How to get feedback you can use

1 The bottom line

Pick a few readers who fit the job, ask them where the draft hurts, and diagnose the cause yourself. Readers can feel a problem but rarely know its cause, so copying their fixes changes the wrong things.

2 Key takeaways

• Choose a few readers on purpose and give each one a job. Open invitations produce a pile of contradictory notes.
  • Jack Carr1:00:19 sends each draft to friends before his publisher, including a photographer who checks weapon details and two attorneys who catch small errors.
  • When Ana Lorena Fabrega34:10 opened her draft to six readers, everyone said something different. She then assigned roles: her husband did line edits, and another reader gave high-level notes on stories.
• Add at least one reader who is not like you. You cannot see your own jargon or gaps, and experts cannot say what confused them.
  • Steven Pinker7:41 used to show drafts to his mother, who was bright and well read but not a cognitive psychologist.
  • Nat Eliason26:53 took feedback from women readers and family more seriously than from crypto readers, who could correct facts but could not say what was confusing.
• Treat every note as a symptom and find the cause yourself. The thing being made is in your head, and the reader's fix is a guess.
  • Editor Lee Boudreaux suggested that Andrew Sean Greer1:09:45 cut the revelatory ending of Less. He rewrote the rest of the novel so the ending landed.
  • Everyone said the father in Finding Nemo was too annoying. Andrew Stanton34:02 moved the scattered flashbacks into one opening scene instead, and the complaints about the father's scenes and lines vanished.
• Act on what several readers repeat. Agreement filters out one person's taste.
  • Rob Henderson21:15 followed a note when at least two of about ten readers raised it, like interrater reliability.
  • When every reader had something to say about two early chapters, Alex Hormozi25:48 cut three chapters, and the book became like reading downhill.
• Believe what readers do over what they say. Faces, laughs and stumbles are honest, and polite answers are not.
  • At family test screenings a parent often worries a film is too dark, and the child shakes their head Andrew Stanton4:44.
  • A Blue Jays riff David Sedaris8:57 improvised in a Q&A got a huge laugh and became the biggest laugh in the essay.

The editor is the patient and you are the doctor.

Andrew Sean Greer, author of Less

2.1 Choose your readers

Show the draft when it can stand up, to honest readers who want it to work, some who know your aim and some who do not.

• Wait until the draft is finished enough to judge, as Amor Towles48:57 does.
  • Towles shares no first draft until he is done. Then it goes to his editors in New York and London, his agent, and about three friends who read carefully.
  • R.F. Kuang48:00 keeps early ideas private. She fears others will project their visions onto them and she will write to suit their taste. Once she has an acceptable first draft, she sends it out to be torn apart.
  • Atul Gawande41:00 goes earlier, when a book is about three quarters done. He buys five or six writer friends dinner and wine, and the meal time boxes the feedback.
• Pick readers who are honest and rooting for you.
  • Bill Browder43:51 gives his book to ten friends of different types, not only literary people. He wants them to read carefully, mark it up and not blow smoke.
  • Derek Thompson17:40 asks readers to disagree and point out his confusions, as long as they want him to get it right. He sees little value in people who only want to dunk.
  • Brie Wolfson54:02 wants appreciators who look for the spark and pull it out. Curmudgeons treat everything as trash, which she calls exhausting and a fool's errand.
  • On Browder's second book, his wife, son and agent all said it sucked. He asked for three more months and rewrote about a third of it.
  • Jayne Anne Phillips24:55 has sent everything for years to one reader, a friend she met in a childbirth class. She says you don't cultivate such people, you run into them.
• Mix readers who know your intent with readers who do not.
  • Clint Bentley and Greg Kwedar54:25 keep a small think tank who have read nearly every draft. They also want cold readers who say what the film is about. Informed readers tell you what isn't coming across, and cold ones tell you what is.
  • Michael Mauboussin29:16 had a non-finance journalist, Lawrence Gonzalez, edit his books. Gonzalez challenged technical terms where a simpler one would do.
  • Peter Attia14:39 and his co-author argued for a month over the word extracorporeal, which Attia thought was normal. He settled on simpler wording and a footnote.

2.2 Ask the right question

Say what you are trying to do, then ask narrow questions about where the reader got bored, lost or unconvinced.

• Tell readers what you are making before you ask for notes.
  • Ryan Holiday57:49 says to work out what you are trying to do, say it to anyone giving feedback, and ask for notes that help. Treat the others as not useful.
  • Suleika Jaouad19:06 learned to be careful about who reads first and what she wants from them. She borrows Perell's labels: trunk, branch or leaf, meaning big structure, mid-level notes or small line edits.
  • On Conspiracy, Holiday went some rounds with his editor by stating the kind of book it was and what it was not.
  • The labels stop leaf comments from landing too early on an idea, or trunk comments after four months of work.
• Ask narrow questions about the reading experience.
  • Paul Graham19:58 asks only two things: what parts bore you, and what parts seem unconvincing. Boring parts can usually be cut.
  • For unconvincing parts, Graham talks the idea through with other people, since their questions show where the explanation is weak.
  • Marc Andreessen1:09:57 asks whether the piece will make too many people mad for unintended reasons, whether it is clear, and whether it moves or bogs down.
  • For Andreessen, intentional anger is fine and accidental anger is not.
  • Jaouad40:18 shows how not to ask. She asked her first reader "Is it as bad as I think?", got back "It's not working", and felt defeated.
• Give readers permission to be harsh.
  • David Perell6:43 describes a question from Tim Ferriss: what 20% must stay no matter what, then what 20% would go. The first lets friends compliment you, and the second gives them an easy way to be harsh.
  • Tim Ferriss48:43 says people will not volunteer cuts without permission, so a number gives them air cover. If a percentage feels abstract, say only one page in five can stay.
  • Dana Gioia1:49:31 asked one or two peers at his level to tell him what was wrong, not to praise it.
  • Jon Yaged30:32 says some authors do not want the unvarnished truth, so say up front what you expect.

2.3 Collect reactions cleanly

Keep readers apart, ask for quick reactions, and watch what people do as they read.

• Keep readers separate and the format simple.
  • Michael Dean1:00:12 asks for gut reactions only: green for what they like, an emoji or LOL, and a flag for confusion. Each reader gets a separate document.
  • Dean merges the responses into a master document where deeper green means more agreement, then fixes the red first. Stretches nobody comments on are candidates to cut or compress.
  • Sam Corcos8:13 asks for comments in full at the bottom of a memo. The margin forces one sentence at a time.
  • Howard Marks10:42 lets his nitpicker read first, then incorporates the changes before four or five others see it.
  • Dean says comedians test the same joke on isolated audiences, so after 50 separate rooms they know it works.
  • Marks's readers find roughly 6, 4, 10 and 200 problems. The 200 are nits like hyphens and commas.
  • Chris Dixon42:20 sent forked copies of a Google Doc to roughly 20 people, who commented on almost every paragraph. He took two weeks alone in a beach house to work through them.
• Watch a cold reader work, and say nothing.
  • Garry Tan31:36 sits next to someone using the draft and lets them struggle until they ask a question. Whatever he says in answer is what needs to go back into the writing.
  • Tish Rabe27:27 has someone who has never seen the draft read it aloud, because writers are too immersed to see what is unclear.
  • David Grann56:01 hands his first chapter to his wife and peeks at her reactions as she reads with a pen.
  • Rabe's kids stumbled over words like meteor and family, and she rewrote those lines.
  • Her groan at the ship chapter led Grann to cut thousands of words on shipbuilding from chapter one of The Wager.
• Be in the room with the reader, even if they say nothing.
  • David Gelb43:39 says to watch a cut with a couple of trusted people who are not cruel. Their presence makes you feel where it drags.
  • Morgan Housel2:55 says Mark Twain read aloud to his family. When they leaned forward he knew he was onto something, and when they got bored he cut.
  • Gelb feels responsible for the other person's time, and if they are bored it kills him.
  • Sam Parr2:16 tries phrases on calls like a comedian. When someone's eyebrows moved at "business group therapy," he kept it and built the copy around it.
• Test in public where the answer comes fast.
  • Taylor Tomlinson13:09 loves that she can write at 4:00 and learn by 8:00 whether it works. The book took two years to answer.
  • Jerry Seinfeld32:20 says Chris Rock works out a set in small clubs, keeps the jokes that get real laughs and replaces the rest.
  • Andy Weir36:45 posted The Martian chapter by chapter, and about 3,000 regular readers flagged science errors.
  • Rock's tests go down to single words. Changing "take a sip" to "take a swig" can make a joke noticeably funnier.
  • David Zucker12:07 previewed Airplane! at about 100 minutes and cut it to around 80.

2.4 Diagnose each note

A note tells you where it hurts. The cause is yours to find, and it often sits somewhere else.

• Take the problem and leave the proposed fix.
  • David Gelb49:05 compares feedback to doctor and patient. People can say what hurts but should not prescribe. He asks what the story meant to them and which parts they did not understand, then fixes it through his own taste.
  • Neil Strauss1:16:38 says to ignore the solution and focus on the problem it was meant to address. Something is not working even if the fix is not you, and the editor is a teammate.
  • Noah Hawley1:08:45 welcomes "that's confusing" and "I see what you're going for but I don't think you got there." He pushes back on "I would do it differently," because that is a different story.
  • In studio test screenings, Gelb says, viewers suddenly become directors and writers and hand you answers.
  • Strauss has stopped reading friends' manuscripts because they defend every line.
• Look earlier and deeper than the flagged line.
  • Matt Haig4:28 pulls at a small editor note and asks what the problem is really about. Often the whole character or premise does not need to be there.
  • Bentley and Kwedar55:38 say the cause of a wrong-feeling spot may sit much earlier in the script. Their rule is to try it and watch the faces of the readers.
  • When two readers flag the same line, Gioia does not just cut it. He changes the context so the line becomes stronger.
  • On The Humans, Haig cut about 40,000 of 80,000 words, then added about 55,000 between the first and second draft.
• Defend what works and change what you cannot defend.
  • Jodi Picoult27:44 steps back and stays objective. Her test is that if she can't defend something to her editor, it shouldn't be in there.
  • Joy Harjo44:34 answered a question about word order with "No, this is sound sense." Moving the word to the grammatical place would flatten the sentence.
  • Yann Martel36:56 resisted when his editor called his floating dialogue in Life of Pi clever showing off. He then admitted she was right, and the dialogue went flush left.
  • Picoult says she sometimes holds firm and was right to.
  • In Harjo's poem Shape Shifter, grammar would flip two words around the image of a deer, but the sound carries the perception.
• Learn the taste of the person giving the note.
  • Haig has worked with one editor, Francis Bickmore, for 15 years. That lets him judge when a note is one person's taste and when many readers would say it, as he explains3:07.
  • When the editor is too nice, Haig translates the polite note and asks for more brutal.
  • Ezra Klein4:40 judges an editor by taste, not line edits. He needs confidence that the person has good judgment about what is good.

2.5 Decide what to act on

Count the repeats, sort the criticism, limit the voices and keep your own direction.

• Count the notes before you change anything.
  • Jack Carr1:00:19 may ignore a note from one or two readers. If five, six, seven or eight say something, he looks again.
  • Hormozi fixes small one-off comments he agrees with and ignores the rest. When many readers hit the same section, he assumes something is wrong with it.
  • For his first book, Carr sent the draft to 20 to 25 people.
• Sort the criticism that comes back.
  • Tim Urban20:21 posted the first ten chapters of What's Our Problem? as blog posts and read hundreds of comments. He discarded the tribal and ad hominem ones.
  • He kept the substantive ones that pointed out inconsistencies, bias or missing the forest for the trees.
  • The ten chapters came out of that gauntlet tattered, and he rebuilt them tighter for the book.
• Limit the voices and trust your own taste.
  • David Shelley28:51 has watched authors go off track from too many well-meaning opinions. He says to pick a few trusted readers and keep a clear sense of your own direction.
  • Ezra Klein11:40 wishes he had trusted his editorial judgment and said no more often. When you are not confident, others cannot tell what you want from them.
  • Riva Tez13:26 agrees with Perell that fast feedback loops get in the way of writing what you think.
  • Perell says the best thing he did was disengage completely for two years.
• Review each finished piece and learn from it.
  • David Baldacci1:17:03 breaks down each book like game film. He looks at which plot setups worked and which he telegraphed.
  • He uses reader emails and friends' reactions, and treats it as his problem if enough people did not see the story as he intended.

2.6 Where writers disagree

Writers differ on when to share, how many readers to use and whether to ask for praise or problems.

• Some writers share only when done, and others share all through the draft.
  • Paul Harding33:19 avoids his agent and editor for years. He wants to ask the work what it means, not guess what readers want.
  • David Zucker1:08:53 needs friends laughing in real time because he can't test jokes alone.
  • Jimmy Soni1:24:39 texted paragraphs to a college friend almost every morning for five and a half years. He calls writing a team sport.
  • Harding adds that some brilliant writers thrive on a dozen readers. The point is to know what you need and drop the idea of what a real writer would do.
• Some writers use a few readers and others use many.
  • Paul Millerd43:40 used only five readers for the whole book and skips feedback on 99 percent of what he writes.
  • Dixon, with a long argument-driven book, used about 20 readers and says nothing substitutes for good readers telling you what they think.
• Some writers ask for praise first and others ask only for problems.
  • Anne Lamott1:13:20 and her writing partner start with what they love, name what does not work, then point to where the piece hooks them. When he led with problems, she cried.
  • Steven Levitan1:17:30 admits he was too direct and now says so when something works, page by page.
  • Jaouad, borrowing Melissa Febos's heat mapping, asks readers where the heat or electricity is in a first draft, then writes into it.
  • Millerd asks early readers only what they love, then tells his editor not to delete those parts.
  • Gioia asks for what is wrong, not praise, and Browder does not want smoke blown at him. Both fit a late draft with a structure to test.