High Conviction Founder Series: They optimize for learning per dollar spent
Big bets disproportionate to the evidence secured – kill startups. Habit #5: buy every lesson at the cheapest experiment you can “buy”.
Hey friends 👋
The wire hits on a Thursday. $500K pre-seed, eleven months of pitching to close it.
Friday you make the list. A sales hire, because founders can’t do sales forever. A $30K booth at the industry’s biggest conference. A $45K agency rebrand, because the current site screams side project. Four months of engineering for the enterprise features every prospect asked about.
Every line item feels like company building.
Six months later the sales rep has quit. The booth produced 212 badge scans and zero buyers. The brand is gorgeous and conversion didn’t move. The enterprise features sit behind two stalled procurement processes, untouched.
$310K gone. Now ask the only question that matters at your stage: what did you learn with the money?
The sales hire couldn’t sell your product. You suspected as much, since you couldn’t either. Everything else on that list, you could have learned for the price of a nice dinner.
Every bet was oversized for the evidence behind it. That gap has a body count. And in this case, likely your startup.
Let’s dive deep 👇
Habit #5 in the series: They optimize for learning per dollar spent
Quick recap for anyone joining late:
Habit #1: high conviction founders test what could kill them.
Habit #2: they fall in love with the customer’s problem, not their own solution.
Habit #3: they think like scientists and act like founders.
Habit #4: they prioritize shortening time-to-customer.
Habit #4 put a clock on your learning. This one puts a price tag on it.
High conviction founders optimize for learning per dollar spent.
Every action a startup takes buys two things at once: a shot at progress (evidence) and a learning. Low conviction founders price the first. High conviction founders price both, and they refuse to pay $180K for a lesson on sale for $300.
Your runway is a research budget
Before product-market fit, your business is a stack of unproven beliefs. Customer, problem, willingness to pay, channel. Beliefs all the way down.
Capital at this stage has one job: convert dollars into evidence until enough of those beliefs become facts that spending on growth stops being a gamble.
Which means the honest unit of runway at pre-seed is the number of experiments remaining, not the number of months remaining.
$400K funding $2K experiments is 200 shots at the truth. The same $400K committed to four big swings is 4 shots. Same bank balance. A 50x difference in what you’ll know when it’s gone.
Size the bet to the evidence
Habit #1 gave you the Ladder of Evidence and the line that the lesson never changes, only the price does. Habit #5 turns that into a spending policy, and we wrote the long version of it in Stop betting the house on a maybe.
The short version comes from poker. Amateurs go big on weak hands because excitement or desperation says to. Professionals size every bet to the strength of the hand in front of them. Your hand is your evidence, and a compliment at a demo is a pair of twos.
Traction is a trendline, and your bets should match the slope. A single data point is a clue. A consistent curve is a case. When the story is still unclear, make small bets. When the signals compound, invest more. When the market starts pulling you in, that’s when you sprint.
In dollars, proportional betting looks like an earned ladder. $500 of coffee and cold calls earns the right to spend $5K on a concierge test. Five customers paying full price for the concierge version earns the right to spend $50K on a real build. Paying customers who stick around earn the right to spend $500K acquiring more of them.
Skip a rung and the difference rides on a guess. A $50K build backed by conversation-level evidence is a $45K gamble wearing a $5K experiment’s clothes. Better to underbuild for a no than to overbuild for a maybe.
The moment this discipline matters most is the week after a raise. Money in the bank creates pressure to make moves that match the new number. Investors expect motion. Your co-founder expects hires. Resist all of it. The wire changed your balance. It did not change your evidence. Bet on the slope, not the spike.
Risky theater
Big actions taken on evidence too thin to warrant them have a name: risky theater.
Risky theater is the expensive cousin of procrastivity from Habit #1. Procrastivity hides in cheap busywork. Risky theater hides in bold moves. Both produce motion the market never gets to falsify. Only one of them torches six figures doing it.
Three founders you will recognize:
The conference founder. They drop $30K on a booth and a sponsorship tier before ever running their pitch past ten cold prospects, which costs nothing. The booth stress-tests their logistics, their banner printer, and their small talk. The message walks in untested and walks out untested, now with a $30K receipt.
The early hirer. They bring on a $140K enterprise sales rep to escape founder-led sales before the motion is predictable. Predictability comes first, then repeatability, then scale. A rep handed an unpredictable motion can only multiply the ambiguity, and in month five everyone agrees to call it a bad fit.
The big builder. They turn two unpaid pilots into an eight-month enterprise roadmap. The evidence on hand supports a $3K concierge test. The spend assumes a validated product. The $147K between those two numbers is pure exposure.
Each bet ran 50x to 100x ahead of the evidence behind it. And each returned close to zero falsifiable information per dollar, which is the tell. Theater spends big and learns small.
The math
Put one hypothesis on trial: independent veterinary clinic owners will pay $350 a month for software that automatically rebooks no-shows.
Four experiments can test it, at four price points.
Experiment one: conversations. Twenty calls with clinic owners, each ending in one of the four asks: TEAM (Time, Effort, Access, Money). Will they put down a deposit, intro you to their office manager, commit to a pilot start date? Cost: about $300 and a week.
Experiment two: a payment link. A landing page, $200 in ads, and a checkout that captures a card for a founding-member rate. Cost: about $700.
Experiment three: concierge. You rebook the no-shows manually for five clinics at full price. No product exists. The clinics don’t care. Cost: about $3K, mostly your time.
Experiment four: build first, ask second. Eight months of engineering, then the same question. Cost: $180K or a large part of the cap table.
All four experiments test the same hypothesis and return the same binary: owners pay or they pass. Each step up buys a little more certainty. The price climbs 600x from the first experiment to the last. The certainty of the experiment doesn’t.
Founders default to experiment four, because the product feels safer to stand behind and the launch reads better on LinkedIn. And when the answer comes back no, they’ve burned the better part of a pre-seed round to hear it.
Now allocate the research budget. Two founders, $300K each. Founder A makes two $150K bets. Founder B spends $60K running experiments one through three against their riskiest beliefs, then puts $240K behind whatever survives. If each big bet rests on even one coin-flip assumption, Founder A clears both flips 25% of the time. Founder B bought dozens of chances to be wrong cheaply and saved the big check for a belief that already survived contact with the market.
Same money. Wildly different information and learning gained.
Build the habit this week
Open your budget. Include the spends you’ve already committed to in your head. Those count.
List every planned spend over $1K for the next 90 days. Price time honestly. A month of engineering is a spend whether or not it shows up in QuickBooks.
Next to each line, write the belief that has to be true for that spend to pay off. Then score your evidence for that belief, 0 to 10. No 7s. 7 is where founders hide from themselves.
Flag every line where the dollars are big and the score is under 6. That’s your risky theater list, and every founder has one.
For each flagged line, write the cheapest experiment that buys the same learning. Use the tenth rule to size it. Run the cheap version first, and let the result size the next bet.
One budget line at a time: spend like a scientist until the evidence lets you spend like a CEO.
Want the worksheet we use with real companies?
Access the Validation Roadmap template → email cam@tractionlab.io
Until next week,
Cam