Over the last four years I moved into founder and leadership roles. That forced me to make a lot of tactical mistakes that might help others. One thing I have learned is that you rarely internalize a lesson until you live it.
TL;DR of what matters most when building:
- Learn that you can learn anything.
- Learn to hire people who are better than you.
- Operate with a high sense of urgency.
There are also more action-oriented learnings. This is a living list. I will keep iterating and adding as I remember more and learn new things.
As a Founder
First-principles thinking. Learning how to learn and how to think critically are two of the most important skills you will develop. In Silicon Valley, “first principles” got popular through Elon Musk. It is essentially the Socratic method applied to problems. Question what you take as true and separate assumptions from facts.
Questions that help me:
- Am I reasoning by analogy, or do I have evidence and data that this is true?
- Is there a real constraint here, or is this a limiting belief that I created?
- What would make this 10x better or faster, and what would have to be true for that to happen?
This mindset avoids cargo-cult decisions. If your answer is “because others do it,” push deeper until you find causal reasons.
Hiring and Culture
Delegate outcomes, not tasks.
Delegating tasks assumes you already know the exact path to the result. Often you do not. If your goal is to 2x qualified leads, delegating “write 10 blog posts” limits the solution space. Delegating the responsibility to 2x leads forces the owner to design and run experiments you may not have considered. Tasks are fine for freelancers. Owners need outcomes and accountability.
Fire faster.
Many founders carry people they no longer believe in because they fear the uncomfortable conversation. Ask yourself: if this person resigned tomorrow, would I feel relieved or worried about the gap? If the honest answer is relief, you already decided.
Do not try to convert people.
Trying to turn a low-drive person into a high-drive one is a time sink. The best hires do not need to be educated by you. They end up educating you and surprising you.
Be brutally honest with candidates.
Startups are chaotic. Roles change, hours are long, pressure is constant. Try to talk candidates out of joining. Spell out the hard parts: changing scopes, long hours, shifting priorities, high pressure. If they still want in, you likely found the right person.
Recognize effort and wins.
Be demanding and appreciative. Lack of recognition burns teams out. I often hear that what people miss, more than salary or scope, is someone explicitly acknowledging results. Do this in public when possible and be specific about what was great.
Sales
Everything starts with a clear ICP.
The most common early problem I see is not being able to answer these clearly:
- Who exactly has the problem we solve?
- What problem do we solve for them?
- How do we communicate the benefit in their words, not ours?
These sound obvious. They are not. Most early teams cannot answer them crisply.
Mental limits block growth.
This may sound soft, but it is real. Most teams are held back by fear and shame. I feel it too. I fear cold calls, cold WhatsApp messages, and asking for referrals. The real issue is fear of hearing “no.” Fewer “no’s” per day means fewer attempts per day, which compounds weekly, monthly, and yearly. Often strategies do not fail. Execution intensity is too low.
Do not evangelize the wrong buyers.
Especially in B2B it is common to hide behind “long processes.” Sometimes that is true. Often you are trying to sell something the other side does not need.
Remember:
- If you need to educate the market, the market will end up educating you.
- Sales is not about convincing. It is about prospecting, qualifying, and disqualifying quickly.
Give permission to say no.
In some cultures people avoid saying no. Make it easy:
“Totally okay if this is not a fit right now. If that is the case, please let me know and we can reconnect later.”
This either frees your pipeline or surfaces the real blocker.
Growth
Define a success metric.
If you run a subscription product and cannot predict who will stay, you are guessing. Define a behavioral threshold that correlates with retention.
Example: two people join Netflix. One watches one movie and never returns. The other watches daily. Who is more likely to renew? Obvious. Your job is to discover the activation threshold and design the product to push users past it. With more data you can set a rule like: “If the user watches three movies per month there is a 90 percent chance they renew.” Then optimize onboarding, notifications, and recommendations to get users to that threshold.
Fire customers.
Losing revenue hurts, but firing the wrong customers improves long-term retention and frees focus for the right ones. The short-term dip buys long-term compounding.
The best growth hack is obvious value.
Build something so good it grows on its own. Aim for a referral coefficient above 1. Every new user should help bring another. If that is not happening, revisit ICP and the core value.
Pricing
Founders obsess about pricing only at the beginning because we do not know what to charge. Then we set a number and forget it. That is a trap.
As your product improves, the value delivered increases, so you should capture more of it. Experiment with packages, anchors, and value metrics. Move prices up until most qualified non-buyers are rejecting primarily on price rather than fit. Your goal is to find the ceiling for your true ICP, not to please everyone.
Remember that value is perception. It varies by buyer and is usually higher than you think. Back price changes with better onboarding, stronger value communication, and clear ROI stories.
Marketing
Share the lessons you wish you had years ago.
This is the best type of content for LinkedIn or Twitter. Teach your past self.
Do not put external links in the main post body.
Platforms optimize for time on site. External links reduce reach. Place links in the first comment or use a light call to action without a clickable link when reach matters.
Awareness and engagement campaigns on Meta rarely paid off for me.
Leaving this open to my own error, but I have not seen good returns there. Your mileage may vary.
For conversion campaigns on Meta, add positive friction.
For tickets of 200 dollars or more, especially in Latin America, impulse buys are rare. Use forms to qualify leads. If you optimize for cheap form fills, the pixel learns to bring you more low-quality leads. Positive friction improves both lead quality and pixel learning.
Creative is king.
Meta’s ML is strong and aligned with your success. The 80/20 is the ad itself. A good ad self-segments. If someone sees a protein supplement ad and never trains, they ignore it. A gym-goer clicks. Most failed ads do not clearly communicate the benefit, the problem, and for whom. That is an ICP problem.
LinkedIn ads are usually a bad use of money.
I have not personally seen strong results. I have heard some positive anecdotes in the United States, but I cannot confirm.
Engineering
Write a thoughtful PRD before coding.
Clarify scope, process, constraints, risks, and success criteria. Many delays and mistakes can be avoided with a solid PRD reviewed by peers. Treat review feedback seriously. It is cheaper to iterate in docs than in code.
Learn CS fundamentals.
This is what separates a software developer from a software engineer. One executes. The other understands the why. Knowing algorithmic complexity prevents inefficient systems and saves real money at scale.
Avoid strict labels.
There is no such thing as only front end or only back end. With fundamentals you can learn what you need and take any role. If you must specialize, do it after you can reason about the full stack.
Final note
These are snapshots, not commandments. I will keep adding to this list as I remember more and as I learn new things in the field.