The reality of Framer templates

A client came to me in 2025 with a template called Kim already bought, already live, already breaking under its own weight. My job was to restructure it and wire up a real CMS. One week, start to finish.

That job is more common than people think. Templates get sold as a shortcut. For a lot of teams they become the opposite: a structure they now have to pay someone to undo before they can move forward.

What a template actually sells you

A template preview is a demo, not a foundation. Someone built it to look finished in a screenshot, not to survive a real content team adding, deleting, and reordering things for the next two years.

That difference shows up fast. Sections that look dynamic are often just styled boxes with text typed directly into them. Nothing pulls from a collection. Add one more item to a row built for a fixed number and the whole layout shifts in ways nobody intended. The template photographs beautifully and behaves like a stack of cards with no rules holding it together.

Breakpoints are the other quiet failure point. A template gets demoed on one screen size, tuned until it looks perfect there, and shipped. Nobody stress-tests it at the widths real visitors actually use: a laptop with a slightly smaller window, a tablet held sideways, a phone that isn't the newest model. The layout holds until it doesn't, and by then the client has already told their own customers the site is live.

None of this is a scandal. A template is priced and built for someone setting up a single static page once and rarely touching it again. The moment a team wants to publish new work regularly, assign different people to update different sections, or grow past exactly what the demo shipped with, the thing that sold them starts working against them.

Rebuilding Kim

Kim was already live when I got it, which is its own kind of pressure. You're not building from nothing. You're taking something a client already approved and already pointed people to, and changing its bones without changing what it looks like to a visitor who refreshes the page mid-week.

The work was structural, not decorative. Content that had been typed directly into layers needed to live in a CMS collection instead, so the client could add a new entry without opening the canvas. Components that had been duplicated and hand-edited needed to become actual variants, so a change in one place didn't require hunting down every copy of the same section. None of that shows up in a before-and-after screenshot. It shows up later, once the client has added more to the site than the original demo ever held, and nothing breaks.

One week is tight for that kind of surgery. It's enough time when the visual design is already settled and the job is architecture, not aesthetics. You're not deciding what Kim should look like. You're deciding how it should be built underneath the look it already has.

When a template is the right call

I still recommend templates. Constantly.

If someone needs a single page live this week and they're not touching it again after, buying a template and customizing the copy is the correct decision, not a compromise. My own three-day landing page build exists for a similar reason: speed, for a team that needs something live now, not a philosophical stance against starting from scratch.

Custom build earns its cost when the site has to keep growing after launch. When a marketing team needs to publish without a developer in the loop. When the brand has to hold up across far more pages than a template was ever built for. When the CMS structure has to be planned before content exists, not retrofitted after the fact. That's a different job than picking a nice-looking template off a marketplace, and it takes longer for a reason.

This is also why a good chunk of my work is fixes and sections on things that already exist, rather than new builds from a blank canvas. A team buys a template, launches fast, and eventually needs someone to go back through it and make it behave: not redesign it, just make the thing they already committed to actually hold up under real use. That's a different job than either buying a template or building custom from scratch, and it's just as real.

The mistake isn't buying a template. It's buying one and expecting it to behave like a custom system just because it looks like one in the preview. Those are two different products wearing the same interface, and only one of them was built to survive contact with a real team.