What Voodoo Actually Looks For in a Prototype Submission

A hyper-casual publisher evaluates a prototype in stages: concept, playable build, a paid creative test, then retention. Each stage answers one question and most submissions end at the first two. Understanding which question a stage asks is what turns a rejection into information you can use.

Why this matters

Teams routinely optimise for the wrong stage. They polish art for a reviewer who is deciding on a ten-second video, or build progression for a test that only measures whether anyone clicks.

Knowing the sequence lets you spend effort where the decision is actually made. It also tells you what a rejection means, which is the part most teams never get.

The process described here is broadly common across the major casual publishers. Details differ, and the differences are worth reading before you choose where to send something.

The submission funnel, stage by stage

Four gates, in order.

1. Concept review. A human looks at a video or a description. The question is whether this is distinct enough and simple enough to be worth testing.
2. Build review. Somebody plays it. The question is whether the loop works and the build is stable enough to advertise.
3. Creative test. Real money buys real installs against real creative. The question is what an install costs.
4. Retention check. The installed players are watched for a few days. The question is whether they come back.

Each gate has a different audience and a different currency. Stage one is judgement, stage two is craft, stages three and four are numbers.

Most submissions are decided in the first ninety seconds of stage one. The rest of the funnel only exists for concepts that survive that.

The cost bar and how it is measured

Stage three is where the money question is settled, and it is measured more narrowly than founders expect.

The test buys installs in a small number of markets, against a handful of creatives, over a short window. The output is a cost per install for that specific setup.

Four things change that number, and none of them are the game.

This is why publishers compare your result against their own recent tests rather than a fixed public number. What they are looking for is whether your concept beats what they already have in the same conditions.

The practical consequence: a strong game with weak creative fails this stage. Creative is not marketing that happens later, it is part of the submission. There is more on that in Three Ad Creatives That Beat the Game.

Retention and playtime gates

Passing the cost gate buys you the retention question, which is now the one that decides most deals.

Publishers look at three things together.

• Day one return. Whether the game gave a reason to open it again.
• The shape from day one to day seven. A steep drop after a strong first day usually means the loop is thinner than the hook suggested.
• Session behaviour. Sessions per day and length, which move earlier than retention and act as an early signal.

The current market weights retention more heavily than it did a few years ago, for the reasons covered in The Hybrid-Casual Shift Explained. A prototype that wins on cost and loses on retention is now a much harder sell than the reverse.

What makes a concept ownable

Publishers are choosing something to invest in for years, which changes what they value at stage one.

Four properties come up repeatedly.

1. It reads in three seconds. A viewer scrolling past understands the goal without narration.
2. It has a satisfying moment that repeats. One clear payoff the game can deliver again and again.
3. It is not already in their portfolio. Distinctiveness is measured against what they already publish, not against the whole market.
4. It has somewhere to grow. A visible path to progression and purchases, even if none of it is built yet.

That fourth point is newer, and it is where many technically strong prototypes now fail. A concept with no plausible depth is a short-lived title, and publishers plan for longer.

Common instant rejections

These end a submission at stage one or two, before anything is measured.

• The build crashes or stalls on launch. The reviewer moves on.
• The mechanic is a direct copy of something in their portfolio or a current market leader.
• Trademarked content appears in the build or the video.
• The video is edited to hide the game, with cuts that never show sustained play.
• The first playable moment is buried behind a menu, a tutorial, or a loading sequence.
• No metrics and no explanation of why there are none.
• The concept needs a paragraph to explain. If it does, it will not work in an ad.

Almost all of these are packaging rather than design, and the full list to work through before sending is in The Publisher Submission Checklist.

What we would do

Build for stage one first. Before writing progression or polishing art, capture ten seconds of silent footage and show it to someone unfamiliar with the game. If they cannot describe the goal, nothing further matters yet.

Treat creative as a deliverable of the prototype, not an afterthought. Two or three concepts cut from real footage should exist before you submit.

Then be honest in the submission about what you have and have not tested. Publishers see a great deal of data and read an unqualified number as either inexperience or spin. A clearly stated small test reads far better than an impressive number with no context.

Send to one publisher, learn from the response, and improve the next submission before sending it elsewhere.

The short version

• Four gates: concept, build, creative test, retention.
• Most submissions are decided in the first ninety seconds of stage one.
• Cost per install is judged against the publisher's own recent tests, not a public figure.
• Retention now carries more weight than it used to, and decides most deals.
• Ownable means readable in three seconds, repeatable, distinct from their portfolio, and able to grow.
• Nearly all instant rejections are packaging problems, which are fixable.

Capture ten silent seconds of your prototype and show it to someone cold. If they cannot tell you the goal, fix that before anything else. When you are ready for a second read, send it to us.

Related reading: Homa, Voodoo, Supersonic, CrazyLabs, Why Most Hyper-Casual Prototypes Fail in the First Ten Seconds, and What to Do When Every Publisher Passes.