References & working notes

Audience direction

Owner direction, 13 October: adults with a Chinese-speaking partner or Chinese heritage have a personal reason to learn. Repeatedly stopping is the shared problem. Other restarting learners are an adjacent audience.

This is a working hypothesis, not a validated customer segment. Auntie Lin is a created hook, not proof of audience fit.

Joseph Choi · 13 October

Start with a specific customer problem. Work closely with a few credible creators. Test formats systematically. Separate creator performance from format performance. Use customer results to judge the work, alongside views. Increase spend after a format converts repeatedly.

The deck’s 2–5 creator test and Content Rewards sequence are our proposal based on these lessons. They are not a campaign commitment or a direct instruction from Joseph.

Brain source: references/marketing/joseph-choi-organic-ugc-workshop-2026-10-13. The original transcript contains repetitions and local timestamp resets.

Content Rewards rep · 8 October

The rep recommended self-service and building our own operation for the proposed starting budgets. The rep said the existing brief was suitable. The rep advised an active Discord community and close work with creators. Agency work includes ideas, formats, sourcing, and submission review.

The earlier plan covered several app campaigns. The revised deck proposes a focused Lock In Chinese test first. We want Kaitlin’s advice on the right place for Content Rewards in that test.

Brain source: meetings/2026-10-08-content-rewards-rep-call. This is a summary of the recorded call, not a current platform pricing claim.

Existing UGC work

The herdr UGC workspace contains work on the “Study with me” creator guide. It includes real app clips, three edited examples, recording instructions, and a post check. A separate discussion covers the wider agency plan and naming.

The introduction uses the 3.3-second Dynamic Island section. The quiz is an alternate example. The owner rejected the Lock Screen and word-walk selections. The audience slide uses a silent 3.65-second door hook. The audience slide uses lin_peep_C.mp4 from the archived creative assets.

Tool status

The research slide uses the owner’s screenshot of the live feed, supplied on 13 October. The feed contains 2,456 posts. Lab analysis is separate from collection and review. A batch of 12 previously liked or shortlisted language-learning posts is requested for analysis. Completion must be checked separately.

The outreach rollout records the core work as built. Preparation and live checks remain before the first send. Dash owns creator records, drafts, and reply review. Instantly is the selected delivery service.

Dash links require an authorized sign-in.

Repo sources: docs/playbooks/creator-outreach-rollout.md; docs/superpowers/specs/2026-10-09-attention-lab-design.md; docs/superpowers/plans/2026-10-12-attention-lab-phase-4-lab-pages.md.

Ideas to test later

Owner update: a small farm at home is being set up. Its scale, equipment, and operating status are not specified.

Owner clarification: the idea behind slide 3 comes from seeing formats repeated across large UGC campaigns. A generator could help many creators make variations after we identify a winning format.

The supplied links identify Hevy, not Strong: Hevy Clipping [Slideshow] by Clipvex and Hevy CTA Slide Generator. The generator offers leaderboard, workout history, gym comparison, PR tracking, and streak formats.

These are external references. Their existence does not prove conversion performance. Our proposed sequence is to validate a format first, then explore templates and wider creator production.

Founder network and the wider plan

The 8 October Content Rewards meeting records Neco’s intention to build an operation focused on mobile app distribution. The plan starts with his own apps, then extends to founders in Hacka Network. The owner corrected the network name from the meeting transcription.

Owner clarification for the Kaitlin call: “I want to be the one who solves distribution. Once I get this going, I will bring more apps and projects.” This describes the intended next step after proving the process, not committed customers.

The meeting also names Pretty Progress and other possible projects. The deck keeps the first campaign focused on Lock In Chinese.

Source: meetings/2026-10-08-content-rewards-rep-call, plus the owner’s current direction.

Review against gbrain · 13 October

The gbrain review covers the original eight slides. A ninth slide now adds research-based content proposals. The revisions make the product mechanism, creator test, outcome measures, and requested call decisions more specific.

• Product: vocabulary on the phone and a quiz before selected apps. This does not promise conversation fluency.
• Audience: the initial overlap is a personal reason to learn plus repeatedly stopping. Partner and heritage messages remain separate hypotheses.
• Creative: credible people speaking directly to camera are a proposed first test. Auntie Lin remains an AI hook experiment, not a validated customer format.
• Measurement: views, installs, useful app activity, and purchases serve different purposes. Keep the same promise through the video, store page, and app. Establish attribution before claiming a video caused purchases.
• Email: the rollout record marks the build complete, with draft review and live sending checks still to complete. It does not establish a live campaign.
• Management: daily creator feedback and an owned Discord community follow the Content Rewards call. Paid trials are proposed, not approved employment terms.
• Sideshift: Joseph’s workshop offer mentions three months of access. Confirm written terms with Kaitlin before treating this as included access.
• Call outcome: select an audience, two initial formats, a hiring plan, and a review date.

Gbrain sources: references/marketing/joseph-choi-organic-ugc-workshop-2026-10-13; meetings/2026-10-08-content-rewards-rep-call; playbooks/consumer-app-growth/promise-continuity; docs/playbooks/creator-outreach-rollout; playbooks/creator-cold-email-outreach-learnings; playbooks/creator-research-and-outreach-workflow.

Latest owner corrections govern: Hacka Network is the correct name; the rejected Lock Screen and word-walk extracts stay excluded. The phone farm is an early setup, not a proven distribution channel. Network size and prospective projects are not presented as committed customers.

New format selection · replaces the first examples

The owner rejects the first selection as insufficiently repeatable. The new search requires videos under 30 seconds, a clear opening, and a real product mechanism that fits Lock In Chinese. See seven formats, timings, variation rules, and local video examples.

The earlier selection below remains a research record. It is no longer the recommended deck content.

Product–content fit · Dash evidence

Added 13 October. These three proposals combine saved Dash examples with the current audience direction and Joseph’s test approach. They are adaptations, not proven conversion formats. Draft openings are our proposed copy, not quotes from creators or customer testimonials.

1. Personal reason. @cantowitheunice is owner-liked. Completed Lab analysis identifies a partner-specific opening followed by useful phrases. The source teaches Cantonese. Our adaptation concerns Mandarin and the habit of reviewing words, not a promise of fluent family conversation. Use a creator whose partner story is true. Test heritage separately.
2. Question reply and routine. @thisrromericanlife is owner-shortlisted; the owner specifically likes the question widget. @learningwithsirine is owner-shortlisted. Completed Lab analysis identifies numbered tips, a product recommendation, and an app overlay. Borrow that sequence, not the source’s four-month fluency claim. A genuine user can show how Lock In fits their routine. Label a drafted question as a question, rather than fabricating a viewer comment.
3. A visible rule. @freya.loren is owner-shortlisted with “Format we can copy” and “Fits Lock In.” @christinewrutzen is agent-shortlisted for the sequence of long avoidance, short action, and completion. The latter has no completed Lab analysis. Our proposed adaptation shows the actual restriction, quiz, and app access. It is not a claim that either reference uses Lock In Chinese.

Suggested first test: choose the partner audience and compare the personal-reason opening with the question reply. Keep the creator, product proof, and call to action consistent. Test the visible-rule version next. Measure installs and useful app activity alongside views; assess purchases when attribution is available.

Saved research feed · Lab analysis

Evidence: Dash shortlist records and current source analysis, read 13 October. Automated Lab observations describe sampled video frames and available audio; they do not establish effectiveness. Public engagement counts are omitted because they do not demonstrate product–content fit.

Existing campaign brief and research · 13 October review

Live creator page · Full creator brief · Recruitment kit · Campaign research

What is already written

• A 30-day “Day N locked in on Mandarin” series. The five shapes are shield moment, quiz fail, reframe, study with me, and receipts.
• Five approved videos each week, distributed across TikTok, Reels, Shorts, and Facebook. The trial covers five days and five shapes.
• $25 per approved video regardless of views. Published milestones add bonuses, measured seven days after posting. The brief also describes a separate future revenue pool.
• Real app use, original footage, licensed audio, disclosure, and no invented results or AI faces and voices. Editing and captions may use AI.
• A shot sequence, 15 hooks, approval criteria, a reshoot path, a submission process, and a first-week checklist.

Where it differs from the current direction

• Audience. Dash’s standing brief still targets students aged 16–25, guilty doomscrollers, and efficiency hackers. The current proposal targets adults who repeatedly stop, with a Mandarin-speaking partner or Chinese heritage. Research collected under the old brief is useful, but does not validate the new audience.
• Format. The existing brief requires a Day N series before testing. The deck proposes testing two openings with a few credible learners before choosing a repeating format.
• Test design. The old trial changes five video shapes across five days. That can identify ideas, but it does not isolate the reason a video works. The proposed test holds the audience, creator, product proof, and call to action steady while changing the opening.
• Measures. The old checkpoints rely heavily on view thresholds. The deck adds installs, useful app use, and purchases where attribution permits. A view threshold alone cannot establish product–content fit.
• Conflicting copy. The recruitment kit says “No face required”; the full brief requires a face in each video. The kit also calls the full brief public, while the live page says acceptance provides it.
• Product details. Both the live page and full brief say the app has not launched. The brief gives precise quiz and feature claims. Check these against the build creators receive before reusing the copy.
• Separate experiments. Auntie Lin and the later generator idea are separate proposals. The existing creator campaign excludes AI faces and requires actual app use.

What the campaign research contributes

The saved study compares Lorcana Scan, Trepa, Bluffs, Keymagine, Takeback, and Hevy. Its main operational lesson is to separate the recruitment page, governing brief, format examples, and resources. Takeback is the closest recorded reference for real use and original footage. Hevy supplies the reference for a repeatable slideshow recipe and generator.

For our smaller initial test, retain clear approval and payment rules, specific filming instructions, two or three finished examples, and one resource index. The historical checklist is not a current completion report: for example, the current brief already includes the day-seven bonus rule that the older research lists as missing.

Proposed revision: keep the useful operating material; replace the broad audience and mandatory 30-day format with the focused learner test. Resolve the face rule, product details, and distribution requirements before sending a revised creator pack. This review does not change the public campaign, payment terms, or Dash standing brief.

Sources read: current local creator brief and recruitment material, live creators.nokta.dev page, Dash UGC research brief v1, and the saved campaign-material study in gbrain. Historical external campaign counts and rates are not treated as current.

Later plan: paid ads and an internal agency

Owner direction, 13 October 2026: paid advertising is a definite part of the plan. Build an internal marketing operation with its own team, process, and tooling wherever practical. The purpose is to run more tests, collect useful data, and improve marketing decisions. Consulting income is not the objective. “100%” describes commitment to paid ads; it is not a reimbursement rate.

Research tools already exist. Creator operations, paid acquisition, and measurement form the proposed next stages. Test promising creative with paid traffic and judge results against installs, useful app use, and purchases. An organic winner is a candidate for a paid test, not proof of profitable acquisition.

Owner clarification: the agency structure is a practical way to collect evidence. More apps and projects create more opportunities to test ideas. More observations alone do not guarantee accuracy: use consistent measures, record the variable changed, repeat tests, and retain app and audience context. Treat patterns from other apps as hypotheses to test.

Turkish advertising support

The Ministry of Trade lists advertising, promotion, and marketing support under its service-sector programme, with pre-approval and support application forms. Official programme page and forms ↗ (checked 13 October 2026).

The programme supports up to 70% of qualifying spend in designated target countries: Decision 10962, Article 17 provides a 50% base; Article 44 permits an increase, and circular Article 102 applies 20 additional percentage points to covered activities. For multi-country activities, all countries must be on the designated list to receive the increase. Circular Article 26 covers eligible overseas promotion for software, games, and mobile apps and places the application after payment. This is reimbursement of qualifying expenses, not instant cashback on all cash spend. We have not established Nokta’s eligibility or payment schedule. Confirm the applicant company, covered app and services, target markets, eligible expenses, approval sequence, and evidence requirements before committing costs in reliance on support. The deck treats reimbursement as a later opportunity, not approved funding.

Official Decision 10962, Articles 17 and 44 · Ministry circular, Articles 26 and 102. Checked 13 October 2026. Article 17 specifies annual support caps of TRY 15 million per product and TRY 50 million per beneficiary, up to 10 products and five years; confirm applicable updated limits when applying.

Turkish studio examples

Royal Match and Toon Blast provide concrete context, not a claim that Turkish studios control the App Store or that subsidies cause their success. Dream Games identifies its Istanbul office and global hit games, and links reporting on Royal Match leading app-store charts. Peak’s games page describes Toon Blast among leading US iOS games. These are publisher descriptions, not a live chart audit. Artwork comes from the linked publishers’ websites.

Technical approach: own the operating system

Owner direction: use a technical product-building background to solve individual marketing and UGC operating problems quickly, automate repeat work, and design for scale. The purpose is more useful tests and evidence. Audience, creative, and customer acquisition still require validation.

The deck overlays the actual Dash interface and a rendered image of the supplied outreach blueprint. A comparison slider reveals the two layers. The interactive original remains available through a separate link. Open the portable blueprint. Source: creator-outreach-p2 worktree, docs/superpowers/specs/2026-10-09-creator-outreach-blueprint.html.

The blueprint is a dated architecture snapshot, not a fresh production audit. It marks checks, dispatcher, API integration, event processing, and scheduled operations as built. It marks mailbox setup and the main switch as pending. Import, verification, drafting, reply options, and later automation stages remain planned in this snapshot. The original links to implementation PRs remain clickable. No email is sent by this deck.