How to read recommended books critically

Q: How many should I read?

Fewer, better, with [notes](/communication-skills-metrics). One book read against a real decision and recorded

This page recommends no books, not because none deserve it, but because a list of recommended books is the least useful thing a business reading guide can offer.

It tells you what one person liked at one time, with reasons never stated. It goes stale the moment it is published.

What does not go stale is the ability to look at any business book from any year. Within a few pages, you can know what kind of thing it is.

You can see what it can and cannot show you and how to read it to keep the useful part.

What to take away

Business books come in a handful of types, and each type can establish a different kind of claim. A memoir cannot prove a method; a survey cannot give you a feel for a decision. Know which you are holding.
The genre has recurring failures, and they are structural, not the author's fault. Survivorship, hindsight, and the single idea stretched to book length will be in most of what you read. Reading well means seeing them.
Choose a book by the decision in front of you, not by the author's fame or the cover's promise. Then read it looking for what would have to be true for its advice to be wrong.

Five types, and what each can show

Start by classing the book. The type is usually clear from the first chapter,

Five book types and their limits

Type

Memoir: What author believed
Case collection: Pattern exists somewhere
Method book: How to do procedure
Research summary: What studies found
Textbook: Settled vocabulary and methods

Can establish

Memoir: Whether it caused result
Case collection: Whether pattern is common
Method book: Whether it beats alternatives
Research summary: More than studies showed
Textbook: Why it matters to you

Cannot

Memoir
Case collection
Method book
Research summary
Textbook

and it determines what you can legitimately take from it.

What it is

Memoir: One person's account of what they did
Case collection: Several businesses described, usually chosen to illustrate a point
Method book: A procedure the author proposes, with steps
Research summary: A writer explaining a body of studies
Textbook or reference: Definitions, techniques and standard practice, organized for lookup

What it can establish

Memoir: What that person believed, and what a decision felt like from inside
Case collection: That the pattern the author saw exists somewhere
Method book: How to do the procedure
Research summary: What the studies found, filtered through the writer's reading of them
Textbook or reference: What the settled vocabulary and methods are

What it cannot

Memoir: Whether the decision caused the result, or would work for you
Case collection: That the pattern is common, or that the cases were not picked to fit
Method book: Whether the procedure works better than the alternatives, unless it says how that was tested
Research summary: More than the studies themselves showed, and often the writer claims more
Textbook or reference: Why any of it matters to your situation

Two of these are routinely misread. A memoir is read as a method: the author did these things and succeeded, so the things are the method. A case collection is read as a survey: twelve firms did X and did well, so X works.

Neither inference holds, and this page's first section is already most of the reason why.

Examples help. Shoe Dog by Phil Knight is a memoir; In Search of Excellence by Tom Peters and Robert H. Waterman Jr. is a case collection; The Lean Startup by Eric Ries is a method book; Good to Great by Jim Collins is a research summary; and a standard financial accounting textbook is a reference. Read each as its type, not as the type you wish it were.

The recurring failures

These are features of the genre. The best books in it manage some of them; no

book manages all of them.

Survivorship. The businesses that get written about are the ones that lasted. The ones that did the same things and failed are not in the book, not because anyone hid them, but because nobody knew to look. Any lesson drawn from successful cases alone is a lesson drawn from half the evidence, and it is the half that was always going to agree.

Hindsight. Once the outcome is known, the decisions that led to it look inevitable and the alternatives look foolish. Written afterwards, the story acquires a shape it did not have at the time, and the reader inherits a confidence the decision-maker never had.

The halo. A firm that has done well is assumed to have done everything well, so its culture, its office layout and its hiring practice all become causes of its success. Take the same practices in a firm that later fails and they become the causes of that instead. The practices did not change; the outcome did.

One idea, book length. Many business books contain a single insight that fits in an essay. The remaining chapters are the insight applied to further examples, each of which was chosen because the insight applies. The book gets longer and the evidence does not.

The proprietary framework. A set of steps or a diagram, named and often trademarked, presented as a discovery. Sometimes it is a genuinely useful way of organizing a problem. The test is whether it makes a prediction that could fail. A framework that explains every case, including opposite ones, has not explained anything.

The missing denominator. "Firms that did X grew" is not a finding until you know how many firms did X, how many of those grew, and how many that did not do X also grew. Most business narrative gives the numerator only.

Author incentive. Many business books exist to sell something else: consulting, speaking, a further book. That does not make them wrong. It means the advice will tend toward the thing being sold, and the reader should know what that is.

Two well-known examples show the patterns. In Search of Excellence is the usual teaching example of the halo and survivorship problems, because several of the companies it praised later struggled. Good to Great is the usual teaching example of the missing denominator, because its cases are selected after the outcome was known.

How to read one critically

A method that fits on a card.

Class it first.Use the table above. Decide what kind of claim this book is capable of supporting before you read a chapter.
Find the thesis, in one sentence.If you cannot state it, the book may not have one, and you are reading a collection of anecdotes.
Ask what would falsify it.What would the world have to look like for the thesis to be wrong? If the book never describes such a world, it is not making a claim, it is telling a story.
Look for the denominator.Every time a success is described, ask how many attempts there were. Every time a practice is praised, ask whether the book mentions anyone who used it and failed.
Separate the observation from the prescription.An author may have seen something real and drawn the wrong instruction from it. Keep the observation, question the instruction.
Date it.When was it written, and what has changed? A method that depended on conditions that no longer hold is history, and history is useful as long as it is labeled.
Write down what you will do differently.One sentence. If there is nothing, the book was entertainment, which is fine, as long as it was not bought as instruction.

The vocabulary needed to do this is small but exact, and the difference between a claim and a fact, or between evidence and information, is set out under decision skills. A book that blurs those pairs is asking you to blur them too.

How to choose one

The wrong way is by reputation. The bestseller list measures sales, which

measure marketing, title, timing and the appetite of other readers, none of which is the book's usefulness to you.

The right way starts from the other end.

Name the decision you are facing.Pricing something. Hiring the first person. Deciding whether to borrow. A book chosen against a live decision is read differently: you are looking for something, so you notice whether it is there.
Match the type to the need.If you need to know how to do a procedure, a method book or a reference. If you need to know what a decision feels like from inside, a memoir, read as one. If you need to know whether something works in general, a research summary, checked against the studies it cites.
Check the author's distance from the evidence.Did they do the thing, study the thing, or interview people who did? Each is legitimate; each produces a different kind of knowledge, and the book should be honest about which.
Read the contents page and one middle chapter before buying.The introduction is where the promise lives. The middle is where you find out whether the book keeps it.
Prefer books that argue with themselves.A chapter on where the method fails is the strongest signal an author knows what they are claiming.

For the financial subjects specifically, the words a book uses should match the definitions in business vocabulary and financial literacy. A business book that uses profit and cash interchangeably has told you something about its precision before it has told you anything about business.

Reading as a practice rather than a purchase

The habit that separates people who learn from business books from people who

The habit that separates people who learn from business books from people who merely accumulate them is note-taking of a particular kind. Not summaries; the book already is one. Three lists, kept across every book you read:

Disagreements.When a book contradicts one, you have a live disagreement to resolve rather than a vague sense of having read both sides.
Decisions.What the book changes about the decision you named before you bought it.
Mistakes.A mistake the book names, and whether you have made it or could. If the book names a mistake, the mistake is probably worth recognizing in your own work.

Three lists kept across every book

Claims I now believe, and which book made me believe them
Things I will do or stop doing, dated
Patterns across books

That third list is where the genre earns its keep. One business book proves little, but thirty, read with the failures above in mind, give you the base rate no single book contains.

Asking which number a claim actually rests on is the same habit that business acumen treats as the core skill. It applies that habit to a page instead of a spreadsheet.

Common questions

Are classic business books better than recent ones?

They are different. Older books survive because something in them kept being useful, which is a real signal, and their examples are old enough that the outcomes are fully known. Recent books have current conditions and unresolved outcomes. Read both, and label which is which.

Is a book by a successful founder more reliable?

More reliable about what they experienced. Not more reliable about what caused it, because they are inside the survivorship problem, and a good founder-author will say so.

What about summaries and abridgements?

A summary keeps the thesis and discards the evidence. That is exactly the wrong way round for critical reading, since the evidence is where the weaknesses show. Use a summary to decide whether to read the book, not instead of reading it.

How many should I read?

Fewer, better, with notes. One book read against a real decision and recorded in the three lists is worth a shelf that was read on holiday and forgotten.