Templates fail when they look pretty in the gallery and fight your content in production.
In testing, flexible section builders beat locked “artistic” templates when content changes weekly.
Demo content is written to flatter the template. Before you fall for one, write down your real pages, your longest page title, your actual navigation, and the number of products or posts you will have in a year. Then judge the template against those, not against the photography.
The most common failure we see is a template built for six polished pages being used for a site that grows to forty. Navigation collapses, the homepage stops representing the site, and the fix is a migration nobody budgeted for.
Colours and fonts are trivial to change on any modern builder. What you cannot easily change is how the template thinks: whether sections are independent blocks you can reorder, whether the header supports the menu depth you need, whether templates exist for the page types you will actually produce. Choose for structure and treat the styling as a bonus.
A template chooses your colour contrast, your heading order and your image weight before you write a word. Check that headings descend in order rather than being picked for size, that text on coloured overlays is genuinely readable, and that the hero is not a huge uncompressed image. These are painful to retrofit and nearly free to check up front.
Custom work pays when the site is the product, when the brand is the differentiator, or when a flow — booking, configuring, quoting — is unusual enough that no template models it. For most marketing sites a well-chosen template plus disciplined content beats a mediocre custom build, because it ships sooner and can be maintained by the person who owns it.
On most builders, switching means rebuilding pages: content usually carries over, layout rarely does. Treat the first choice as a medium-term commitment and test it against your real content first.
Not by existing. They hurt when they bury text in images, produce a broken heading order, or ship heavy uncompressed media. All three are visible before you commit.
Judge support and update history rather than price. A maintained free template beats an abandoned paid one, and abandonment is the risk that actually costs you.