all notes
2026-08-14Naman Barkiya

Guiding Elephant build postmortem: ten pages, one CMS, and the platform call that saved a rebuild.

Guiding Elephant, a finance consulting firm, went from no site and no matching identity to a ten-page, single-author Framer build in 2024: a brand system matched to the firm's credibility, a CMS scoped to what one editor needed (not a headless setup), and a deliberate Framer-over-custom-code call because the site never crossed into high content volume or structured data.

A finance consulting firm whose advice was ahead of its website. Ten pages, one author, a Framer build over a custom one, and a CMS scoped to what the team could actually run. The platform decision, not the design, was the one that mattered.

Guiding Elephant is a finance consulting firm whose advice was ahead of its website. We shipped a ten-page, single-author Framer site in 2024: a full identity system matched to the firm's credibility, an efficient CMS the team could run without us, and a deliberate refusal to build it in custom code. The platform call, not the design, was the decision that mattered most.

The brief was not "build us a nice site." It was "make the site match the advice." A finance consultancy sells judgment. If the site looks thinner than the counsel behind it, a prospect discounts the counsel before the first call.

What was actually broken?

Not the firm's expertise — the gap between it and what a prospect saw online. No site, no identity system, nothing that signaled the depth of the advice being given. A consultant selling financial judgment cannot afford to look like a hobby project, and the prior state read as one.

We did not start with pages. We started with the identity: a brand system built to carry the weight of "trust us with this decision," then a ten-page structure — home, services, team, and a handful of supporting pages — built to hold it.

Why Framer, and not a custom build or WordPress?

Because the job the site had to do fit Framer's strengths exactly, and every other option would have paid for capability nobody asked for.

Ten pages. One author. A design language that wanted room to breathe — expressive layout, custom motion, a look that reads as considered rather than templated. That is the Framer case, not the WordPress case or the Webflow case. We've written the full decision tree elsewhere: Framer wins on design-led, low-content-volume, single-editor sites; it loses the moment content volume or structured data goes up. Guiding Elephant never crossed that line, so we didn't build past it.

The platform decision in one table:

SignalGuiding ElephantWould have pushed to Webflow/WordPress
Page count1020+
Editors13+
Content typeMarketing, mostly staticBlog, case library, e-commerce
Design barHigh, expressiveTemplated is fine

Every row landed on the Framer side. A custom-coded stack would have added weeks of engineering for design flexibility Framer already gives out of the box, for a client who was never going to need headless CMS queries or e-commerce logic.

What did the CMS decision actually solve?

The team needed to update services copy and team bios without opening a ticket with us. That's the whole requirement — not a blog, not a content library, not multi-language. We scoped a CMS sized to that requirement inside Framer's own content model instead of reaching for a headless setup that would have needed its own hosting, its own auth, and its own maintenance conversation.

The refusal here was the same one that shows up on every build: don't build the CMS the client might grow into, build the one the team can run on day one. If Guiding Elephant's content needs outgrow Framer's model, that's a real conversation to have later, priced against what's actually needed then — not a hypothetical to architect around now.

What did the identity system have to prove?

That the firm's depth was visible before a prospect read a word of copy. A finance consultancy's site is a credibility instrument first and an information source second. We built the identity to match the register of the advice: no template gloss, no generic "trusted advisor" visual language — a system specific enough that it could only belong to this firm.

The team pages carry more design weight than is typical for a services site. That was deliberate. Prospects vetting a consultancy look at who is giving the advice before they look at the service list.

What would we do differently?

Push the CMS-scoping conversation earlier in the timeline. It landed right — the team runs the site without us today — but we settled the content model after the page structure was mostly locked, which meant one late pass to fit the CMS fields to pages that weren't originally built around them. On a ten-page site the cost was a day, not a rebuild. On a larger one, that ordering mistake gets expensive fast.

The platform call — Framer, not a custom build — was right the first time. Ten pages, one editor, a design bar that mattered more than content depth: that's the profile every time, and we'd make the same call again without a second look.


The instinct behind that platform call — matching the tool to what the brief actually requires, not to what's most impressive — is the same one behind how we scope every build before a line of code gets written.


Written 2026-08-14 by Naman Barkiya.

FAQ

Questions this usually surfaces.

What would you redo on the Guiding Elephant build?
Scope the CMS content model earlier, before the page structure locked. It landed right — the team runs the site without us today — but settling it late meant one extra pass to fit CMS fields to pages that weren't originally built around them. On a ten-page site that cost a day; on a larger one, that ordering mistake gets expensive fast.
How long did the Guiding Elephant site take to build?
We don't publish a single number for this build the way we do for fixed-scope sprints — the identity system and the ten-page Framer build ran together as one engagement rather than a strict sequential timeline. What's fixed is the scope that shipped: ten pages, one CMS, one brand system, no features the single-editor team couldn't run itself.
What does a brand-and-website build like this cost?
Cost scales with what's actually being built — an identity system paired with a ten-page Framer site is a smaller engagement than the same identity work paired with a CMS-heavy or e-commerce build. We scope and quote per engagement rather than publishing a flat number, because the honest answer depends on page count, editor count, and how much of the identity has to be built from zero.