What's Already Settled

You arrived at this sprint with more of the work done than most founders finish in six months. This page records what is decided, so the rest of the document does not spend your money telling you things you already know.

Settled, and we agree:

• The product. Hymind, a personal audio system. Seven connected formats with HY Vision at the core.
• The positioning. "Feel your future before you live it."
• One app, not several. Correct — the connected system is the differentiator.
• The avatar. Maya, 38. Specific enough to write for.
• Pricing. $24.99 monthly, $179.99 annually, no weekly plan. Well judged (§06).
• The free/paid split. Generic content free, personalisation paid. This is also a build order (§09).
• The brand direction. Quiet Precision. Palette set. Explicitly rejecting purple wellness styling and

the visual language of conventional meditation apps — a genuine positioning advantage in a category where everything looks alike.

• Transparency in HY Liminals. Every sentence visible and approved. This is not only ethical, it is

your mitigation for a real store-review risk (§10).

• No outcome guarantees anywhere. Your document is disciplined about describing experience rather

than promising results. That discipline is what keeps you out of trouble at review.

Three things you got right that are worth naming:

The wake-state insight. Listening while walking, cooking, commuting rather than in a dedicated wellness window. It is the most original idea in the concept and nobody in your competitive set is building for it.

The connected-format system. Vision produces Future Memories; Minder returns them; approved statements become Liminals. This is much harder to copy than any single feature, and it is the answer to "just affirmations."

Both platforms at launch. You reached this from the Stella clone evidence before we looked at data. We verified it and you were right (§03).

What this document adds: the numbers, the market reality, and the build path. Everything from here is work you could not have done without paid data and campaign experience — what the category actually costs to enter, what your fair-use policy does to your margin, what will get you rejected at review, and what to build in what order.

Executive Summary

Six decisions, our recommendation on each, and where the detail sits.

1. 🔴 Your fair-use allowance and your voice engine are one decision

Finding. Personal Premium grants three personal creations a day. At premium voice quality that costs roughly $468 per subscriber per year against $339 of revenue — you would lose $180 a year on your most engaged users while profiting on casual ones.

Recommendation. Tier it: premium voice for the first one or two creations daily, budget voice beyond. Preserves quality where users notice, protects margin where it scales.

Detail: §05

2. The economics work — the answer to your July question is yes

Finding. At typical usage, LTV/CAC is 5.7× with payback in about 3.2 months. Even a pessimistic case with a weak funnel and high churn clears 2×. Cold, paid traffic is viable in this category.

Recommendation. Proceed, but instrument scenes-per-subscriber from day one. The whole model rests on that number.

Detail: §05

3. Launch in the US, not Germany

Finding. The US is 7× the UK and 9× Germany in category demand. 82% of German demand sits on a single word. German meditation traffic costs $9.36 a click on 5,400 searches.

Recommendation. US, English, both platforms. Then UK, Canada, Australia — free expansion, same content. Germany only if the numbers later justify a second language.

Detail: §02, §04

4. App store search will not launch this app

Finding. None of your three reference apps won through ASO. Stella ranks #1 for "manifestation" — on 216 searches a month. Only "I am" has real search volume, and it wins on quote keywords, not affirmations.

Recommendation. Treat the avatar channel as the acquisition strategy, not marketing support. Own "manifestation" in the store because it is uncontested and free; compete in "affirmations" patiently over twelve months; point paid spend at manifestation at $1.92 rather than affirmations at $7.03.

Detail: §03, §11

5. Two things can delay your launch, and both are avoidable

Finding. Voice cloning collects biometric data under GDPR — explicit consent, a DPIA and deletion rights are required, not optional. Separately, hypnosis and subliminal framing attracts store review scrutiny.

Recommendation. Scope the GDPR work into the build now and raise it with your lawyer. Keep the disciplined claim-free language you already use, extend it to store metadata and ad copy, and consider "transparent affirmations" rather than "subliminals" in the listing.

Detail: §10

6. Phase the build — using the plan you already wrote

Finding. Your Masterconcept §21 already stages the build. Launching everything at once means guessing at everything at once.

Recommendation. Phase 1: HY Vision, My Context, HY Boost, Minder + widgets, generic library, 8 voices, paywall, analytics — both platforms. Phase 2 (weeks 4–10): voice cloning, Liminals, Shift. Phase 3: full HY³ wake logic, once it always works.

Nothing leaves the concept. It is a sequence, not a reduction.

Detail: §08, §09

The one-line version

The business works. The product is well conceived. Two things need fixing before you build — the generation-cost ceiling and the GDPR scope — and the launch should be American, phased, and carried by your audience rather than the app store.

Market Demand, Measured

Your research established that the category is emotionally engaged and commercially proven. What it could not do — because the data sits behind paid tools — is size the demand and compare markets.

We pulled monthly search volume for 22 category terms across six countries.

Total category demand by market

The United States is roughly seven times the UK and nine times Germany.

What this means for you specifically

Launch in the US, not DACH. This is worth stating directly because the natural instinct — building from your home market where you understand the language and culture — would be the wrong call here. The US market is not slightly bigger. It is an order of magnitude bigger.

Germany's demand is thinner than it appears. Of 60,380 monthly searches, 49,500 sit on the single word "manifestation." Strip that out and the entire German category is around 11,000 searches a month. That is not a market you can build a launch on.

The UAE is negligible at 8,060 searches a month. This matters because you are moving to Dubai. That move makes sense for tax, cost of living and lifestyle — it does not put you closer to your customers. Your customers are in America. Worth knowing so you plan the business around it rather than assuming local proximity helps.

Where the money is being spent

Cost-per-click is a useful proxy for how hard advertisers are competing:

Two things stand out.

"Daily affirmations" at $7.03 in the US is expensive — that is a well-monetised term with established apps bidding hard. It is not where a new app should buy its first users.

"Manifestation" at $1.92 in the US carries 201,000 searches at a quarter of the cost. Higher volume, lower competition. That is where paid acquisition should start, and it aligns with the uncontested ASO position identified in §03.

German meditation traffic at $9.36 per click on 5,400 searches is the worst combination available: expensive and tiny. Another reason DACH is not the launch market.

The honest caveat

This is Google web search data, not App Store search data. It tells you how many people are looking for this topic, which is the right measure for sizing a market and for planning paid acquisition and content. It is not a direct measure of app store discovery — that is handled separately in §03, and the finding there is that app store search barely matters in this category anyway.

ASO Reality Check

The finding: none of your three reference apps won through app store search. Not one.

This changes where your launch effort should go, so it is worth showing the data rather than asserting it.

What we did

We pulled the complete keyword profile for Stella, "I am" and Activations from the US App Store and Google Play: every term each app ranks for, the search volume behind it, and its exact rank position.

One limitation, stated plainly. This data is US-only — the endpoint returns an error for every other country, so no geographic ASO ranking data is available at any price. It also maps web search volume to store rankings; it is not Apple's internal in-store search index. Apple restricted that data for all vendors in late 2025, which is why no tool sells reliable in-store volume today. The genuinely accurate route is running a small Apple Search Ads campaign and reading the impression data back — which you will want at launch anyway. We recommend it as a pre-launch step in §11.

What the three apps actually rank for

Stella — the app you most want to emulate — has almost no search presence. It ranks #1 for "manifestation," but that term carries 216 searches a month in the App Store data. It ranks #1 for "stella app" and "stella manifestation app" — branded terms only people who already know it would type.

Activations is the same story: its best genuine term is "daily motivational" at 115 searches a month, where it sits at position #53.

"I am" is the exception, and how it wins is the interesting part. Its traffic does not come from affirmation keywords. It comes from quotes:

People search for quotes. "I am" captures them, then converts them into affirmation users. That is a deliberate and effective ASO strategy, and it is invisible unless you look at the keyword data.

What this means for Hymind

Stella's $340K a month did not come from the App Store. It came from the founder's audience. If Hymind copies Stella's product without Stella's distribution, there is no traffic — the store will not supply it.

This is the same point raised in July, now with data behind it: distribution decides the outcome in this category, not the build. Your avatar channel is not a nice-to-have alongside the app. It is the acquisition strategy. Everything in §11 follows from that.

The three niches you are spanning

Your competitive set splits into three markets of very different shape:

Nearly every app in the affirmations niche is rated 4.8. It is a mature, well-executed, widget-driven market. Manifestation, by contrast, has essentially no search demand and no dominant player.

The keyword strategy this implies

Rank in affirmations, where the traffic is. Own manifestation, where nobody is competing.

• Affirmations is where the searches are, and where you must earn visibility over time. This is a

long game against established apps with thousands of ratings.

• Manifestation is uncontested. Ranking #1 there is achievable quickly, but it delivers a few

hundred visits a month. Worth owning, not worth relying on.

• Quotes is the largest untapped surface, and "I am" has proven it converts. Worth testing as a

discovery layer even though it sits slightly outside your positioning.

Practical implication for the store listing: your title and subtitle should carry affirmation and manifestation terms, not "Hymind" alone. Nobody searches for a brand that does not yet exist.

The clone problem — your point, verified

You flagged that a copycat took Stella's name on Android. We checked, and you are right.

Searching "stella manifest" on Google Play returns "Stella: Manifest & Affirmation" by Stackwares at rank #1, rated 4.7. The original Stella is not on Google Play at all.

A clone owns the Android search result for the original's own brand name, purely because the original left the platform open.

That settles the platform question. Launching iOS-only leaves the same door open for someone to do this to Hymind, and the name is more distinctive and therefore easier to squat. Launch both platforms together. This costs more up front, and it is the correct decision — you were right about it before we looked.

(We verified the ranking directly. Her reported ~100K download figure for the clone we could not independently confirm, so we cite the ranking, not the number.)

Where to Launch

The recommendation

Launch in the United States. English only. Both platforms.

Why

§02 established the size difference: the US is roughly 7× the UK and 9× Germany in category search demand. But size alone is not the whole argument.

Germany looks tempting because it is home and because "manifestation" shows 49,500 searches. But that single term is 82% of the entire German category. Outside it, demand is around 11,000 searches a month, and German meditation traffic costs $9.36 per click — expensive and small at once.

There is also a content cost. Hymind is an audio product where the words carry the entire experience. Launching in German means writing, voicing and quality-checking every script twice, plus a second voice set. That is a significant multiplier on the content work in §08, for a market a ninth the size.

On the UAE

You are moving to Dubai. The UAE category demand is 8,060 searches a month — roughly 1.5% of the US.

That does not make the move wrong. Tax treatment, cost of living and quality of life are entirely valid reasons, and the Dubai company gives you a clean structure for app store payouts. It just means the move is a business-structure decision, not a market-access one. Plan on serving American customers from Dubai rather than expecting local demand.

Sequencing

Phase 1 — US, English, iOS + Android. All acquisition effort in one market. One language, one content set, one set of learnings.

Phase 2 — UK, Canada, Australia. Free expansion: same language, same content, same store listing with minor localisation. Together they add roughly 165,000 monthly searches for almost no production cost. Do this once US unit economics are proven, not before.

Phase 3 — Germany, if the numbers justify it. Only after Phase 1 and 2 confirm the model, and only with a deliberate decision to fund a full second-language content set. It is the natural market for you personally and the weakest one commercially. That tension is worth naming now so the decision is made on evidence rather than instinct.

One thing to hold on to

Every market in Phase 2 speaks English and shares your content. That is unusually cheap expansion. Resisting the pull to launch in German first is worth roughly a ninefold difference in addressable demand for the same production effort.

Unit Economics

The question you actually bought: can you acquire a user for less than they are worth, and how long until the money comes back?

You asked this in a different form back in July, when you said Stella and Activations grew on warm traffic and wondered whether the same product converts on cold, paid traffic. That was the sharpest question in the whole conversation. Here is the answer, with the working shown.

How to read this

Every input below is either sourced or flagged as an assumption. Where we could not find reliable data, we say so rather than fill the gap. You will recognise this discipline from media buying: a model built on invented inputs produces confident nonsense.

The inputs

Blended CPI: $3.96. Install → paying subscriber: ~10%. CAC per paying subscriber: $39.65.

The result

The answer to your July question is yes — the maths works on cold traffic. At typical usage, a subscriber is worth roughly 5.7 times what they cost to acquire, and pays back in about three months. That is a healthy business by any standard; anything above 3× is considered good.

Even the pessimistic case (D), where the funnel underperforms and churn runs high, still clears 2×. The model is robust to bad luck.

🔴 But there is one thing that breaks it

Look again at Scenario C.

Your Masterconcept grants Personal Premium users three personal creations per day — up to 90 per month. At premium voice quality, generating those costs roughly $468 per subscriber per year against $339 of revenue.

You would profit on casual users and lose $180 a year on every engaged one. The people who love Hymind most would be the ones costing you money. That is the opposite of how a subscription business should behave, and it is invisible until you are at scale and the bill arrives.

This is not a flaw in your thinking. Your document explicitly says the thresholds will be "validated against real production costs." You anticipated the question correctly. You just did not have the number. Here it is.

The break-even ceiling

Holding LTV/CAC at 3× or better:

Your 3-per-day allowance is safe at any cost up to about $0.10 per scene. Above that it is not.

Three ways forward — your call, not ours

1. Budget voice for personal generation. Three per day works comfortably at 5.4× LTV/CAC. The cost

is audio quality, which matters in a product where voice is the experience.

1. Keep premium voice, reduce the allowance to one per day. Protects margin, but weakens the core

promise of unlimited personalisation.

1. Tier it. Premium voice for the first one or two creations a day, budget voice beyond that. Best

commercial outcome, more build complexity. This is our recommendation — it preserves the quality of the experience where users notice it most, and protects margin where they do not.

The voice engine and the fair-use policy are not two decisions. They are one.

What actually moves the outcome

We tested each variable independently:

• Generation cost dominates everything. Switching voice engines moves LTV/CAC from 5.7× to 7.0× at

typical use — and from −4.5× to +5.4× at your stated cap. Nothing else comes close.

• Churn matters more than acquisition cost. A worse funnel and worse churn together still clear

2×. Runaway usage cost does not.

• The 15% Small Business rate is worth about $17 per subscriber versus the standard 30%. Enrol

before launch, not after.

Caveats we owe you

• Install → trial at 25% is our assumption, not measured data. RevenueCat's widely quoted 89%

trial-start figure is per session, not per install, and quoting it here would flatter the model.

• Non-US CPI has no tier-1 source. Any figure outside the US carries a +15–25% assumption. Treat

§04 geography as directional.

• Twelve scenes per month as "typical use" is an estimate. It is also the number the entire model

hangs on. Instrument it from day one — actual usage per subscriber is the single most valuable number you will collect in your first month, and it will tell you within weeks whether the allowance is set correctly.

Pricing and Packaging

Your pricing is right

$24.99 monthly and $179.99 annually sits correctly against the market, and the reasoning you gave on the call was sound. For reference:

You are positioned just below Activations and well above "I am." Given Hymind does more than either, that is defensible.

Your decision to skip weekly pricing is correct. Weekly billing suits utility apps with short usage windows. For a habit product it produces high churn, frequent payment friction, and a subscriber who re-evaluates the cost 52 times a year rather than once or twice.

The annual discount deserves a deliberate decision

$179.99 annually against $24.99 monthly is a 40% discount — you are pricing the year at the equivalent of $15/month.

That is on the generous side of normal. The tradeoffs:

In favour: RevenueCat's data shows 68% of health and fitness subscribers choose annual when offered, and annual subscribers churn far less — 2.2 years average life versus 8.3 months on monthly. Cash arrives up front, which matters when you are funding generation costs.

Against: a subscriber who would have paid $24.99 monthly for eight months ($200) instead pays $180 for a year. The discount is buying commitment you might have got anyway.

Recommendation: keep it. The retention difference outweighs the discount, and the up-front cash materially improves the payback picture in §05. But treat it as a decision you have made, not a default, and revisit it once you can see actual monthly-versus-annual behaviour.

Trial mechanics

RevenueCat's data points to a clear answer: 17–32 day trials convert best, at around 45.7%, versus 39.9% for the median.

Recommendation: a 21-day free trial on the annual plan, with the monthly plan available without a trial. Long enough for the habit to form and for the personalisation to accumulate context — which is precisely when Hymind gets better and cancelling gets harder.

Pair it with the §07 flow: the paywall appears after the first personalised piece, never before.

Win-back

Stella shows a $25 one-month offer when a user tries to cancel. That is a well-judged mechanic and worth copying in structure.

Recommendation: on cancellation, offer a reduced-price month rather than a discount on the annual. The goal is to keep the habit alive through a wobble, not to permanently reprice the customer.

What to watch after launch

• Monthly vs annual split. If annual runs well below 68%, the discount is not doing its job.
• Trial-to-paid rate. Below 30% points at the onboarding flow, not the price.
• Scenes generated per subscriber. The §05 model hangs on this. If it runs far above 12/month,

revisit the allowance before it becomes expensive.

Onboarding Architecture

You raised this on the call and it is the right thing to worry about. You found that Stella's onboarding runs 25–30 minutes and draws complaints, and you asked how to collect enough for personalisation without making people work for it before they have felt anything.

The tension, stated properly

Personalisation needs data. Data collection costs patience. Patience is what a new user has least of.

Stella resolved it in the wrong direction: ask everything up front, then reveal the price at the end. That maximises data and sunk-cost feeling, and it is why the reviews complain. Worse, it has since been copied by the apps in your competitive set — which means doing it differently is itself a differentiator.

What the data says

RevenueCat's benchmark data shows a paywall presented at a value moment converts 2.1× better than a hard paywall shown before the user has experienced anything.

That is the whole answer. Not "shorter onboarding" — value first, then the ask.

The recommended flow

Screen 1 — Choose one thing you want. Six tiles: money, confidence, relationships, career, calm, self-worth. One tap. No account, no email, no questions.

Screen 2 — Hear something immediately. A generic but genuinely good 60–90 second HY Boost or a short generic HY Vision on the chosen theme. This is the moment the product has to land. Everything before it is friction, everything after it is easier.

Screen 3 — The honest offer. "That was our generic version. Hymind can write these about your actual life — your home, your work, the people in it. Want to tell us a little?"

Now the user has felt the difference between generic and personal, so the questions have an obvious purpose. This is where Stella loses people and where you can win them.

Screens 4–6 — Progressive profiling, three questions maximum. Only what the chosen theme needs. Someone who picked "career" does not need to answer questions about their home.

Screen 7 — The personalised piece. Generated from what they just gave. This is the second value moment, and the strongest possible position from which to show a price.

Screen 8 — Price, clearly, with the free tier visible.

Then: continuous context building. Every subsequent creation asks for one more detail. After a month the profile is richer than a 30-minute questionnaire would have produced, because each question arrived when it was obviously useful.

Why this beats the alternative

Your Masterconcept already describes this. §13 says value before intimate questions, and phrases the invitation as "Would you like Hymind to know your life well enough to make this yours?" That is the right instinct and the right sentence. This section is the operational version of what you already wrote.

Free tier design

The purpose of the free tier is not to be generous. It is to demonstrate the gap between generic and personal clearly enough that upgrading feels obvious.

Give away: a real library of generic content, unlimited replay, widgets and reminders. Enough to build a daily habit.

Hold back: anything that uses their life — personal Visions, personal Liminals, the voice clone.

This is exactly the split you specified. It works because the free product is genuinely useful and the paid product is obviously different. A free tier that feels crippled produces uninstalls; one that feels complete produces no upgrades. Yours sits in the right place.

What to measure from day one

• Drop-off per onboarding screen
• Percentage reaching the first audio (the value moment)
• Percentage who accept the personalisation invitation
• Trial start rate by entry theme
• Scenes generated per subscriber per month — the number the §05 model depends on

Instrument these before launch, not after. They are cheap to add during the build and expensive to retrofit.

Content Library Sizing

Your other open question from the call: how many pre-recorded sessions and categories should exist at launch.

The principle

Launch with enough that the free tier feels complete, and no more. Every additional generic session costs money and time before you have any evidence about what people actually use. Once real usage data exists, you will produce far better content far more cheaply, because you will know what people choose.

What production actually costs

Using the generation costs established in §05:

Generation is not the constraint. Even a large library costs less than a hundred dollars to synthesise. The real costs are scriptwriting, review and quality control — particularly for HY Deep, where scripts need professional hypnotherapy review before release.

That reframes the decision. The question is not "what can we afford to generate" but "how many scripts can be written well, reviewed properly, and quality-checked before launch."

Recommended launch library

Categories: six at launch, not sixteen. Money, confidence, relationships, career, calm, self-worth. These match your onboarding themes and the highest-volume search terms from §02. Add categories once usage shows which ones people actually pick.

Why fewer than your reference apps

Activations has 700+ sessions. "I am" has 1,000+ themes. Those libraries were built over years, funded by revenue, guided by usage data.

Matching them at launch would be the most expensive possible way to guess. You would spend months producing content before knowing which themes people choose, then discover that a handful carry most of the usage.

Your differentiator is not library size — you cannot win that race, and §03 shows the apps that tried are not winning on discovery anyway. Your differentiator is personalisation. The generic library exists to demonstrate quality and seed the habit. It is the shop window, not the shop.

A cost lever worth knowing

Consider using premium voice for the generic library and budget voice for high-volume personal generation. The generic library is heard by everyone, is produced once, and defines the perceived quality of the product — spending $25 rather than $4 there is trivially worth it.

Personal generation is where volume costs accumulate, and where a user is hearing content about their own life. Attention is on the words rather than the polish of the delivery.

This gives you premium perceived quality where it is judged and controlled cost where it scales.

What to Build First

The tension, named honestly

On the call you said you would love to launch with all components, and your reasoning was good: "I like I am the app, but it's just affirmations." You believe the connected system is the product, and that shipping a subset makes Hymind just another affirmations app.

You are right about that, and this section does not argue otherwise. Nothing is being removed from the concept.

The only question is what your first users can hold in their hands, and what follows once it is earning.

Your own document already answers this

Section 21 of your Masterconcept reads:

"Hymind is built in deliberate stages. First priority goes to a high-quality generic experience, HY Vision, My Context, voices, Voice Clone, library, transcript, clear pricing and cost measurement. The full public core, personal extensions, reliable HY³ wake logic and controlled referral and one-time-purchase tests follow."

That is a phased build plan, and it is a good one. It was written by you. What follows is that plan with costs, sequence and reasoning attached.

Why phasing is the winning move, not the compromise

Stella launched focused and reached roughly $340K a month within two months. It did not launch complete. It launched with a strong core and a founder who could put it in front of people.

The apps in your competitive set that launched with enormous libraries — Activations with 700+ sessions, "I am" with 1,000+ themes — built those over years, funded by revenue.

Launching everything at once means guessing at everything at once. Phasing means the parts you build second are informed by how people used the parts you built first.

The build order

Phase 1 — Launch (both platforms, simultaneously)

The core loop, complete and excellent.

Phase 2 — Weeks 4–10 after launch

Phase 3 — Once retention is proven

On voice cloning specifically

You confirmed it stays in, plus eight selectable voices. Those are two separate line items with very different costs.

The eight stock voices are comparatively cheap — licensed TTS, integrated once, no per-user overhead. They belong in Phase 1.

Per-user voice cloning is a different order of work: consent flow, secure storage, per-user model handling, moderation, GDPR obligations, and an ongoing cost per user. It belongs in Phase 2, not because it is unimportant but because it is the single most complex thing in the product and it should not delay revenue.

A user hearing their affirmations in one of eight beautiful voices is not a compromised experience. Their own voice, arriving six weeks later as a headline update, is also excellent marketing.

Both platforms, together

Confirmed by the clone finding in §03. Launching iOS-first leaves the Android door open exactly as Stella did, and "Hymind" is more distinctive and therefore easier to squat than a generic term.

This raises Phase 1 cost. It is still the right call, and you identified it before we looked at the data.

Risk Register

These are the things that can delay your launch or cost you money, ordered by how much damage they do. Most are avoidable if handled during the build rather than discovered at review.

🔴 HIGH — Voice cloning is biometric data

Under GDPR, a voice recording is special-category biometric data (Article 9). It is not treated like an email address. Because you are currently EU-based and will have EU users regardless of where you move, this applies to Hymind.

What is required, not optional:

• Explicit opt-in consent before any recording, separate from general terms acceptance
• A Data Protection Impact Assessment completed before launch
• A stated retention period, and the ability for a user to delete their voice data in-app
• Disclosure of where processing happens and which processor handles it

Cost of getting it wrong: app rejection at review, and genuine regulatory exposure afterwards.

Recommendation: budget for this in the build and raise it with the lawyer already handling your trademark. It is a known, solvable requirement — but it needs to be scoped now, not discovered later.

🔴 HIGH — Health and therapeutic claims

Both stores restrict apps that imply treatment of medical conditions. Hypnosis, subliminal audio and "rewiring limiting beliefs" all sit close to that line.

What triggers rejection: claiming the app treats anxiety, insomnia, depression or trauma. Framing hypnosis as therapy. Promising specific outcomes.

What you are already doing right: your Masterconcept explicitly avoids outcome guarantees, states the app does not replace professional support, and describes what the listener experiences rather than what it will fix. That discipline is exactly correct and should carry into your store listing, your avatar channel content and your ads — the store reviews your marketing, not only your app.

Watch: the "brainwashing" and "reprogramming" language on your brand board. It works as social copy; it reads badly to a store reviewer.

🟠 MEDIUM-HIGH — HY³ overnight reliability

This is not a policy risk. It is a technical one, and you flagged it yourself.

Playing audio through a locked night and waking a user reliably fights both operating systems. iOS low-power mode and Android Doze can interrupt background playback. It works, but not with the certainty an alarm clock needs.

Mitigation: build a native alarm as the guaranteed fallback so the wake always happens even if audio is interrupted. Set expectations in the UI. Do not market HY³ as a fully automatic core feature until it has been tested on locked devices, in low-power mode, on both platforms. Your own document says exactly this — it is right, and worth protecting during the build when the temptation is to ship it early.

🟠 MEDIUM — Subliminal audio positioning

Historically, "subliminal" content attracts scrutiny as potentially manipulative because the user supposedly cannot perceive it.

Your design already solves this. Every sentence is visible, editable and approved before playback. Nothing is hidden. That is a genuine, defensible distinction.

Recommendation: make the transparency explicit in store metadata, and consider naming the feature "transparent affirmations" rather than "subliminals" in the listing itself. Keep "HY Liminals" as the in-product name — it is good branding. The store listing is a different audience with different sensitivities.

🟠 MEDIUM — AI content disclosure

Requirements for labelling AI-generated content have tightened across both stores and under EU rules. Hymind generates audio with AI and, with voice cloning, does so in the user's own voice.

Mitigation: label AI-generated scenes in-app, disclose AI processors in the privacy policy, and show a clear notice before the first generation. Low cost if designed in, awkward to retrofit.

🟡 LOW — Manifestation framing

Manifestation and law-of-attraction apps ship on both stores routinely. The risk only appears with guaranteed-outcome language — "manifest anything," "the universe responds."

Your positioning already avoids this. No action needed beyond keeping marketing consistent with the product.

🟡 LOW — Subscription mechanics

Your pricing, free tier and cancellation model are compliant as designed. Ensure enrolment in the Small Business Programme before launch — it reduces store commission from 30% to 15% under $1M in revenue, worth roughly $17 per subscriber in the model in §05.

A note on what we could not verify

Some published guidance on AI-disclosure enforcement dates could not be confirmed against primary sources. We have presented the risks and mitigations rather than cite dates we cannot stand behind. Before submission, have whoever handles your legal work review the current text of both stores' policies — they change, and the version at submission is the one that matters.

Go To Market

The strategic conclusion from §03

None of your three reference apps won through app store search. Stella won on the founder's audience. Activations won on a recognisable narrator. "I am" won by capturing quote-seekers at enormous scale over several years.

In this category, distribution decides the outcome — not the build, and not ASO.

Which means the avatar channel you have already started is not marketing support for the app. It is the acquisition strategy. Everything below follows from that.

Now until launch (2–3 months)

This is the highest-leverage period, and most of it costs nothing but consistency.

Build the audience before the product. Stella's advantage was an existing audience at launch day. You are doing the right thing already by producing hypnosis audios for the avatar account before the app exists. Keep going, and increase frequency.

Publish the format, not the product. Short HY Boosts and HY Shifts for specific everyday situations are the content. People experience the actual product mechanic for free, repeatedly, before there is anything to buy.

Start the waitlist now. Every piece of content should have somewhere to send interested people. A single page and an email capture is enough. A waitlist that hears from you weekly through launch is worth many times one collected in the final fortnight.

Tag every signup by source. Avatar channel, organic, paid — you need to know which audience converts to installs, not just which grows fastest.

Run a small Apple Search Ads campaign before launch. Two purposes: it gives you real Apple keyword impression data, which no third-party tool can sell you (§03), and it tests messaging cheaply.

Launch

Store listing built on §03. Title and subtitle carry affirmation and manifestation terms. "Hymind" alone earns nothing — nobody searches for a brand that does not exist yet.

Own manifestation immediately. It is uncontested. Ranking #1 is achievable within weeks. It delivers modest volume, but it is free and defensible.

Compete in affirmations patiently. That is where the traffic is (§03) and where established apps with thousands of ratings sit. This is a twelve-month project, not a launch-week one.

Point paid acquisition at manifestation, not affirmations. §02: "manifestation" carries 201,000 US searches at $1.92, while "daily affirmations" costs $7.03. Same category, a quarter of the price.

After launch

Prove the model before scaling spend. §05 gives a payback of roughly three months at typical usage. Do not scale acquisition until real cohort data confirms the assumptions — particularly scenes generated per subscriber, which is the number the entire model depends on.

Then scale what already works organically. Content that performed on the avatar channel is the best possible source of paid creative, because it has already been validated by an audience.

Micro-influencers after launch, not before. Additional reach without their voice becoming part of the product, exactly as your Masterconcept specifies.

What we would not do

Do not buy installs at launch to look successful. Vanity installs from cold traffic to an unproven funnel burn cash and teach you nothing.

Do not launch in German first. §04.

Do not depend on ASO for launch traffic. §03 shows the category does not work that way. Treat any store traffic as a bonus.

The Next 90 Days

Assumes a build start in early September and a launch window in late November or early December. Your instinct of 2–3 months was right, and this is what has to happen inside it.

Weeks 1–2 · Foundations

Weeks 3–6 · Core build

Weeks 7–9 · Completion and hardening

Weeks 10–12 · Launch

After launch — the first real decision point

Around week 4 post-launch you will have enough data to answer three things:

1. Is scenes-per-subscriber near 12/month? If materially higher, revisit the allowance before it

becomes expensive.

1. Is trial→paid near 40%? Below 30% points at onboarding, not price.
2. Is D7 retention above 10%? This is the number that decides whether the model works at all.

Phase 2 — voice cloning, HY Liminals, HY Shift — starts once those answers are in hand, not on a calendar date.

What could move the timeline

Realistic risks, not pessimism:

• Store rejection on wording — budget one cycle (§10)
• HY Deep professional review taking longer than expected
• Voice cloning consent and DPIA work, if pulled into Phase 1 rather than Phase 2

Your position was that quality matters more than speed. That is correct, and the phasing in §09 is what makes both possible: a smaller Phase 1 done properly, rather than everything done adequately.

Assumptions and Sources

Every number in this document is either sourced or flagged as an assumption. Where data does not exist, we say so rather than estimate and present it as fact.

Data we pulled ourselves

Published benchmarks

Assumptions — our judgement, not measured data

These are the numbers to challenge, and the ones to replace with real data as soon as you have it.

Limitations we want to be explicit about

App store keyword data is US-only. The endpoint returns an error for every other country. No geographic ASO ranking data was available at any price, which is why §04 is built on search volume rather than store rankings.

Search volume is web, not in-store. Apple restricted in-store search data for all vendors in late

1. No tool currently sells reliable App Store search volume. The accurate route is Apple Search Ads,

recommended in §11.

Stella's revenue figures are unaudited. The ~$340K/month figure comes from a published interview, not verified accounts. The implied revenue per subscriber is far above category norms, so we have used it as directional evidence only and modelled on category medians instead.

The clone's download count is unverified. We confirmed the Google Play ranking directly. The ~100K figure we could not independently confirm, so §03 cites the ranking rather than the number.

Store policy interpretation is not legal advice. §10 reflects published guidelines and observed practice. Policies change, and the version in force at submission is the one that matters. Have your lawyer review current policy text before you submit.