The 37% Rule: When Math Says Stop Looking and Decide
You’ve interviewed six candidates for a key role. Three were decent. One was great but you weren’t sure. Now candidate seven walks in and you’re still wondering if someone better is coming.
Sound familiar? You’re stuck in the most common decision trap in business: the fear of commitment when you don’t know if something better exists.
Mathematics solved this problem decades ago. The answer is called the 37% Rule, and it works for hiring, vendor selection, pricing, investment timing, and pretty much any decision where you’re choosing from a sequence of options.
The Secretary Problem, Explained Without the PhD
In the 1960s, mathematicians formalized what they called the Secretary Problem. The setup is simple:
- You have n candidates to interview, one at a time
- After each interview, you must accept or reject immediately — no callbacks
- You want to pick the single best candidate
The optimal strategy? Reject the first 37% of candidates automatically. Don’t hire any of them, no matter how good they seem. Use them purely as your calibration set. Then, hire the next candidate who is better than every single one you’ve seen so far.
This strategy gives you a 37% probability of selecting the absolute best candidate from the entire pool. That might not sound impressive until you consider the alternative: random guessing gives you a 1/n chance. With 10 candidates, that’s 10%. The 37% rule nearly quadruples your odds.
The math behind it converges on 1/e, where e is Euler’s number (≈ 2.718). So 1/e ≈ 0.3679 — roughly 37%.
The Fibonacci Connection That Shouldn’t Exist
Here’s where it gets interesting. The Fibonacci sequence produces a ratio that approaches 0.618 (the golden ratio φ). The complement of that ratio, 1 - 0.618. equals 0.382.
Compare:
- Optimal stopping threshold: 1/e ≈ 0.3679
- Fibonacci retracement level: 0.382
The difference is 0.0141. Less than 1.5 percentage points.
These numbers come from completely different branches of mathematics. One emerges from calculus and probability theory. The other from a recursive sequence a medieval Italian wrote down to model rabbit populations. Yet they land within spitting distance of each other.
Why? Both numbers describe natural transition points - moments where a system shifts from exploration to exploitation. The golden ratio appears in phyllotaxis (leaf arrangements), spiral galaxies, and financial market retracements precisely because it represents an efficient boundary between growth phases. The 1/e threshold appears in optimal stopping because it represents the exact boundary between gathering information and acting on it.
Nature, markets, and mathematics all seem to agree: somewhere around 37-38% is the point where looking becomes procrastinating.
Fibonacci traders already use the 0.382 retracement level as a decision point, a price that has corrected “enough” to signal a re-entry. They’re running the same algorithm as the secretary problem, just in a different domain. Retrace 38% of your gains, and you’ve explored enough of the downside. Time to commit.
How to Actually Use This in Business
Hiring
If you’re interviewing 10 candidates for a position, reject the first 4. Don’t make offers. Don’t get attached. Use them to calibrate what “good” looks like for this role, at this salary, in this market.
Starting with candidate 5, hire the first person who’s better than all four you rejected.
For a cleaning company hiring a crew leader. maybe you get 8 applicants. Reject the first 3. Hire the next one who beats them all. Stop agonizing over whether someone better might apply next week. The math says this is your best shot.
Vendor Selection
Getting quotes from 6 suppliers? Review the first 2 purely for benchmarking. Then pick the next vendor that beats both on your weighted criteria (price, reliability, terms). You’ll waste far less time in procurement cycles.
Pricing Decisions
Testing price points for a new service? Try 5-6 different prices in sequence. The first 2 are data collection. After that, lock in the first price point that outperforms your test set on conversion × margin.
Investment Timing
Looking at 12 months of potential entry points? Let the first 4-5 months pass. When you see a month that offers better conditions than everything prior, deploy capital. This is remarkably similar to dollar-cost averaging with an intelligent trigger - and it maps directly to the 0.382 Fibonacci retracement that technical traders already rely on.
House Hunting
Planning to see 15 homes? Tour the first 6 without making offers. They’re your education. Then bid on the first house that tops everything you’ve seen. Real estate agents hate this approach because it looks like indecision at first. But the data says it maximizes your chance of getting the best home.
When the Threshold Drops: The Cost Adjustment
The 37% rule assumes search is free. In reality, every interview costs time. Every vendor meeting burns hours. Every month of waiting has opportunity cost.
When search costs exceed roughly 5% of the expected value of the decision, the optimal threshold drops to around 25%. You should explore less and commit sooner.
Hiring a $50K/year employee? If the process costs you more than $2,500 per candidate cycle (recruiter fees, lost productivity, interview time), shift from 37% to 25%. Interview 10 people? Reject 2-3, then take the next best.
For high-cost decisions, executive hires, major equipment purchases, facility leases. the math pushes you toward faster commitment. The expensive mistake isn’t picking a B+ option. It’s spending six months looking for the A+ while the B+ goes to your competitor.
Connecting the Framework: Optimal Stopping + Kelly Criterion
In a previous article on applied decision frameworks, I covered how quantitative models outperform gut instinct in operations. The 37% rule fits into a broader decision framework:
- Optimal Stopping (37% Rule): Answers WHEN to commit - the timing question
- Kelly Criterion: Answers HOW MUCH to bet, the sizing question
Together, they form a complete decision system. First, use the 37% rule to determine when you’ve seen enough options. Then, use Kelly sizing to determine how much capital, time, or resources to allocate.
Hiring example: The 37% rule tells you candidate 5 is your pick. Kelly tells you whether to offer them 80% of budget (high confidence) or 60% (hedging against uncertainty).
Vendor example: Optimal stopping says Supplier C beats your benchmarks. Kelly sizing says allocate 70% of volume to them, keep 30% with your backup. Because even the best decision has variance.
This is how the 8-Step Problem-Solving Method works at the quantitative level. structured frameworks that remove emotion from decisions and replace it with math.
The Anti-Agonizing Rule
Most business owners I work with share the same failure mode: they keep searching because they’re afraid of regret. “What if there’s someone better?” “What if prices drop next month?” “What if I see a better property?”
The 37% rule doesn’t eliminate regret. It minimizes it mathematically. No other strategy - no amount of extra interviewing, shopping, or deliberation, produces better results over time.
Here’s the practical version for a 10-person hiring process:
- Interview candidates 1-4. Take notes. Score them. Reject all of them.
- Starting with candidate 5, make an offer to the first person who scores higher than your best from the first group.
- If you reach candidate 10 without finding someone better, hire candidate 10.
That’s it. No hand-wringing. No “let me think about it.” The math already thought about it.
The Bottom Line
The 37% rule is one of the few mathematical results that translates directly into business operations without simplification. Explore 37% of your options, then commit to the next one that beats everything you’ve seen.
When costs are high, drop to 25%. When stakes are extreme, pair it with Kelly sizing. When someone tells you to “trust your gut,” show them the math.
Nature figured out the 0.382 threshold through billions of years of evolution. Markets rediscovered it through millions of trades. You can adopt it in your next hiring decision.
Stop agonizing. Start deciding.