11 weeks ago I had a blank Figma file and a question.
Today there are 5,200 creators on the waitlist, 12 agencies in LOI, and a category-defining moat that I am betting my next decade on.
This post is the operator notes. The 80 interviews. The bets that worked. The bets that did not. The math that changed mid-build. The version of these 11 weeks that you cannot get from the deck.
I am writing this for the next builder, the next operator, the next founder thinking about a 12th video-tool startup. If any of these notes saves you a month, the post did its job.
Week 1 to 2: the question that broke the curve
The original question was wrong.
I started with "how do I build a better video editor." I spent week 1 looking at CapCut, Submagic, OpusClip, every wrapper that had raised a round in 2023 to 2024. I built a feature list. I started wireframing.
Then I made the move that changed everything. I stopped wireframing and started interviewing.
The interviewing question that broke the curve: "what is the actual number of minutes you spend per branded video, from idea to live post."
The first 12 creators averaged 3 hours 11 minutes. The last 68 creators averaged 3 hours 15 minutes. The number was so consistent across creators, niches, follower counts, and platforms that I knew it was a structural number, not a personal one.
That number is what I was actually building against. Not "a better editor." Not "a faster AI." A different shape of workflow that moved a structural 3 hours 15 minutes to a structural 90 seconds.
The full breakdown of where the 3 hours 15 minutes goes is in this post.
Week 3 to 4: the category insight
By week 3 I had heard enough to see something.
Every creator named a different bottleneck. The editor took too long. The captions were off-brand. The scheduling required platform-by-platform attention. The analytics did not feed back into anything. The brand kit drifted across team members.
Each was a real complaint. None of them was the actual problem.
The actual problem was that the creator was the integration layer between 12 tools. Solving any single tool did nothing. Solving the integration layer changed everything.
This is where the loop thesis emerged. Not "build a better tool." Build the thing that absorbs the integration layer entirely.
This was also the week I realized that "generation is commoditizing" was the most important sentence in the deck. The category was racing to the wrong destination. The full thesis is in this post.
Week 5: the bet that did not work
I built a prototype for what I called "stack hub." A tool that would orchestrate the 12 other tools through APIs.
The idea: you stay in CapCut, Submagic, Later. Stack Hub talks to all of them. You get a unified view without leaving your existing tools.
I burned 4 days on it. It worked. Then I tested it with 6 creators. None of them adopted it.
The reason: it lowered the integration tax but did not eliminate it. The tools still had separate brand kits. Different feature roadmaps. Different export quirks. Stack Hub made the chaos visible but did not actually fix it.
Lesson: a meaningful product change has to remove the underlying problem, not just paper over it. Compromised products do not compound.
Week 6: the Twin Mode insight
The Twin Mode pivot came from interview #54, a real estate creator in Austin named Sarah K who you might have read about elsewhere.
She said: "I would pay anything for a version of me that could shoot when I cannot."
Until that interview, Twin Mode was a feature I was unsure about. The HeyGen and Synthesia categories felt like sketchy territory. I was reluctant to put it on the roadmap.
Sarah K's comment plus 18 follow-up interviews showed me that the ethics framework, not the feature, was the question. Creators wanted twins. They were scared of getting burned. If we built the consent-first version, the demand was massive.
That week, Twin Mode went from "maybe optional" to "the killer feature." The ethics framework that made it work is here.
Week 7: the agency surprise
I had not been targeting agencies. The agency interviews were initially just to validate that the creator product was differently positioned from the agency category.
What happened instead: agencies asked to be on the waitlist before creators did. Of the 22 agency calls I had in weeks 5 to 7, 19 of them ended with "when can we sign up."
The agency math is just better than the creator math. Margin pressure is higher. Scale problems are bigger. The integration tax compounds across N clients. ENCORE solves more pain for an agency than for a single creator.
We restructured the pricing tiers that week. Agency tier got priority. The $349 price point became the second anchor (with Studio at $129 as the first). 12 of those 22 agencies became LOI partners. The agency math is in this post.
Week 8: the survey and the $129 anchor
I needed a price.
The deck math said $99 for Studio. Investors said $149. I had no idea. So I ran a survey to 800 creators (a sample built from the 80 interviews and their referrals).
The survey asked one question with five price points: $29, $49, $79, $99, $129, $149, $199. For each, "would you pay this monthly for the ENCORE Studio tier described below."
The conversion rate at $129 was 63%. The next-best was $99 at 67%. The drop from $99 to $129 was small. The willingness-to-pay was real.
The math: $99 × 67% = $66.33 expected revenue per visitor. $129 × 63% = $81.27. The $129 price wins on expected revenue per visitor by 22%.
We anchored Studio at $129. The 63% number became the marketing line. The full pricing math is in this post.
Week 9 to 10: the waitlist climb
The waitlist went from 800 (the survey audience) to 5,200 in two weeks.
Three things drove it.
One. The 3-hours-15-minutes post on LinkedIn got 412,000 impressions in 8 days. We grew the waitlist by 1,800 just from that single piece of content. The hook was the number. The conversion was the math.
Two. Two agency LOIs went public. Their teams shared the waitlist link with their networks. We added 1,400 creators from agency-driven referrals.
Three. A creator with 740k TikTok followers found the early-access landing page (we did not approach her). She made one video about her own 11pm creator problem. We added 1,200 creators in 36 hours.
Lessons:
- A strong shareable hook compounds faster than paid acquisition
- Agencies are top-of-funnel for creators because creators trust agencies
- Organic creator endorsement is worth 10x paid acquisition
Cost per waitlist signup ended up at $1.40 (paid) and $0.00 (organic). We targeted $5. The build-in-public strategy materially beat the budget.
Week 11: the moat lock
This week I made the most important architectural decision of the build.
The temptation was to ship faster by being a wrapper. Pick one or two video gen vendors. Wrap them. Get to market. Worry about the multi-vendor router and the Brand Brain later.
The decision: ship the moat from day 1, even if it means a longer pre-launch.
We built the multi-vendor router and the Brand Brain into the v1 architecture. We are 3 weeks slower to public launch than the wrapper path would have been. But we are also 12 months ahead of any wrapper that tries to retrofit a moat into their architecture later.
The argument for picking the loop over a better generator is in this post.
What 5,200 means
5,200 creators on the waitlist is not the goal.
It is a leading indicator that the thesis is right. It also means we cannot launch faster than the Twin Mode enrollment infrastructure can handle. The cap is 10,000 for a reason. Past that, the onboarding queue stretches past 5 days, which breaks the 5-minute Twin enrollment promise.
We are intentionally bottlenecked by the promise we will not break. That feels right to me.
What I would tell the room
Three things, in order.
Find the structural number. Not the slogan. The number that nobody is solving. Mine was 3 hours 15 minutes. Yours will be different. Without it, you are just iterating on features.
Build the moat from day 1, even if it costs time. Wrappers die. Loops compound. The architectural difference has to be in the foundation.
Trust the survey data over the investor instinct. I almost shipped at $99 because that is what advisors told me. The survey said $129. The survey was right.
The next 11 weeks will tell us whether the bet works at production scale. I think it will. The math is consistent. The audience signal is real. The category is moving the way we predicted.
If you want to follow along, the Builder Story is here and the waitlist is open until we hit 10,000.