Experimentation and optionality
Cheap trial and error beats confident top down planning.
Plan carefully in advance, then execute the plan.
Run cheap experiments and let the market decide. Most winning ideas couldn't have been predicted in advance.
Sutherland's case against planning is that it's built on a false premise: that success can and should be predicted in advance. Most of the genuinely successful ideas in his corpus, like the pumpkin spice latte, tested poorly on paper and only worked because someone tried it anyway and let the market decide. He treats the cost of testing as effectively trivial compared to the cost of being wrong with confidence, and argues that businesses should optimise for reducing the risk of catastrophic downside rather than for predicting a single optimal outcome, since "all big data comes from the same place: the past."
He's candid that this makes the early stages of any genuinely new idea look like failure (a sigmoid adoption curve is, by definition, flat and unconvincing right up until it isn't), and that most people only retrospectively construct the confident sounding logic for a decision that was really a hunch followed by a lucky break.
'You can't know in advance' is also a convenient alibi for anyone whose bet didn't pay off, or for organisations that want to avoid being held accountable for a plan at all. Optionality without some red line for failure can just as easily describe recklessness dressed up as philosophy.
300 verified insights in this theme
300 verified insights in this theme
On the false hindsight-bias picture of how new technologies actually get adopted.
On why marketing returns are fat-tailed, like venture capital or film.
On advertising's reluctance to sell fame itself as a benefit, despite it being one of the most valuable things it can deliver.
On why publicity generates opportunities a business could never have deliberately planned for.
On serendipity as the real engine of both scientific and creative discovery.
Closing thought of the talk, on the danger of over-optimizing away all randomness from a business.
Continuing the repeat-use argument, framing the key adoption question as whether people who try something ever go back.
Giving a personal example of innovation adoption from his own life.
Arguing agencies should proactively create work rather than always waiting on commissioned client briefs.
Explaining why some creators (Picasso, the Beatles) can make work unprompted, unlike costlier crafts like Fabergé eggs or watches.
Proposing that Cannes-style award shows should become a marketplace where clients license existing creative work.
Arguing the market for creativity could be ten times larger since the real constraint is cost, not the availability of creative ideas.
On why the best clients are often seen internally as risk-takers precisely because they let upside opportunity outweigh downside risk.
Distinguishing reputationally-costly small failures from financially catastrophic large-scale ones.
Citing Jeff Bezos's 'two-way door' concept to argue for reversible, low-risk experimentation.
Rory's own elaboration of the 'fat tailed' idea Taleb introduced to him, in his own words.
Arguing experimentation and paying for things reveal what people actually want, unlike stated preferences.
His closing 'moonshot goal' answer: scaling up psychological solutions to major world problems like polarization.
Rory on unpredictable restaurant-location success, before the graveyard-site anecdote.
Rory on the benefit of staying founder-led rather than financially controlled.
Rory on the freedom to experiment that comes with ownership.
Rory on progress happening backwards, citing penicillin and Viagra.
Rory's example of progress happening backwards from observation.
Rory recounting advice he gave a client agonizing over commercial innovation.
On why organizations over-plan instead of experimenting.
On the surprising history and initial resistance to Amazon Prime.
On Amazon Prime's paid-membership mechanic as an underexplored innovation.
Describing a subscription-refund idea he proposed to Kagi.
Rory quoting his own advice to Kagi about instituting a generous refund policy early.
On why marketing and R&D should be budgeted and evaluated identically.
Contrasting incremental marketing gains with rare, game-changing breakthroughs.
Contrasting marketing's "less wrong" logic with the single-right-answer mindset of exam maths.
On the strategic value of trying ideas competitors are too serious to attempt.
Rory continuing the Newton/Darwin contrast, on how to approach complexity.
Rory continuing the Newton/Darwin contrast, on how to approach complexity.
Rory on entrepreneurial and professional bravery.
Rory on the underrated business education value of hospitality jobs.
Rory on how growing up around a family business shapes financial intuition.
Rory's proposal that young people should hold back a financial reserve as a shock absorber.
Rory on creativity, citing what he learned from mathematicians.
Rory recounting the Berkeley study of creative versus workaday architects (via John Cleese).
Rory's bee-hive/waggle-dance analogy about efficiency versus exploration.
Rory's bee-hive analogy, continuing the point about the need for exploratory slack.
Rory on the role of self-belief and arrogance in entrepreneurial success.
Rory citing Nassim Taleb's point about designing for variance rather than the average.
On why unexpected, surprising information (gossip, anecdote) carries evolutionary and business value.
On why entrepreneurs favor counterintuitive theories, using the bacon-price economics example.
Continuing on why contrarian, counterintuitive positions have business value.
Responding to Srdjan's mention of Charlie Munger's 'less stupid.'
On using revealed preference (spending) rather than stated preference in transport policy.
On the Heathrow pod-parking anomaly as evidence of unmeasured psychological value in travel.
Advising a client on how to approach commercial innovation without a fixed process.
Describing Nassim Taleb's barbell strategy for investment.
Introducing Brian Klaas's book Fluke and the role of contingency in life outcomes.
Reflecting on survivorship bias and luck in business narratives, following the Fluke/Mick Jagger/Archduke Ferdinand anecdotes.
Continuing on why reasoning-first approaches to problems can only reach a fraction of the possible solution space, versus intuitive or associative leaps.
Explaining why hiring in batches (as Ogilvy did when it hired him) permits riskier choices than hiring one at a time.
Explaining why roughly 20% of bees ignore the waggle dance and forage at random.
Drawing the explore/exploit distinction from the bee waggle-dance example to human foraging.
Note: likely an ASR slip for 'love exploit... hate explore' given the surrounding contrast — flagged for human review before verification.
Introducing the Asda vs TK Maxx analogy for thin-tailed 'exploit' versus fat-tailed 'explore' behavior.
Continuing the TK Maxx analogy for fat-tailed exploration.
Closing the Asda/TK Maxx analogy on why relying only on safe, predictable choices forecloses upside.
Summarizing the dual explore/exploit mandate immediately after the Asda/TK Maxx analogy.
Applying the explore/exploit distinction to two contrasting uses of AI.
Contrasting the Austrian-school view of a business as a discovery mechanism with the standard "efficiency mechanism" view.
Comparing marketing's payoff structure to the pharmaceutical and Hollywood industries.
Describing the asymmetric way marketing's costs versus benefits get evaluated inside businesses.
Applying the loose-fitness-function idea to warn against overly narrow CFO-driven metrics.
On why customer service autonomy generates lucky discoveries that scripts foreclose.
Introducing the explore/exploit trade-off as the core dual mandate for a great business.
Defining the 'exploit' side of the explore/exploit trade-off in business resource allocation.
Warning about the cost of over-exploiting known strategies without exploring new ones.
Listing the costs of a business that never explores beyond what already works.
Arguing that electric cars, freed from mechanical reliability constraints, should unlock more design variety.
Explaining why he's relaxed about others copying his ideas (e.g. 'flat white or eff off' coffee concept).
Framing capitalism, per Austrian economics, as discovery rather than optimization.
Comparing public speaking improvement to a comedian refining a set through repeated performance.
Closing reflection on how the pursuit of certainty forecloses fortunate accidents.
Cold open, on the freedom of self-employment.
On staging reconstructions/press conferences to acquire new data.
On the pace of adoption of significant innovations.
Concluding the Google Glass / marketing patience discussion.
On the sigmoid adoption curve and misjudging early-stage marketing effectiveness.
On critical mass and long-term survival as strategy.
Continuing the evolutionary-biology-as-business-strategy point.
On business as an exploratory, value-discovery process rather than a cost center.
On working in a shop/cafe/restaurant as a 'free MBA'.
On the educational value of frontline work.
Recounting Adam Smith's story of the lazy boy who rigged a steam engine flap with string, and Smith's broader point that innovation often comes from workers themselves.
Contrasting hypothesis-driven experimentation with relying on pre-existing data alone to make decisions.
Arguing marketing returns resemble venture investing or A&R, not a linear cost-to-value ratio.
Arguing that businesses over-invest in cost-cutting because it's measurable, while value creation is deferred and harder to quantify.
Closing argument that institutions over-optimize for measurable efficiency at the expense of undiscovered opportunity.
On why electric-vehicle and video-conferencing technology should be judged by future potential, not current performance.
Concluding the anecdote about a tech company calling its own direct mail 'annoyingly effective'.
On the rich funding early adoption of new consumer technology through status spending.
On why markets need trial and error rather than top-down design.
On evolution as the ultimate trial-and-error process.
Concluding the trial-and-error/evolution analogy.