Visual novels look inexpensive to make because most of the screen is text, and this is one of the most persistent misunderstandings in independent game production. The writing is rarely the constraint. The art is, and the art costs are structured in a way that surprises almost everybody the first time they build a full project.

Understanding the exchange rate between different kinds of asset is what lets a small team scope a project it can finish. Teams that do not understand it produce a beautiful first hour and then run out, which is by far the most common way these projects die.

Sprites and the expression problem

A character sprite is drawn once and then needs expressions, and expressions multiply. A cast of six with eight expressions each is forty-eight pieces of art before anything else exists, and eight is a low number for a story that runs for hours. Add pose variants and outfit changes and the count rises quickly.

The standard economy is layered sprites: one body, swappable heads or facial parts, separate mouth and eye layers. That reduces the count dramatically and is why so many visual novels have characters who change expression without changing posture. It is a budget decision that became a genre convention.

Backgrounds cost more than people expect

A background is a full illustration with perspective, lighting and detail, and it appears behind everything, so errors are visible for a long time. It also cannot be reused across projects the way a character can, since it belongs to a specific place. Most teams end up with far fewer locations than their script originally implied.

This is why script and art must be planned together. A writer who invents a new location for a single scene has committed the team to a full illustration for four lines of dialogue. Writing to a fixed location list is a constraint that improves scripts as often as it hurts them, in the same way a built set does in the piece on set reuse.

Event art is where budgets vanish

Event illustrations, the full-screen images for significant moments, are the most expensive art in the project per unit. They are also the ones players remember and screenshot, so cutting them feels like cutting the point of the game. Teams consistently plan too many, deliver a fraction, and feel the shortfall keenly.

The workable approach is to fix the number early, allocate them to the moments that carry the story, and treat that list as closed. Adding one later should mean removing one, which is a rule that sounds harsh and saves projects. Adult projects face this most sharply, since explicit scenes are almost always event art rather than sprite work.

Scope is set by art, not by words

A hundred thousand words of script is a large writing job and a manageable one for a single person over a year. The art implied by a hundred thousand words is usually several times more work than that. Estimating a project by word count is therefore close to meaningless, and it is why so many announced projects quietly disappear.

Estimate by asset list instead: sprites, expressions, backgrounds, event illustrations, user interface. Multiply by an honest per-asset time. Then compare the total against the hours actually available, and cut until the numbers agree before writing the script rather than after.

The user interface is the asset nobody budgets and everybody needs. Text boxes, menus, save screens, name plates, buttons in several states and a title screen add up to a real illustration job, and they cannot be cut, because a game without a save menu is not shippable. Put them on the list at the start, where they can influence scope, rather than discovering them in the final month.

Planning a project you can actually finish

Fix the location list, fix the cast size, fix the expression count, fix the event illustration count, and write inside those limits. Constraints of this kind are not a compromise on ambition; they are what allows a finished thing to exist. The finished small project also teaches more than the abandoned large one.

The income routes that support this kind of work are covered in the piece on where the money comes from, and the distribution constraints on adult projects specifically in the piece on why publishability varies.