Notes on Charles Fisher's "The Relentless Startup"

The mechanics of building a VC-funded startup

Charles Fisher co-founded Unlearn.AI, a deeptech startup building digital twins for clinical trials. He raised $135M and left the company in 2024 after eight years.

He’s written this great book called The Relentless Startup on how to build a VC-funded startup. It’s very boots-on-the-ground and has good insight if you’re new to the startup game.

It also has lots of clear prompts at the end to help you understand your leadership archetype, business vision & strategy, etc. This isn’t sponsored but I’d buy the book for that section alone. In any case, here are my condensed notes from the book.

What you should know as a founder

• Startup = uncertainty + ambition. Speed is your USP.
• Founder aim: test hypotheses about the business as quickly as possible.
• For VC funding, a big vision is necessary. You shouldn’t be able to write it down to every detail - that’s a feature of a big vision, not a bug.
  • A big vision also allows you to pivot and helps you hire a strong team, despite them being paid less than big tech or finance, because they believe in the company’s mission.
  • Even though you can’t write it down, you need to tell the story well: 10s soundbite, 1m elevator pitch, 10m story.
• Build a company to your strengths and optimise it for rapid decision-making.
• Six traits as a founder to set the tone: visionary, passionate, decisive, hands on, set the bar high, integrity.

Leadership and culture

• Culture = who you hire, fire, and promote. Culture = performance; culture != fun.
• Cultural principles should not work for everyone. That’s the point. You want strong filters.
• For your skills/traits, create a 2x2 table of: avoid entirely (you’re relatively bad at that others are also relatively bad at), focus (you’re relatively good and others relatively bad at), delegate (you’re relatively bad and others are relatively good), delegate/collaborate (you’re relatively good and others are relatively good). This will tell you who you need to hire.
• Founder mode for areas of comparative advantage, manager mode outside them. I.e., hands-on for strongest areas, delegate your weakest areas. 2 days/week for your strengths/individual contributions (IC), never cave to pressure from board or employees to delegate your areas of strength.
• Strategy is a set of principles, not plans.
  • Strategy is set of guiding principles, a plan is a set of instructions. Can’t plan more than 1-2 steps ahead but you/employee can problem solve using principles to figure out right approach.
  • To come up with your business strategy, start from the vision and work backwards to come up with a set of 3-5 core principles.
• Hiring and firing
  • Want diversity of skills, complementary personalities, strongly-aligned values. And ideally, strongly-aligned aesthetics too.
    • All companies have aesthetics - Apple’s is obvious but Microsoft also has one, it’s just more utilitarian.
  • Hire for attitude and teach skills. High agency, high competency, humble grafters.
  • Hire A-players and treat them well.
    • An A-player = an A-player relative to you.
    • High bar to entry: they self-assess for company fit then apply, then multi-step interview with culture fit assessment (resume-blinded) and 1.5h skills check (presentation on relevant technical topic + Q&A from team), then 2-3 great references.
    • Give them high compensation relative to industry (make use of equity-based compensation too).
    • Be as transparent as you can (e.g. patents, KPIs, etc.).
  • Fire with empathy: 90d decision, above market severance to those terminated + offer recruiting services + offer recommendation letter.

Management and decision-making

• Talk directly to ICs of company all the time.
• If it goes well, it’s on the team. If it goes badly, it’s on you. CEO is the most stressful job. It’s not fun. That’s why you’re paid well.
• Updates and solving day-to-day blockers via Teams message/emails, not meetings.
• Meetings for: sensitive issues, brainstorms, make a decision, feedback/user testing, inform or rally the troops.
• Productivity of a team begins to decrease beyond 4 people.
• Decision-making
  • Success running a company = make difficult decisions quickly + stick with them long enough to see if right or wrong.
  • CEO makes all hard decisions, but not all decisions. Keep emotions in check when making decisions and don’t second guess once you’ve made it.
  • Template: classify as reversible (one-way door or two-way door), make pros and cons, set deadline to make the call.
• Prioritise ruthlessly
  • Sequential not parallel. This is an extreme, continuous form of decision-making. Make extra effort to keep your to-do list under control.
  • Use the Metropolis algorithm for simple continuous experimentation: (1) guesstimate and measure, (2) adjust one variable and measure, (3) if better then carry on, else go back and adjust the same variable another way or change the variable you’re adjusting.
  • Company has one annual objective, team has one quarterly objective, each person has one top priority task at any given moment.
  • Inability to prioritise repeatedly means getting fired.
• Building the product
  • Differentiation is necessary.
  • The first version always sucks.
  • MVP for deeptech can be long (2-3y). Stay persistent.

Raising money from investors

• Most entrepreneurs should not raise VC. Best way to build a company is raise as little money as possible. Bootstrap then VC when scaling if needed. Exception is deeptech which needs a lot of money to just get started.
• Term sheet: should be around 1 page, run away if it’s much longer than that.
• VC firm jobs: analyst < associate < principal < vice president < partner < GP. Also venture partners, operational partners, and entrepreneurs-in-residence (EIRs).
• Fundraising tips: (1) raise when hit some inflection point, (2) build relationships with investors long before you need to raise money (always talking to them over apps, in-person drinks), (3) only raise from investors who are committed.
• Docs needed
  • Non-confidential 15-minute pitch deck: story, vision, market opportunity, business model, team, no sensitive/financial info.
  • Confidential pitch deck: needs NDA, has info on your proprietary technology, customers, financials. Don’t share before second meeting.
  • Financial model for next few years.

Interacting with your board

• Board seats: typically 3-7 members, usually includes founders, investors, independent directors. Typically meet quarterly.
• Will struggle with board as first-time founder because: power dynamics + they know less than you think they should (they spend 10h/quarter, you spend 10h/day).
• Build relationships with them.
  • Try to get dinner the night before.
  • Also 1-on-1s outside of this main meeting + informally message over apps frequently + coffee/drinks occasionally.
• Meet quarterly for 3h meeting.
  • Don’t surprise them with anything.
  • Pre-read: give them a (non-excessive) pre-read for the meeting. Ignore their complaints about having to read it.
  • First half: start with a 30m presentation on what company does with last 5m on most important thing for company right now. Then have people from company present to board on interesting work they’re (or their team is) doing.
  • Second half: voting on resolutions (grants of stock options to new employees and approval of minutes from previous board meeting), then questions you asked board for input on in your memo. Members will ask you to step outside for a few mins to discuss your performance as CEO.
• Remember: you don’t have to listen to them, and don’t let them disrespect you.