The Collaborative Book Beta
Expose a draft to shared editing and let readers improve each other's fixes
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 94%
Gawdat says he writes books like software by issuing a beta version to about 270 readers in a shared Google Doc. Readers receive editing privileges, so they can propose changes directly. Because everyone sees the document, later readers can edit or respond to earlier readers' changes, producing a conversation around difficult passages. The author then uses those interactions and reader behaviour to improve the book. Gawdat says a sentence declaring happiness a choice caused about eight percent of his Solve for Happy beta readers to leave, which helped him inspect the claim's effect. The numbers are his personal process, not required sample sizes. The reusable mechanism is shared, observable iteration while the author remains responsible for evidence, ethics, permissions, and the final text.
Origin
Gawdat says his software background led him to issue books as beta versions and refine them through a shared Google Doc before publication.
Core principles
- 01A draft can be treated as a beta rather than a finished object
- 02Readers need enough access to propose concrete changes
- 03Shared visibility allows feedback to build on earlier feedback
- 04Reader drop-off can reveal where an argument loses people
How to run it
- 1
Build the beta
Prepare a coherent draft that readers can experience from beginning to end. Label it clearly as unfinished and specify the questions you want tested.
Pro tip Include a version date so feedback refers to the same draft.
Watch out Remove confidential or legally sensitive material before sharing.
- 2
Recruit test readers
Invite readers who can represent relevant perspectives and give concrete feedback. Explain access, attribution, and confidentiality expectations.
Pro tip Prioritise relevance and diversity over copying Gawdat's reader count.
Watch out Obtain consent before tracking or using personal reader information.
- 3
Enable shared revision
Use a collaborative document where readers can suggest edits and see existing suggestions. Preserve version history and a recoverable original.
Pro tip Suggestion mode can protect the base text while keeping changes visible.
Watch out Unrestricted editor access can damage or leak a draft.
- 4
Watch the conversation
Observe where readers reinforce, challenge, or rewrite one another's feedback. Treat repeated confusion and disagreement as signals to inspect the passage.
Pro tip Group comments by the underlying reader problem.
- 5
Review reader behaviour
Where ethically and transparently available, note where readers stop, skip, or disengage. Investigate the cause instead of assuming the argument is wrong.
Pro tip Pair behaviour with direct reader explanation.
Watch out Do not infer mental-health status or motives from drop-off alone.
- 6
Ship the author's revision
Resolve suggestions into one deliberate version. Fact-check claims independently and keep final editorial responsibility with the author.
Pro tip Record why repeated suggestions were accepted or rejected.
Watch out Consensus among beta readers does not establish factual truth.
In the wild
Gawdat gives roughly 270 beta readers editing privileges in Google Docs. He says some readers modify changes proposed by others, turning separate feedback into a visible discussion that improves the manuscript.
→ Feedback becomes iterative and relational rather than a pile of isolated notes.
Gawdat says about eight percent of early readers left around a sentence stating that happiness is a choice. He reports that many had disclosed depression, but this was his interpretation of beta-reader information, not clinical research.
→ A high-friction claim became visible for further scrutiny rather than passing unnoticed.
Common mistakes
Copying the 270-reader count
The transcript supports that as Gawdat's practice, not as a validated or necessary sample size.
Letting consensus replace evidence
Reader agreement can improve clarity but does not prove factual, scientific, or medical claims.
Sharing without safeguards
Collaborative access requires permissions, privacy controls, version history, and clear ownership expectations.
Is it for you?
Best for
Authors with a substantial draft who can safely share it with a trusted and sufficiently varied test group.
Not ideal for
Confidential manuscripts, untrusted groups, or projects where uncontrolled editing would create legal, privacy, or ownership risks.
From the transcript
“when I write books in general, I write them like software. So I issue a beta version”
“there is a conversation happening, and basically it takes the book to its best possible version”
From the episode
The Happiness Expert That Made 51 Million People Happier: Mo Gawdat