From the Nerd News Archive: our road retold in the voice of the moment. A story written today.
Every website project starts with a question nobody asks out loud: how much of this site should exist before we begin? A template already exists entirely. A component library exists in parts. A blank canvas does not exist at all. I have built from all three starting points, and I have watched money get wasted in every direction, so I finally wrote down the framework I use to choose.
The framework is not a philosophy. It is a sorting function with two inputs: what the site must accomplish, and what the budget can honestly carry. Goal and budget. Everything else is detail.
When a template is the honest choice
A template is honest when the goal is presence and the budget is small. A new business that needs to exist online this month, look credible, list its services, and take a phone call does not need custom anything. It needs a good template, real photos, real words, and a fast launch.
The discipline is admitting the trade out loud. A template buys speed and a proven layout, and it costs distinctiveness. The site will resemble other sites built from the same bones, and for a business whose customers are comparing plumbers rather than design portfolios, that trade is usually correct.
The tell that a template was the wrong call: the business starts bending to fit it. Services get renamed to match the menu slots. A section gets cut because the theme has no home for it. When a company reshapes itself around someone else's layout, the cheap choice has started charging rent.
The sweet spot: wireframes plus a component library
The middle path is the one I recommend most often. Wireframe the site's argument first: what each page must say, to whom, and in what order. That thinking is fully custom, and it is where most of the value lives. Then build the wireframes out of proven components, the accordions and sliders and card grids that have been solved a thousand times and do not need solving again.
This splits the budget along the right seam. Custom money goes to structure, copy, and flow, the things that decide whether the site works. Borrowed execution covers the parts a visitor never judges you on. Nobody has ever chosen a vendor because their accordion was artisanal.
The tell that the middle path was wrong: every page looks eighty percent right and nothing feels like the brand. If review meetings keep circling a vague sense that the site could belong to anyone, the components are wearing the company instead of the other way around, and the project actually needed the next tier.
When scratch is the only way
Building from scratch is not the luxury tier. It is what the requirements demand when the goal cannot be met by arranging existing parts. An interaction nobody's library ships. A catalog or a tool that behaves like a product instead of a brochure. A brand experience specific enough that borrowed bones would show through the skin. When the goal is one of those, scratch is not the expensive option. Everything else is, because everything else fails.
The tell that scratch was the wrong call is quieter: the invoices are custom but the requirements never were. If the finished site could have been assembled from parts and no visitor would ever know the difference, the project paid engineering prices for a problem templates solved years ago. Scratch should exist because the goal insisted, not because it sounded impressive in the proposal.
The framework on one page
- Small budget, standard goal: take the template, and spend the savings on photography and copy
- Real budget, custom message, standard mechanics: wireframes plus a component library
- A goal no existing parts can meet: build from scratch, and only then
- In every case, ask what the goal requires before asking what the money allows
The order of those questions matters. Budget-first thinking picks a tier and then trims the goal to fit it, which is how businesses end up with sites that are affordable and useless. Goal-first thinking sometimes concludes that the honest answer is a smaller goal, and that is fine. A modest site that does its one job beats an ambitious site that almost does three.
None of this is about tools, which is why I expect it to outlast every tool I currently use. Goals and budgets are permanent. The shelf of pre-built things only gets longer. The judgment about when to reach for that shelf and when to refuse it: that is the actual service I sell.
