Dobelli catalogues around 100 common cognitive biases and thinking errors — from confirmation bias to sunk cost fallacy — in short, two-to-three-page chapters, each with a clear real-world example, making it an easy reference to dip into rather than a book requiring cover-to-cover reading.
A catalogue, not an argument
Dobelli doesn't build a single thesis the way Kahneman or Ariely do. This is deliberately a reference book — 99 short chapters, each covering one cognitive bias or thinking error in two or three pages with a single clear example — and it began life as a newspaper column, which explains both the format and its virtues. The chapters are self-contained, briskly written and end with a usable instruction rather than a discussion. You are not meant to read it cover to cover; you are meant to recognise a bias by name in the moment you are about to commit it.
The organising philosophy behind it is worth stating, because Dobelli does state it and most summaries skip it. His claim is that clear thinking is achieved by subtraction rather than addition — that avoiding a catalogue of predictable errors will improve your decisions more than acquiring any amount of new insight. It's the via negativa applied to judgement, and it sets a modest, achievable bar: you are not trying to be brilliant, you are trying to be less reliably stupid in the handful of ways that everybody is.
That modesty is also the book's honest ceiling. It will not teach you probability, it will not make you a better forecaster, and it contains no method. What it does is give you names for things you already half-noticed, and names are surprisingly powerful in a meeting — 'that's survivorship bias' ends an argument faster than three paragraphs of careful reasoning.
The biases that recur most in business
A handful of entries do the heaviest lifting for an owner. Survivorship bias — drawing conclusions from the successes you can see while ignoring the much larger, invisible pile of failures — explains why 'do what the successful founder did' advice is so often useless: you are studying the one boat that didn't sink, and the seabed is not available for interview. Sunk cost fallacy explains why failing projects, failing hires and failing product lines attract more investment rather than less the longer they have already cost money, because the emotional pull is to justify the past spend rather than to evaluate the decision fresh.
The swimmer's body illusion is the one most worth internalising. Elite swimmers do not have those physiques because they swim; people with those physiques get selected into elite swimming. Confusing a selection criterion with an effect underlies a vast amount of casual business advice — the MBA, the accelerator, the networking group, all of which may be filters rather than causes. Once you have the concept, you cannot unsee it in marketing copy.
Chauffeur knowledge is the other keeper, and Dobelli tells it via Max Planck's driver, who had heard the lecture so many times he could deliver it flawlessly and was undone by the first real question. The distinction between people who genuinely know something and people who can perform knowing it is directly relevant to hiring, to choosing advisers and to assessing anyone confident on a stage. Dobelli's tell is simple: someone who genuinely knows a subject will say 'I don't know' without discomfort.
Overconfidence, control and the stories we tell afterwards
Several chapters attack the illusion that we understand and control more than we do. The illusion of control is the tendency to believe our actions influence outcomes they don't — pressing a lift button repeatedly, or believing a particular routine caused a good sales month that was mostly noise. The overconfidence effect shows people are systematically more certain of their forecasts than their accuracy justifies, and worse outside their genuine expertise than inside it, which matters directly for anyone projecting revenue or committing to a launch date.
Regression to the mean is the quiet troublemaker among them. Exceptional results — a terrible month, a stellar quarter — tend to be followed by more ordinary ones for purely statistical reasons, and because we intervene after the extremes, whatever we did gets the credit or the blame. This is why the consultant hired after the worst quarter always looks effective, and why praising good performance appears to make it worse.
A related cluster covers the ways we mispredict our own behaviour and everyone else's. Action bias is the compulsion to do something in the face of uncertainty even when waiting is demonstrably better — Dobelli's example is a goalkeeper facing a penalty, who dives because standing still looks like failure even when standing still saves more. Its mirror, omission bias, is why we judge harm caused by doing something more harshly than identical harm caused by doing nothing, which is roughly how most businesses end up with a legacy product nobody would launch today. And the incentive super-response tendency — people respond to the incentive rather than to the intention behind it — is the single most useful chapter for anyone about to design a commission scheme.
Story bias and hindsight bias then finish the job. We compress messy sequences of events into narratives with causes and lessons, and once we know how something turned out, it feels as though it was always obvious. Together they make business history close to useless as evidence and make everyone, including you, feel a great deal smarter about the past than the record supports. Dobelli's practical countermeasure is the only one he really insists on: keep a written record of what you predicted at the time, because your memory of your own forecasts is not a reliable witness.
Confirmation bias, social proof and how to actually use the book
Dobelli treats confirmation bias — noticing and remembering evidence that fits what we already believe, and quietly discounting what doesn't — as close to the master error underlying the rest, because it is what stops people updating even after reading a corrective chapter. His fix is deliberately effortful: go looking for the disconfirming evidence on purpose, write down what would change your mind before you gather it, and treat comfortable agreement as a warning sign.
Social proof and authority bias round out the set most relevant to running a business: the tendency to copy what others, especially high-status others, are doing and to treat that as evidence in itself, independent of whether it is working for them. In a small business this shows up as competitor-watching — matching a rival's discount, their opening hours, their new service line — on the assumption that they know something you don't, when they are frequently copying somebody else.
For anyone who prices or negotiates, three more entries earn their keep. The anchoring effect means the first number mentioned drags every subsequent number towards it, whoever mentions it and however arbitrary it is. The contrast effect means the same price feels cheap or expensive depending entirely on what it was shown next to, which is why the option nobody buys still earns its place on the menu. And the endowment effect means people value what they already own well above what they would pay to acquire it — the reason free trials convert, and the reason your own stock, your own service line and your own original idea all feel more valuable to you than to the market.
As for how to read it: not straight through. Ninety-nine chapters of 'here is another way you are wrong' becomes wearing and repetitive in one sitting, and several entries overlap enough that the distinctions blur. The book works as a desk reference consulted at the moment of a real decision — before renewing a lease, before doubling down on a struggling hire, before believing a success story. Read three chapters, apply one, and it will have paid for itself; read all ninety-nine in a weekend and you will retain almost none of them.
Key lessons
- Sunk cost fallacy leads people to keep investing in a failing decision purely because of what's already been spent.
- Confirmation bias means we notice and remember evidence that supports what we already believe, and overlook what contradicts it.
- Survivorship bias skews our sense of what works, because we only see the successes, not the much larger number of failures.
- Social proof and authority bias make us weight other people's choices and opinions more heavily than we should in our own decisions.
Most bad decisions follow a small, well-documented set of recurring thinking errors — being able to name and recognise them in the moment is the single most useful defence against them.
What this means for a UK small business
This is the most practical book on the list to keep as a genuine desk reference rather than a one-read, precisely because of its format. Sunk cost is worth checking at the exact point a UK owner is deciding whether to keep funding a struggling service line, a hire who isn't working past probation, or a lease that no longer fits — ask what you would do starting fresh today, ignoring everything already spent, and treat any gap between that answer and your instinct as the bias talking.
Survivorship bias earns its place every time a franchise pitch, a 'how I built my seven-figure agency' course or a competitor's public success story lands in your inbox. You are seeing the one visible winner, not the quieter pile of businesses that tried the same thing and dissolved before anyone wrote about them. The same applies to the case studies on any supplier's website.
Putting real numbers on the sunk cost test, illustratively: say you have spent £18,000 over nine months on a service line that has brought in £4,000. The instinct is to reason that £18,000 is too much to walk away from. The correct question ignores it entirely — starting today, with nothing spent, would you commit the next £2,000 a month for a line billing under £500 a month? If the honest answer is no, the £18,000 is gone either way, and every further month costs you £1,500 you could put somewhere that works. The £18,000 is not a reason to continue; it is the reason it feels impossible to stop.
Chauffeur knowledge is the one to apply to advisers. Plenty of people can talk fluently about R&D tax relief, marketing funnels or employment law without having done any of it; the tell is what happens when you ask a specific question about your own situation. A genuine specialist narrows, qualifies and occasionally says they'll check. A chauffeur keeps talking.
What’s aged well
The catalogue of biases draws on well-established behavioural science and remains a useful, evergreen reference.
What feels outdated
Nothing significant; short standalone chapters age individually rather than as a whole.
Where it falls short
It's a popular summary rather than primary research, and it shows: entries oversimplify genuinely contested areas of psychology, sources are relegated to the back, and a fair number of the 99 biases overlap so closely that the distinctions feel like padding to reach a round number. It has also aged into the replication crisis — several effects presented here as settled have since been weakened or seriously questioned, and the book gives no sense of which findings are rock solid and which are shaky.
There is also the provenance question. Dobelli was publicly accused, most vocally by Nassim Nicholas Taleb, of leaning far too heavily on other writers' ideas without adequate credit — worth knowing before you cite it as a source rather than as a prompt.
The Business Stuff verdict
A genuinely useful quick-reference for common thinking errors, best dipped into rather than read in one sitting.
Three things to actually do after reading it
- Identify one current business decision at risk of sunk cost fallacy, and evaluate it fresh, ignoring what's already spent.
- Actively seek out one piece of evidence that contradicts your current strategy, rather than only confirming evidence.
- Keep the book as a reference and check a live decision against the relevant bias before committing to it.
If you liked this, read next
Five similar books
- Thinking, Fast and Slow (Daniel Kahneman)
- Predictably Irrational (Dan Ariely)
- Nudge (Thaler & Sunstein)
- Superforecasting (Tetlock & Gardner)
- Noise (Kahneman, Sibony & Sunstein)
Common questions
Should I read this or Thinking, Fast and Slow?
Thinking, Fast and Slow if you want to understand how judgement actually works; The Art of Thinking Clearly if you want a quick-reference list of named errors. Kahneman is the primary source, written by the researcher, and it builds a genuine theory — it is also long, dense and takes real effort. Dobelli is a catalogue: no theory, no argument, ninety-nine two-page entries you can consult in the middle of a decision. Many people get more day-to-day use out of Dobelli and more actual understanding out of Kahneman. If you're only reading one and you'll finish it, Dobelli. If you'll finish either, Kahneman.
Is it true that Dobelli was accused of copying other writers?
Yes, and it's worth knowing. Around the book's English publication, Nassim Nicholas Taleb publicly accused Dobelli of drawing far too heavily on his work — and on other popularisers such as Cialdini — without adequate attribution, and the dispute was covered in the press at the time. Dobelli rejected the characterisation, and the book does list sources at the back rather than in the text. It doesn't make the content wrong, since none of these ideas originate with Dobelli anyway and he never claims they do. But it's a reason to treat the book as a prompt rather than a citation, and to go to the primary sources if something matters.
Have any of the biases in the book been debunked?
Some have been weakened, which is a real limitation of any 2013 popular psychology book. Psychology has been through a replication crisis in the intervening years, and several once-standard findings — priming effects and ego depletion most prominently — have failed to replicate or shrunk substantially under scrutiny, while others such as anchoring survived but with smaller effects than originally reported. The heavy hitters most relevant to business — survivorship bias, sunk cost, confirmation bias, regression to the mean, overconfidence — are on much firmer ground, partly because several are logical or statistical points rather than experimental results. Treat the book as a starting list, not a settled one.
How should I actually read a book of 99 short chapters?
Not in one go. Read the contents page, pick the ten entries whose titles describe a mistake you recognise in yourself, read those properly, and shelve the rest for reference. Then use it reactively: before a decision that involves money you can't easily get back — a lease, a hire, doubling down on something struggling — look up sunk cost, survivorship bias and overconfidence and read the three pages. The whole thing takes about four hours end to end if you insist, but retention from a single sitting is poor because the chapters have no connective tissue and start to blur by the fortieth.
