptoms are poor communication, unclear goals, and micromanagement.

The fix is to communicate key success factors explicitly, provide the tools, and then let people paint the picture themselves.

Acquirers avoid disruption at the top right after buying. Gervazio stayed on as CEO because selling Avrios was not the end of her journey, and management alignment was one of the criteria that shaped which offers were viable.

The a16z-backed startup marketed its AI sales bot as superior to human workers; employees reportedly worked holidays, slept in the office, and faced relentless pressure while the CEO defended the intense culture.

Sifted's Freya Pratty reported the trend with Sophie Winwood of Unlock VC warning about its effects; Scaling Europe's Seb Johnson fired back that Sifted's own fundraise coverage features even fewer women founders. Both sides raise uncomfortable truths.

Recounted at Italian Tech Week: founders have an abundance of ideas, but organizations can only process so many; rationing your own ideas is a leadership act.

As Anybotics grew, Enzo Waelchli's role shifted from managing individuals to leading senior leaders; without that foundation even willing leaders fail to scale.

Great operators make their reasoning legible so others can execute independently. Opacity scales badly.

Across 50+ episodes, declining speaking invites, product features, and distractions disguised as growth showed up again and again as a defining trait of effective founders.

From the investor seat, clarity and humility, not loudness, are the signals Lina Chong watches for in founders who successfully evolve with their company.

The resulting organizational drag cannot be fixed by any amount of model improvement.

The other underestimated compounding effect is regained headspace for what actually moves the mission forward.

Most failure comes from underestimating the distraction and execution load after closing, not from overpaying.

He concedes funding would have meant better sleep during COVID and easier international scaling. But investor-backed agency roll-ups in his space have repeatedly struggled with culture integration, and bootstrapping enforced focus because there were never millions to spend on things they did not really need.

Hanan argues the real-time buzz between social, paid and analytics teams reacting to campaign wiggles cannot happen remotely. He contrasts Webrepublic's one-roof Zurich setup of about 150 people with international network agencies that split social, content and other functions across countries.

First priority is unified finance and go-to-market reporting with shared KPI definitions, because you cannot manage what you cannot see. Delaying tough choices out of discomfort is the integration mistake that hurts most.

Candidates withdraw from perfect step-up roles the moment they hear five days in office. Connolly argues you only get crisis projects and coffee-machine context by being there; flexibility is earned later by building systems that work. She simultaneously calls the 996 craze wild and unnecessary.

Public debate focuses on selling a vision, hiring, and technical acumen, but the compounding advantage comes from being able to take constant mistakes, reflect on them, and convert them into learning speed.

Online is a two-dimensional slice of reality: you miss the team dynamics, gestures, and whether you actually want to be in a room with these people for years. He does financials and business-case discussions online but always meets the team before committing.

Companies often reorganize one department at a time to limit disruption. Perry Reifer argues the bold version, changing the whole structure at once, gets you to the future-fit answer faster and better, and that legacy is useful for storytelling but never for building the future.

Her abandoned practice as a people leader is adding processes. Processes come in to monitor where trust is missing; limiting them to the truly necessary automatically raises trust, relationships and leadership to a different level.

Wildfire raised roughly one tenth of its nearest competitors' capital while performing just as well, kept more control, and built a do-more-with-less culture Victoria traces directly to her asparagus-farm upbringing.

The textbook answer is never to link OKRs and bonus, yet if everything else is compensated and OKRs are not, people rationally deprioritize them. His resolution: a 360-style evaluation including OKRs, KPIs, competencies and values, judged in discussion rather than by formula, since stretch targets and external factors make raw percentages meaningless. OKRs also live at team level while evaluation is individual.

Wanting to let go is the entry requirement; then trust gets built through shared culture and process plus expertise that gets paid back. Without it you micromanage, and your good people leave. He also encourages new senior hires to proactively challenge his because-we-always-did-it-this-way assumptions.

yamo employees had access to all metrics and quarterly deep-dive town halls, but cash position and fundraising progress stayed with management. He argues there should already be enough delivery pressure in a startup without everyone fearing for the company's survival.

Hoelzl agrees Europe is harder on employment law and cultural sentiment than the US, yet insists you can embrace European values and still build an outcome-driven company by actively selecting for that spirit.

Only after the cultural base was set did the company disperse via the US acquisition and remote teams. An argument for early colocation in the remote-versus-office debate.

Herzog held that view herself as a founder and only understood after switching sides: an experienced board member does not need to see every tree to judge whether a forest is healthy or sick.

Most advice contains some truth for some company in some situation; the founder's job is the mapping, not the obedience. The corollary he had to unlearn from consulting: hear everyone, tease out the conflict, then own a decision not everyone likes instead of seeking consensus.

Write the principles down, explain them, and apply them consistently across the organization; the cost of exceptions far outweighs their benefits, because one deviation unravels the system. Consistency is also what makes constraints acceptable to ambitious juniors.

Know what you are good and bad at and complete yourself with the team; trust forms when people feel the way you operate is authentic. Being transparent about what bothers you is not showing weakness, it creates commitment to the big goals.

Against the assumption that young founders default to remote, he calls himself a classic on this: flexibility yes, but almost the whole team is Berlin-based, and physical co-presence produced long-term relationships and new ideas.

Scaling back usually means layoffs, and highly qualified people avoid companies with a history of letting people go out of concern for their careers. Brite has never had to cut because cost-consciousness is built into its DNA, even when the team grumbles about renegotiating offers after a big round.

On keeps its five spirits stable but re-asks annually how they must show up in behaviors as the company changes, for example translating office-born values to shift workers in a Shanghai retail store. Culture is never a set-and-forget exercise.

The bet is that empowerment through removed controls beats the losses from occasional abuse, a position many CFOs would push back on.

The point is headspace, not just time: strategy, vision, and developing people are the things nobody else can do for you. He is comfortable stating 100 percent because nobody actually gets there, but even moving in that direction is the win.

Operational feedback loses its value if discussed a week later; higher-level feedback given between meetings is perceived as unimportant. One-on-ones are for feedback and growth, not just tasks and projects.

Lütke's company-wide memo made reflexive AI usage in as many contexts as possible a job requirement, a glimpse of a workplace fundamentally changed by AI.

Jeff Breunsbach argues that what gets celebrated in CS orgs is reactive heroics, which rewards exactly the wrong behavior.

Directly linking compensation to stretch goals backfires by encouraging sandbagging and gaming.

Stef van Grieken disliked Google's thirty-minute standing 1:1s; at Cradle he does 1:1s with his team only every two to four weeks.

Andreas Klinger cites Rivan founder Harvey Hodd, who taught himself to build machines and spends every minute carrying boxes, fixing bolts, and re-iterating priorities with the roadmap on the wall.

Her answer to which advice founders should ignore: 'probably most of it.' If advice feels like a crutch, the real work is building your own judgment.

Kate Connolly: it might apply to your first 10 customers, but not to how you spend your own hours.