Glossary/Software factory
What is Software factory?
Also called Software factory for go-to-market.
A software factory is a repository and toolchain holding a company's context in one place so that agents can build from it: the standards, the data, the specifications and the history of what was tried. The term comes from software delivery, where it named a standardized pipeline for producing applications, and it moved toward this meaning as the scarce input shifted from writing the code to supplying the context.
The original sense is older than agents and narrower. A software factory was a repeatable path to production: shared templates, a common build, governed deployment, the point being that the second application costs less than the first. That meaning is still in use and is still about shipping code.
What changed is which input is scarce. When an agent can write the implementation, a company's advantage moves from the code it already wrote to the means of producing the next thing, and those means are mostly context: what it sells, to whom, at what price, against which competitors, under what standards, and what happened the last five times it tried. A software factory in the current sense is the repository that holds that and the toolchain that lets an agent act on it.
The go-to-market version applies the same pattern to revenue instead of to product. The plan and its numbers, the durable knowledge about who buys and why, the efforts under way, the plays running in production, the evaluations that gate them and an append-only archive of what happened all live as files in one repository, reviewed as diffs and deployed by pipeline. This is a way into the pattern rather than a category name: someone searching for a software factory usually means shipping code, and go-to-market is the part of the company the idea reached last.
Is a software factory the same as a monorepo?
A monorepo is where it lives. The factory is the repository plus the things that run against it: the tests, the build, the deployment path and the agents that read and change it. A repository of documents with nothing executing against it is a wiki that happens to have git history.
What goes in a go-to-market software factory?
The plan and the number it is aiming at, the durable knowledge about the market, the bounded efforts under way with their owners and deadlines, the operating rhythm, the engine itself declared as code, the tests that gate changes to it, and an archive of every artifact produced. The archive is the part most often missing, and it is what lets the next agent run learn from the last one instead of repeating it.
How is this different from GTM as code?
GTM as code is the practice of declaring the engine in version-controlled files instead of in a vendor interface. A software factory is the wider repository that also carries the context, the plan and the history an agent reads before it changes anything. You can do GTM as code without any of that context; the factory is what makes what comes out of it worth sending.