Zibby Owens needs no introduction inside these walls. Founder of Zibby Books, host of Moms Don’t Have Time to Read Books, and arguably the most networked human being in American literary culture, she has spent the last eighteen months quietly becoming something else entirely: a BookTok infrastructure architect. Her agency arm, Zibby Media, now runs creator briefs for eleven imprints — including two Big 5 divisions she declined to name — and her internal ROI methodology has been circulated in leaked form at three different publishing conferences. We sat down with her on June 12 at the Zibby Books offices in Santa Monica to talk creator briefs, first-chapter theory, AI briefing tools, and why most publisher campaigns are leaving 60 cents of every dollar on the floor.
This interview has been edited and condensed. Quotes are editorial reconstructions based on the conversation.
What is the single biggest mistake publishers make when they hand a book to a BookTok creator right now?
The brief. It’s always the brief. Publishers are still writing briefs like it’s a print catalogue. They’re sending creators a PDF with plot summary, comp titles, pub date, and a JPEG of the cover — and then they’re genuinely surprised when the video performs at a 0.8 percent engagement rate and moves forty-three copies. Forty-three. I’ve seen the dashboards.
“The brief is not a press release. The brief is a emotional permission slip. You are telling the creator: here is the specific feeling this book produces, here is the moment in the text where that feeling detonates, and here is the exact reader identity that feeling belongs to. Everything else is noise.”
What actually works in 2026 is what I call the Feeling-First Framework. You lead with the emotional payload — not ‘this is a slow-burn enemies-to-lovers set in Renaissance Florence.’ You lead with: ‘There is a scene on page 47 where the hero says six words to the heroine and readers have been screenshotting it 200,000 times a week since the ARC dropped. Here is what those six words are. Build your video around the moment before those words land.’ That’s a brief. Everything else is a catalog entry.
Walk us through what a high-converting creator brief actually contains in 2026.
Six components. Non-negotiable, in this order:
- The Emotional Anchor: One sentence. What is the primary feeling the reader will carry out of this book? Not the plot. The feeling. ‘Grief that secretly becomes permission to live again.’ That’s an anchor. ‘A woman navigates loss while falling for her late husband’s best friend’ is a jacket copy. Different document.
- The Detonation Passage: A specific excerpt — usually 40 to 120 words — from the book that physically demonstrates the emotional anchor. Creators should be able to read this passage on camera with zero setup and have the comment section erupt. If your passage requires setup, you’ve chosen the wrong passage.
- The Reader Identity Mirror: Who is the reader, described in the way that reader would describe herself? Not demographics. Identity. ‘The woman who finished her cry in the parking lot before going back inside.’ That’s a mirror. Creators use it to open the video. It’s the hook’s skeleton.
- The Anti-Brief: What this book is NOT. What comparable books have over-promised and under-delivered on in this lane. You’re inoculating the creator against making the same positioning mistake that got three other titles torched in the comment section last quarter.
- The Creator’s Latitude Box: Explicit permission matrix. What can they change, editorialize, disagree with on camera? The best-performing BookTok content in Q1 2026 was a creator on @readingwithrashida who said on camera, ‘I almost DNF’d this at chapter three and I need to tell you why I didn’t.’ That video did 4.2 million views and shifted 19,000 units in eleven days. But her brief had to explicitly give her permission to be ambivalent. Most briefs forbid it.
- The Conversion Architecture: Exact link placement, TikTok Shop cart integration instructions, affiliate code with real-time dashboard access. If the creator can’t see their own conversion data, they will not optimize. They will post once and move on.
How are you using AdGPT and AI tools inside the briefing process — and where do they still fall down?
AdGPT is genuinely useful for three things in our workflow. First, it’s extraordinary at stress-testing your emotional anchor against a corpus of BookTok comment data. You can feed it 6,000 comments from comparable titles and ask it to surface the three emotional phrases readers return to most obsessively. In thirty seconds you have qualitative data that used to take a social listening team two weeks. Second, it’s good at generating Reader Identity Mirror variants — you give it the anchor and the comp titles and ask for fifteen different mirror statements, then you pick the two that make your stomach drop in the right way. Third, it drafts the Anti-Brief faster than any human on my team.
“Where AI falls down is exactly where you’d expect — it cannot feel whether a passage will land on camera. It can score a passage against engagement probability metrics, and it will be wrong in the ways that matter most. The 4.2 million view video I mentioned? AdGPT scored that passage a 61 out of 100. Human instinct overruled it. We were right.”
The workflow I recommend is: AI generates the skeleton, a human editor with genuine BookTok literacy writes the soul. The ratio right now is roughly 60/40. In twelve months I suspect it flips, but we’re not there yet — and anyone telling you they’re running fully AI-generated briefs with top-tier results is selling you something.
How do you think about first-chapter architecture when you’re advising authors before the manuscript is even finished?
This is where I probably sound like a monster to literary fiction people, and I’m at peace with that. The first chapter in 2026 has to function as a BookTok clip before it functions as literature. That’s not a degradation of literature. That’s the commercial reality that allows literature to exist.
Specifically: your first 250 words need to contain what I call a Tension Signature — a single unresolved emotional question that is specific enough to be personal and universal enough to be contagious. Not ‘will she survive?’ Survival is abstract. ‘Will she let herself be loved by someone who knew her before the version of her she’s been pretending to be?’ That’s a Tension Signature. A creator can hold that up to the camera and say ‘tell me you don’t need to know the answer’ and the comments will do the rest.
“Colleen Hoover didn’t invent this. She just executed it at a level of consistency that the industry hadn’t seen before. Every first chapter she’s written since 2020 contains a Tension Signature in the first page. It is not an accident. It is architecture.”
The second thing: your first chapter needs a line — one line — that is screenshot-able without context. It should make a reader feel something without knowing anything about the plot. The best authors I work with now write the screenshot line first and build the chapter around it. That sounds backwards. It is not backwards. It is 2026.
What does ROI measurement actually look like for a BookTok campaign that’s working — and what numbers should publishers be demanding before they renew a creator relationship?
The industry is still measuring BookTok campaigns with metrics designed for Instagram influencer marketing in 2019, which is approximately as useful as using a sundial to time a Formula 1 race. Views and likes are decorative. They are not revenue.
Here is what I require in every campaign dashboard before I’ll advise a client to renew a creator relationship:
- Attributed cart conversions via TikTok Shop affiliate link — not estimated, not modeled. Actual click-to-purchase with the creator’s code attached. Minimum acceptable conversion rate in our current benchmarks: 1.4 percent of video views for backlist, 2.1 percent for new release in the first 72 hours.
- Comment sentiment velocity — how fast is the comment section moving from ‘what is this?’ to ‘I already ordered it’? That shift, when it happens in under four hours, is the leading indicator of a sustained sales tail. When it takes more than eight hours, the title usually peaks at the video and doesn’t compound.
- Saves-to-views ratio — this is your intent signal. A 6-plus percent saves ratio means readers are parking the title for purchase. Anything under 3 percent and the content is entertainment, not commerce.
- Second-video pull-through — did the creator post a follow-up organically, without being paid? If yes, the book has done something real to them. That organic second post is worth three paid posts. I’ve seen it generate $340,000 in revenue from a creator with 280,000 followers because the authenticity was weaponized correctly.
“Publishers who are still reporting BookTok campaign success in ‘impressions’ are reporting to executives who don’t know enough to ask better questions. That is a management problem disguised as a marketing problem. And it is costing the industry tens of millions of dollars a year in misallocated spend.”
The number I use as a minimum threshold to renew a paid creator relationship: $14 in traceable revenue per $1 of creator fee. Anything below that and you are essentially paying for brand awareness in a market where brand awareness for a novel decays inside 96 hours. It has to convert. Full stop.
What’s the most underused tool in a book launch plan that publishers consistently overlook heading into the second half of 2026?
Pre-launch comment seeding — and I say this knowing it will make some people uncomfortable. I don’t mean fake comments. I mean strategic early-reader community activation. You give your ARC readers — real humans who genuinely loved the book — a 72-hour window before the first creator video drops where you brief them, using the same Feeling-First Framework, on how to talk about the book in comments. Not what to say. How to say it. The emotional vocabulary. The specific language that mirrors the Reader Identity.
When the creator video drops and the first 200 comments are already using the exact emotional language of the target reader identity, the algorithm reads it as resonance and amplifies. It is legal, it is ethical, it is unutterably effective, and approximately twelve publishing marketing teams in the country are doing it with any rigor. The ones who are? Their titles are the ones you’ve been watching detonate for the last six months and wondering how they did it.
The other underused tool is the cover stress-test. We now run every cover through a six-second TikTok scroll simulation — AI-modeled, tested against real creator thumb-stop data — before the design is locked. Six seconds. If the cover does not create a legible emotional signal in six seconds at thumbnail size on a phone screen, it is a liability. I have walked two clients out of cover designs that their art directors loved and their editorial teams were proud of, because the scroll test failed. Both subsequent redesigns outsold the original projected first-week numbers by over 200 percent. The data is not subtle.
Zibby Owens is the founder of Zibby Books and Zibby Media. Her podcast, Moms Don’t Have Time to Read Books, releases new episodes weekly.