The myth-busts, hot takes, explainers, and tools worth keeping.
⚡Myth Buster· 2
⚡Myth Buster1:04:30
User imagery is not your target audience
Sutherland untangles a common marketing confusion: the person shown using a product is user imagery, not the actual buyer. Small cars are advertised with 28-year-old female drivers, yet the average new-car buyer is around 54. Show young drivers and older buyers happily buy in; show 65-year-olds and you lose the younger aspiration entirely, so the featured user is chosen strategically, not as a literal customer portrait.
The person featured in an ad is user imagery, not the target audience
New cars are mostly bought by the relatively elderly despite young ad casting
Aspirational young imagery keeps older buyers without alienating the young
Reverse it, and you lose the younger audience entirely
“don't ever get confused that that's actually your target audience”
#marketing#advertising#segmentation#branding
⚡Myth Buster1:36:30
Range anxiety is an imported American fear that doesn't fit the UK
Sutherland argues that electric-car range anxiety is a rational American fear wrongly imported into the UK. America needs about 116,000 petrol stations to serve its geography; the UK has roughly 8,500 because it is small and dense. A broken charger in Britain just means another two miles down the road by a tea shop, not a night stranded in a freezing Idaho truck stop. Ironically, every EV brand advertising its range magnifies a fear that barely applies.
US needs ~116,000 petrol stations for geography; UK has ~8,500
UK density means a failed charger is a minor detour, not a crisis
Home 3kW charging plus trains cover most real UK journeys
Brands competing on range amplify a fear that scarcely applies in Britain
“for most Brits most of the time range anxiety is a perfectly rational American fear which doesn't apply to the UK”
We favour what's quantifiable over what's important
Sutherland argues marketing has become so obsessed with measurable, attributable results that it can no longer do anything it cannot perfectly quantify. Yet many valuable marketing activities are inherently probabilistic: you make enough noise that when someone eventually enters your market, they consider you. He cites the man who bought an Aston Martin because of an advertisement he saw aged twelve, an effect no attribution model could ever capture.
Direct marketing's measurability became an obsession that crowds out the unmeasurable
Much valuable marketing is probabilistic: be known so buyers find you later
Long lags between stimulus and purchase defeat attribution
The Aston Martin bought from an ad seen at age twelve makes the point
“we've disproportionately favored what's quantifiable over what's important”
Why Amazon marketplace brands have gibberish names
Sutherland explains that random consonant-and-vowel brand names like 'Lulu' exist because Amazon's algorithm prioritises trademarked names, and a made-up string is far faster and easier to trademark than a real word. This games the algorithm but serves no consumer interest. He argues brands are the units of selection in consumer capitalism: a real brand lets you reward good experiences and punish bad ones, so nonsense names actively erode market quality.
Gibberish names are easy and fast to legally trademark
Amazon's algorithm rewards trademarked names, so sellers game it
Brands are the 'units of selection' in consumer capitalism
Real brands let consumers reward and punish; nonsense names cannot
The practice degrades markets and consumer confidence
“brands are basically the units of selection in the evolutionary Marketplace which is consumer capitalism”
#branding#amazon#marketplaces#trademarks
◆Hot Take1:31:00
How Google and Amazon started serving themselves, not you
Sutherland describes the 'enshittification' of once-admirable platforms. Google's search degraded as ads crept from a clean right-hand column across everything, so finding a hotel's phone number now surfaces 76 competing hotels. Amazon has moved from serving the consumer to serving the advertiser and then itself, so searching for Samsung televisions no longer reliably returns Samsung televisions. The pattern: platforms serve consumers, then advertisers, then only themselves.
Great platforms decayed from serving users to serving advertisers to serving themselves
Google search filled with competing results for simple lookups
Amazon search for a named brand no longer reliably returns it
The decline creates an opening for a competitor to wrong-foot them
“if I search for Samsung televisions I expect my search to include some Samsung televisions”
#big-tech#platforms#google#amazon
✶Explainer· 5
✶Explainer10:30
Why basic statistics is a business superpower (and confident amateurs are dangerous)
Sutherland argues that a fairly good grasp of statistics is a workplace superpower because so few people have it, while confident but averagely-good statisticians are actively dangerous. He illustrates with the Sally Clark cot-death case, where multiplying probabilities as if independent, and never comparing double-cot-death against double-infanticide, turned a likely-innocent woman into a wrongful conviction. Eminent medics, barristers and judges all made the error.
You only need A-level-standard statistics for it to be a superpower
Averagely-good but confident statisticians are more dangerous than ignorant ones
The Sally Clark case multiplied cot-death odds as if independent
The real test is comparing double-cot-death vs double-infanticide probability
Oxbridge-educated experts and judges still got it catastrophically wrong
“averagely good statisticians particularly if they're confident are actively damaging and dangerous”
Sutherland warns that data is only reliable if it is representative and if the future resembles the past, which is safe over a year but not over a decade. He cites a dairy company with milk delivery down to a fine art until the law let people buy milk from supermarkets, instantly voiding every assumption. Quantification bias compounds this: we over-collect what is easily measurable and mistake it for what matters.
Data is only reliable if the future resembles the past
Unrepresentative data produces similarly biased conclusions
A single law change voided a dairy company's fine-tuned model overnight
Quantification bias favours the measurable over the important
“all big data comes from the same place the past”
#data#forecasting#quantification-bias#risk
✶Explainer44:30
Why 100 people once is not the same as one person 100 times
Sutherland explains that marketers look at the world through the consumer's eyes over time, at 90 degrees to the rest of the organisation, which sees only snapshot aggregates. High Speed 1 saved a few Kent commuters an hour every day, a genuine life-changer, while High Speed 2 saves many people an hour once a year. Mathematically identical in aggregate, psychologically and behaviourally totally different, and the average customer often does not even exist.
Marketers view customers over time; the organisation sees snapshot aggregates
Saving a few people an hour daily is behaviourally unlike saving many an hour yearly
HS1 changed lives for Kent commuters; HS2 does not for Manchester
The 'average customer' is often a statistical fiction
Data is never objective; it depends on the angle you view it from
“the average customer probably doesn't even exist”
“saving a lot of people an hour infrequently looks the same to the statistical model as saving a few people an hour every day”
#marketing#behavioural-science#data#aggregation
✶Explainer1:06:30
Leading brands win on repertoire, not loyalty
Citing Byron Sharp, Bennett and Field, Sutherland explains that most brands are bought as part of a repertoire, and what distinguishes leading brands is simply that more people buy them sometimes and buy them slightly more often. His favourite illustration is the Dos Equis 'Most Interesting Man in the World', who does not always drink beer, and would not stick to one brand, but when he does, he prefers Dos Equis. Perfect loyalty is not the point.
Most brands are bought as part of a repertoire, not exclusively
Leaders simply have more people buying them sometimes, more often
The Dos Equis line captures repertoire buying perfectly
Chasing perfect loyalty misunderstands how brands actually grow
“I do not always drink beer but when I do I prefer Dos Equis”
#branding#marketing#loyalty#growth
✶Explainer1:23:00
How choice architecture distorts markets: the property-price ziggurat
Sutherland shows that the order and attributes by which people eliminate options dramatically shape what they choose, one of behavioural science's most robust findings. Property portals sort by location, price and bedroom count, so aesthetics and architecture, harder to quantify, barely register, meaning a Robert Adam or Gropius house carries near-zero premium. Price filters in £50k bands even create a 'price-demand ziggurat', so a house listed at £675k gets hidden from both searchers.
The order of elimination-by-attribute strongly shapes final choices
Hard-to-quantify aesthetics fall too far down the decision tree to matter
Great architecture adds only 0-2% to a home's price
£50k search bands turn the demand curve into a ziggurat
Bedroom-count metrics incentivise too many tiny rooms
“you don't have a price demand curve you have a price demand ziggurat”
Ask 'why' five times: the childish question that beats intermediate objectives
Sutherland defends the habit of repeatedly asking 'why', which senior civil servants regarded as childish, almost like banging cutlery on the table. It is actually highly intelligent because in complex organisations intermediate objectives start dominating attention at the expense of the ultimate objective. The apparently low-status question forces you back to what actually matters.
Asking 'why' five times looks childish but is a sophisticated tool
Intermediate objectives quietly hijack attention from the real goal
Low-status questioning is a way to expose that drift
Seniority makes people reluctant to ask basic questions
“intermediate objectives start dominating... at the expense of what is the ultimate objective”
Steal the Amazon Prime subscription model into new industries
Sutherland is a fan of learning from unrelated businesses via recurring fractal patterns, and points to how many industries could steal the Amazon Prime membership model. He proposes a hotel-chain crash-pad subscription (pay a small annual fee, get any-vacancy rooms cheaply after 10pm) and first-class train carriages reserved for season-ticket and frequent travellers, mirroring how airlines already treat frequent flyers. Prep coffee's subscription already works this way, and reshaped his loyalty entirely.
Many businesses could transplant the Amazon Prime membership model
A hotel crash-pad subscription would suit commuters who miss their train
Trains could reserve first-class carriages for frequent and season-ticket travellers
Airlines already treat frequent flyers this way; railways never have
Prep coffee's subscription changed the guest's habits and loyalty completely
“there are probably... 10 percent of businesses out there that could steal the idea of Amazon Prime”
#subscriptions#loyalty#business-models#innovation
▲Takeaway2:06:30
Why chains like Travelodge and Starbucks quietly created value
Against the middle-class reflex to sneer at chains, Sutherland defends Travelodge, Premier Inn and Starbucks for raising the floor of quality. Before reliable chains, every coffee or hotel booking was a gamble with a high chance of being terrible; a dependable chain sets a standard everyone else must beat. He quotes an independent coffee-shop owner who credited Starbucks with making it possible to charge £3 for a coffee at all.
Chains created huge value by raising the floor, not just the ceiling
Before them, a coffee or hotel was a gamble with high downside
A reliable chain forces new entrants to be at least that good
An independent owner credited Starbucks with enabling his whole business
Middle-class status-signalling blinds people to this value