- Cold
- never heard of it
- Reached
- the post was on their screen
- Aware
- registered that it exists
- Wishlist
- wants it
What that year cost and bought
One lived-through year with the settings above. Reroll the luck and it changes — that is the point, and the distribution below says by how much.
The arithmetic of it
Run the simulation to see where this lands.
The year, week by week
Wishlists accumulating. Flat stretches are the normal state; the steps are the weeks something landed.
The same plan, 300 times
You get to live one of these. The spread is not error — it is the actual range of outcomes for the exact plan you set.
One video, all the way down
The whole model is this, repeated. A thousand views sounds like a room full of people. Here is what is left of it by the end.
- 1000views
A median clip from a new account. Not a bad one — a normal one.
- 60actually registered a game exists
The rest scrolled past. The video played, the thumb moved, nothing was stored.
- 5.4opened the store page
The only step you fully control from here. Everything above is the algorithm; everything below is your capsule, your trailer, your first three screenshots.
- 1.1wishlisted it
One person. Sometimes two, often none — the number is small enough that luck is louder than skill.
- 0.86told a friend
One in a hundred of the people who noticed mentions it to somebody, and they mention it to two people, not to an audience.
A thousand views is one wishlist. That is not pessimism, it is the exchange rate — and it is why the answer to “how do I get noticed” is a number of posts, not a trick.
The tail is real, and it is not a plan
Reach per post is not a number, it is a draw. The distribution is skewed hard: the median post does what the median post does, and roughly one in a few hundred goes somewhere else entirely.
Every post you make this year, sorted by how far it went. The bar on the left is your normal day. The sliver on the right is the video people will later say changed everything.
Post daily for a year and you will probably get one clip past a hundred thousand views. The simulation says so, and so does everyone who has actually done it. What the simulation also says is that it will not save you: a hundred thousand views on a scroll feed is a few hundred people who registered the game and a few dozen wishlists. The spike is real, the aftermath is a Tuesday.
This is the part that breaks people. They are told to make one great video, they watch someone else’s great video do a million, and they conclude they are bad at this. The honest version: the outcome you are aiming at sits in the top percent of a distribution you get one draw from per post. The only variable you control is how many draws you take — and how far right you can shove the whole curve by making the game itself more worth filming.
The chain breaks on the first step
“One person tells a friend, the friend tells two more.” That is the story of viral growth, and for almost every indie game it is false in a way that matters.
Retelling
k ≈ 0.01
About one in a hundred people who noticed your game mentions it to someone, and they mention it to two or three people in a chat, not to an audience. Every branch dies immediately. Over a whole campaign this adds single-digit percent to your reach.
Somebody films it
k ≈ 0.14
Thousands of times rarer per person — but when it happens, the reach is a post, not a conversation. This is the only channel that can compound, and it only opens if the game is the kind of thing somebody wants to show. That is what a hook actually buys you.
Together
k ≈ 0.15
Below 1, every wave dies out and total reach settles at 1/(1−k) times what you put in.
Word of mouth is not a strategy you can execute. It is a multiplier on volume you already produced, and for most games that multiplier is 1.02.
What each thing was actually worth
Every row is the same year run again with one thing removed, twenty-four times over, and the difference taken. Not opinion about what matters — subtraction on your own plan.
| Lever | Wishlists it added | Share of the total | What it cost | Per wishlist |
|---|---|---|---|---|
| Working it out… | ||||
The shares add up to more than a hundred percent, and that is not an error: these things multiply each other rather than divide the result between them. A better store page converts the traffic the letters brought, the demo converts on the page, and the festival sends people to both. Remove any one of them and the rest get quietly worse — which is also why removing the store page hurts most: everything else ends up there.
Ads are the most honest line here, because their price is public: about six dollars per thousand impressions, a small fraction of which leaves any trace, a fraction of that opens the store page. Multiply it out and a wishlist costs a couple of dollars — fine if a wishlist is worth a couple of dollars to you, which it only is if the game converts at launch. Paid reach on a game with a weak hook is a machine for turning money into impressions and impressions into nothing. Set the hook slider low and watch the cost per wishlist climb while the spend stays exactly the same.
The store page is the strangest line, because it costs nothing and it multiplies everything above it. Every view, every mention, every letter ends at the same capsule image and the same trailer, and whatever fraction of people wishlist there applies to all of it at once. It is the only number on this page you can double in a weekend, and doubling it doubles the year.
Letters are the line most people never try. Fifteen minutes each, no budget, a few percent answer — and the ones who answer bring an audience that already trusts them, which is exactly the thing your own posting cannot buy at any volume. A hundred letters is twenty-five hours, about the same as three days of filming clips, and it usually moves more.
What none of these do is rescue a game nobody wants. Every lever here is a multiplier on how the game lands when somebody finally looks at it, and multipliers on a small number stay small. That is not a reason to skip them — it is the reason to know which number you are multiplying before you spend a year on it.
The threshold where the store starts working for you
Around 7,000 wishlists the storefront begins showing the game on its own. It is the first moment in this entire model where reach arrives without you paying for it in hours or dollars.
Before the threshold, every impression is something you made happen. After it, the platform adds a channel that scales with what you already have — and because it feeds itself, the gap between a game at six thousand wishlists and a game at twelve thousand is much wider than two times. This is the real reason the advice is always “wishlists, wishlists, wishlists”: not because the number is a score, but because it is the input to a system that only turns on above a line.
On the plan you set, that line is somewhere between plausible and out of reach — the distribution above says which. If it says a few percent, the answer is not to try harder on the same plan. It is to change the thing the plan multiplies: more draws, or a game that is worth filming.
What this leaves you with
Marketing is a volume of draws, not a campaign
There is no post that works. There is a number of posts, and a distribution. Plan in hundreds of posts a year or plan to be unknown — those are the two options the arithmetic offers.
The hook is worth more than the effort
Doubling your posting doubles your reach. Making the game visibly worth showing moves the notice rate, the wishlist rate, the retelling rate and the chance somebody films it — all four at once, and the last one multiplicatively. Same year of work, several times the result.
Views are not leads
A view is a video playing near a person. Single-digit percent of that registers as a game that exists, and single-digit percent of that becomes a wishlist. Anyone quoting you a view count as progress is quoting the wrong number.
Density beats duration
People forget at about 4% a week. The same hundred posts spread over two years reach a smaller audience than over six months, because you are refilling a bucket with a hole in it.
Budget is leverage, not a substitute
Money buys impressions at a known price. On a game nobody wants, that is an efficient way to purchase nothing. Fix the conversion first, then pay to point more people at it.
Every number this rests on
A model is only worth as much as its assumptions are checkable. These are all of them. They are calibrated to be right in order of magnitude and in the shape of the dependencies — not to predict your specific game, which nothing can.
| Platform | Median reach | Spread | Notice | Intent | Followers see | Hours |
|---|---|---|---|---|---|---|
| TikTok | 320 | σ 1.55 | 6% | ×0.75 | 11% | 1.5h |
| YouTube Shorts | 260 | σ 1.45 | 7% | ×0.8 | 12% | 1.2h |
| Reels | 190 | σ 1.4 | 5% | ×0.6 | 10% | 1.2h |
| X / Bluesky | 110 | σ 1.25 | 14% | ×0.85 | 28% | 0.5h |
| 800 | σ 1.7 | 18% | ×1.15 | 4% | 1h | |
| YouTube · devlog | 420 | σ 1.35 | 34% | ×1.35 | 30% | 10h |
| Steam · devlog | 90 | σ 0.9 | 55% | ×1.6 | 45% | 1.5h |
- Noticed → store page
- 12% before platform and hook modifiers. This is the step the algorithm and your hook decide together.
- Store page → wishlist
- 8%–32% depending on how good the page is, ×1.25 with a playable demo. This one is entirely yours.
- Second contact
- 3× more likely to open the page. Recognition is most of what converts.
- Letters
- 4%–14% reply and publish, median 700 views each with the same heavy tail, 15 minutes of your time each.
- Forgetting
- 4% of the aware audience per week — half gone in about 17 weeks.
- Retelling
- 0.6%–5.6% of new aware people tell 2.4 others.
- Someone makes their own video
- 0.0015%–0.22% per aware person, scaling with the cube of the hook. Their clip draws from the same heavy tail yours does.
- Store visibility
- Kicks in at 7,000 wishlists and grows sublinearly from there.
- Ads
- $6 CPM, 3.5% notice rate, ×0.55 intent versus organic.
- Launch conversion
- 4.5%–10% of wishlists buy in release week; year one is about 2.8× that.
- Your share of the price
- 52% after store cut, regional pricing, VAT and refunds.
- Reachable audience
- 3 million people for a given genre, across all platforms combined.
And what it deliberately does not model
- Genre size. The reachable audience is a flat three million. In reality it swings by two orders of magnitude between a cozy farming game and a hardcore grand strategy, and that single number moves everything else.
- Review score and release timing. Year one is a fixed multiple of release week. A game that reviews at 95% and a game that reviews at 70% have very different tails, and launching next to a giant costs you a chunk of the week you spent a year building toward.
- Ad attribution. The model spends ad money at a market rate and assumes it lands. In practice you cannot put a pixel on a store page you do not own, so optimising paid reach toward wishlists is guesswork — which makes real-world ads worse than modelled here, not better.
- Burnout. The plan you set runs every week for a year without a missed day. Nobody does that. If the number at the top only works at seven posts a week, it does not work.
The anchors are the checkable part: a thousand views is about one wishlist, the storefront starts helping around 7,000 wishlists, 5–10% of wishlists buy in release week, and you keep about half the price. Everything between those anchors is a mechanism built to reproduce them — plausible in shape, not measured. The whole thing is one readable file with tests that assert what it claims about the domain; change a constant you disagree with and every number on this page moves with it.