00The cost problem
Building each site starts the work again.
Building websites one at a time carries a cost that barely moves no matter how many sites you produce. Each project starts from a blank page, or close to it, and the work of structuring pages, wiring navigation, and setting up deployment happens again for every site. The cost per copy stays roughly constant because the method never changes.
The print shop and scribe comparison makes the difference concrete. A scribe copies each document start to finish, and the effort per copy is about the same whether you need one document or fifty. A print shop works differently. It builds the press and sets the type once, then each later copy costs a fraction of the first. The upfront investment is real, and it only makes sense if you actually need many copies.
A template gives you a starting file. A factory gives you a production system, and the distinction shows up in four concrete parts.
01The system
Four concrete parts.
The data model
The data model defines what varies between sites. City sites, branch sites, or product-line sites each have a set of fields that change: name, address, service list, imagery, contact details. The data model names those fields once so each site supplies values instead of redefining structure.
The template system
The template system holds what stays constant: layout, component behavior, styling, and the rules for how content renders.
The pipeline
The pipeline validates, builds, tests, and deploys. It checks that required fields exist, that links resolve, that the site compiles, and that the output reaches the right host.
The content workflow
The content workflow collects and approves each site's information before it enters the pipeline, so the people who know the content review it in a structured way rather than through ad hoc edits.
Structured content is the load-bearing idea here, and it is not unique to this pattern. Astro content collections are one example of defining content shape separately from presentation. GitHub Actions shows the same thinking on the automation side: a defined workflow that runs the same checks on every change. Neither product is what SKYNET runs on. They are implementation examples of the two halves, structured content and automated pipelines, that a factory depends on.
02The workflow
The work changes when a factory exists.
Launching another site becomes a controlled configuration, review, validation, and deployment process, not merely data entry.
Someone configures the new site's values in the data model, the content workflow routes those values through review and approval, the pipeline validates and builds, and deployment happens only when the checks pass. The human steps remain: decisions about content, review of the rendered result, and sign-off. What disappears is the repeated construction of structure that already exists.
03The investment
The economics run in both directions.
The first site through a factory costs more than hand-building it, because the production system itself has to be designed and built.
The data model, the template system, the pipeline, and the content workflow are real work with real cost. The investment pays back only across enough structurally similar sites, where the per-site cost drops far enough below the hand-built cost to recover what the factory took to create. There is no universal break-even count. The right comparison is concrete: estimate what the factory will cost to build and maintain, estimate what each site would cost built individually, and see where the crossover lands for the number of sites you actually need.
04The boundary
The pattern does not fit everywhere.
A small number of structurally different sites will not recover the factory investment, because the shared structure that makes the factory pay off is absent. Projects requiring genuinely bespoke designs and functions also argue against it. When each site needs its own layout, its own interaction model, or its own data shape, the factory's machinery becomes overhead rather than leverage. Those sites are better handled individually, and that is a legitimate choice rather than a failure of the pattern.
05Production example
SKYNET is the production example.
Sputnik Digital Effects built it as a factory system that shipped many city sites, and the SKYNET case study walks through how the pieces fit together.
06The decision
The decision rule is direct.
If you need several sites that share most of their structure and differ mainly in content, a factory will likely beat building them one at a time. If you need few sites, or sites that differ in structure, build them individually. Compare the factory investment against the cost of building each required site separately, and let that comparison decide.
If you are weighing that tradeoff now, we can look at your site list together and estimate which path costs less. Start with the services overview to see how we scope the work, then bring us the list of sites you need to ship.