Beta readers: what to ask, and when
A beta reader is someone who reads your book before it is finished and tells you what happened to them while they read it. That is the whole job, and it is a useful one.
Most authors get little out of it, for three reasons that have nothing to do with the readers. They send the book too early, they send it to too few people, and they ask a question that cannot be answered honestly.
When they come in
After your own revision. Not during, and never on a first draft you already know is broken.
If you can name what is wrong — the middle sags, the ending is rushed, chapter nine is a placeholder — you do not need readers to tell you. You will spend their goodwill hearing your own diagnosis back, and goodwill is the scarce resource: most people will read your manuscript once, not three times.
So the sequence is: draft, your own revision pass, then beta readers. Their reports either go straight into another revision, or they go to a developmental editor alongside the manuscript, which makes that edit sharper and sometimes shorter. Running betas after the developmental edit instead, as a check that the repairs worked, is equally defensible. What is not defensible is sending out a draft you are still ashamed of and hoping five strangers find the thing you are avoiding.
All of this happens well before typesetting — see The editing passes that have to happen before typesetting for where the paid passes sit.
Who makes a good one
People who read your genre, by choice, for pleasure. A reader who gets through four romances a month knows, without being able to explain it, when your third act breaks the promise the first one made. A literary reader given a cozy mystery will tell you the prose is thin, which is not a finding.
Not writers, by default. Writers are trained to fix things. They hand back line edits, structural theories and an alternative version of chapter two — all of it about craft, none of it about what they felt. A couple of writer friends is fine. A panel made only of writers gives you a workshop, and a workshop is a different tool.
Not people who love you. Your partner, your mother and your closest friend are not lying to you exactly; they are reading a book by someone they care about and cannot separate the two. They will tell you it was great, they will be sincere, and you will learn nothing.
Strangers are better than friends here. Genre reader groups, a newsletter list, a writing community where you have read for other people first.
How many
A handful — five or six who will actually finish. Recruit more than that, because some will not.
One reader's opinion is noise. You cannot tell whether the person who lost interest at chapter eleven found a real fault or was having a bad week, and rebuilding the book on that one report may fix nothing. Five overlapping reactions are data. When four people slow down in the same stretch, that stretch is slow, whatever any of them say about why.
Two readers is the worst number, because they will disagree and you will be left choosing which one to believe.
What to ask
This is the part that decides whether the exercise works.
"What did you think?" produces politeness. It asks the reader to deliver a verdict on months of your work, to your face, and almost everyone answers I really enjoyed it and leaves the conversation.
Useful questions are specific, and they ask about experience, not craft. Your reader is the world expert on what it was like to read your book and a beginner at fixing it. Ask about the first thing.
- Where did you put it down? The single most valuable question you can ask. Not where they were bored — where life interrupted and they let it. That is where the book stopped pulling.
- Where were you confused? Page or chapter, if they can find it. Confusion is nearly always a real defect, and it is usually information delivered too late or in the wrong order.
- Who did you care about, and when did that start? If the answer is nobody, or page 90, you have found the problem with the opening.
- What did you expect to happen that did not? This surfaces promises you made without noticing, which is most of what a disappointing ending actually is.
- Was anything hard to believe? Plot, motivation or research, all in one question.
- Would you have bought this from the blurb? Send the blurb with the book and ask separately. See How to write a blurb that sells the book.
Six questions is about the ceiling. A twenty-item form gets abandoned, or answered carelessly, which is worse.
In the tools. Build the EPUB from your manuscript and send that: it reads properly on a phone or a Kindle app, which is where most of your betas will actually read. Later, for the advance-copy round, ARC Delivery emails each reviewer their own watermarked copy and reports who downloaded it. Start a book → · about the tools
How to read what comes back
Readers are reliable about what they felt and unreliable about how to fix it.
So take the symptom and throw away the prescription. "You should cut the flashback in chapter four" is a prescription. Underneath it is a symptom — the story lost momentum there — and the flashback may be the cause, or the scene before it may have ended on nothing. That call is yours, because only you can see the whole structure.
Then look for the pattern. Lay the reports side by side and count what three people mention. One strong, articulate, confident voice is the most dangerous input you will get: it arrives sounding like a professional assessment and it is still a sample of one. Weight what recurs.
And watch for what nobody mentioned. If no one brings up the character you built the book around, that is a finding, and it will not appear in any of the answers.
The practical side
Send a clean, readable file. Not a marked-up working draft with your notes in it. An EPUB or a PDF that reads comfortably on a phone. Friction loses readers, and a reader who quits for format reasons tells you nothing about the book.
Give a deadline, and make it real. Three to four weeks. An open-ended ask gets read in six months or never.
Give permission to stop. Tell them outright: if you lose interest, stop, and tell me where. A reader who quits at page 60 and says so is worth more than one who grinds through out of obligation and reports that it was fine.
You do not need a watermarked copy at this stage. Betas are a few people you chose. Watermarking belongs to the later advance-copy round, when you send to reviewers you do not know — see Getting early reviews without breaking a rule.
Make yours
Everything here is built into the tools, and they are free to use: name your book, bring your manuscript or your artwork, and work as long as you like. You only pay when you want the files without the preview watermark.
Next: Getting early reviews without breaking a rule · Back to The editing passes that have to happen before typesetting · Also: Preparing a manuscript that typesets cleanly