How smart founders decide which fires to let burn

Executive overview

Fast-scaling companies face more problems than they can solve. Fighting every fire exhausts the team and stalls growth. The skill is intelligent triage: knowing which fires to extinguish now, which to monitor, and which to leave burning entirely.

Neglect is sometimes the strategy. The founders who scale fastest are not the ones who fix everything — they're the ones who commit deliberately to what they won't fix yet.

The core triage logic

• Speed of scaling creates more fires than any team can handle simultaneously
• Letting a fire burn is only valid when the probability and damage of ignoring it are both low
• Assign each fire a probability, a vector (rising or falling?), and an impact if it hits
• If a fire is existential and likely, solve it immediately; if it's cosmetic and low-risk, defer indefinitely
• Trying to fix everything means fixing nothing important fast enough

When to ignore customer complaints

• PayPal grew so fast its three-person customer service team fell 10,000 emails behind per month
• The team turned off desk phone ringers and used cell phones — deliberately ignoring incoming complaints
• Rationale: serving future customers at scale mattered more than serving current ones individually
• The fix came all at once: a 200-person Omaha call centre stood up within two months
• Letting a fire burn only works if you're watching it and have a plan to douse it at the right moment

The Chernobyl test: existential fires must be solved fast

• SurveyMonkey had no database backup — a server failure would have ended the business
• Reid's framework: assign a daily probability to the catastrophic outcome, then calculate cumulative risk
• 0.01% per day is manageable over months; 1% per day means ~15% chance of failure in 30 days
• Once the SurveyMonkey team sized up the real odds, they fixed the backup within days
• Almost every early-stage startup runs without failover — LinkedIn operated for years without a backup database

Deliberate neglect: SurveyMonkey's ugly design

• SurveyMonkey's UI was widely criticised as dated and unattractive for years
• Selina chose not to fix it: the product was functional, loved, and performing
• Easy-to-use and beautiful are not the same thing — when forced to choose, prioritise ease of use
• Amazon's website is deliberately ugly; Bezos locks his team to convenience, price, and speed only
• The post-checkout "big massive heart" thank-you page made everyone cringe — and it stayed, because the customer had already paid
• Ugly design eventually created a secondary fire: recruitment became harder as engineers judged the product on aesthetics

How fires merge — and why that's the real risk

• A fire left burning long enough will merge with another fire
• SurveyMonkey's design debt became a recruitment drag, which slowed the engineering fix of that same design
• The risk is not any single fire — it's the chain reaction when deferred fires compound
• Vigilance means watching for fires that are growing or converging, not just the ones already loud

Communicating deliberate neglect to your team

• When a leader visibly ignores a problem, the team assumes incompetence or indifference
• You must signal clearly: I see it, I'm choosing not to fix it now, here's why
• Selina's Ticketmaster playbook: get everyone on a conference call, ensure full information sharing, install a triage leader
• Communication is the first thing to break down when something goes wrong — stabilise it first
• Patience as a skill: snapping at a team in crisis makes the crisis worse

Seizing growth over perfection: the overcommitment pattern

• Evite's viral growth was discovered by accident when a tripped power cord took the site down and users complained immediately
• iLike launched on Facebook with two servers; within 30 minutes they were doubling capacity over and over, renting U-Hauls to borrow machines from data centres on Memorial Day weekend
• Kuli Kuli said yes to a nationwide Whole Foods launch in January when approached in June — factory problems, snowstorms, and a gritty texture all followed
• The product launched in 435 stores anyway; it now trades in 3,000 stores, quadrupling footprint in a year
• The lesson: saying yes to massive growth and scrambling to deliver beats spending six months perfecting before launch

Advice Selina would give her younger self

• Ship early, get consumer feedback on a subset of users — feedback prevents fires better than planning
• Hire to cover your blind spots; you cannot see every fire coming
• Design was not Selina's strength — her first hire at Gixo was a great designer
• Experience reduces panic, not fires; the fires keep coming, but your reaction time and calm improve

Hiring for fire tolerance

• Recruit people willing to run into a burning building — not just people who can fight fires
• The more useful signal: people who can spot a distant five-alarm blaze while managing the one in front of them
• Let the unfinished parts of your product reveal themselves in interviews — candidates who still want to join are the ones you want
• Complaints from customers and sales reps are a positive metric: people only complain when they care