I rebranded two sites this week — godfreylabs.com and this one.

The first followed the textbook order. Interview against Don Miller's seven SB7 sections. Lock the BrandScript. Audit every persuasion surface against it — homepage, footer, meta, CTAs. Write a rewrites doc that proposes verbatim replacement copy, surface by surface. Then apply.

The second site skipped a step. Brandscript → straight to code. The implementation shipped. It looked right. I assumed it was right.

Today I wrote the rewrites doc anyway. Retroactively. Against a site that was already live.

That ended up being the most useful artifact of the whole rebrand.

What the rewrites doc actually does

A BrandScript is a posture — the customer is the hero, you are the guide, here's the villain, here's the plan. It's a one-page reference you can hold up.

A rewrites doc is the audit. For every persuasion surface — H1, subhead, CTA button, footer colophon, nav, meta description, the alt text on the hero image — you stand in front of it and ask:

• Is this the customer or me?
• Have I named the problem in their language?
• Empathy first, then authority?
• Can they picture the next three steps?
• Is the call to action specific?
• Are the stakes legible on both sides?
• Could a non-expert in this audience follow it?

Without the doc, those questions get asked exactly once — during the interview — and never again. The implementation drifts. Six months later the homepage and the brand doc disagree, and nobody noticed.

With the doc, every surface has a paper-trail answer. Which means every future copy edit answers to the same script, instead of restarting the argument from zero.

Why doing it retroactively still works

The instinct is: if the code is already live, the rewrites doc is overhead. The decisions are made. What's the point?

The point is that implicit decisions are the load-bearing ones, and they're invisible until you write them down.

When I audited this site against its BrandScript today, every persuasion surface scored PASS — the implementation matched the script. Good. But the more useful output was the section I added called "Architectural decisions — already locked." Things like:

• The hero leads with the chat, not a "book a call" button. Why? Because this site converts the vetter, not the stranger — and asking a vetter to convert is the architectural mistake.
• There's no /about page. Why? Because the whole site is the about page. Adding a route would create a second-best surface to read me at length.
• There's no testimonials section. Why? Because authority lives in artifacts on the page, not in third-party validation.

I made those decisions during implementation without writing them down. They were obvious in the moment. They will not be obvious to me in three months when I'm tempted to "improve" the site and add a Book-a-call button to the hero because a tweet I read said heroes should have one.

The retroactive rewrites doc is what makes those decisions defensible later. It says: this absence is a feature, not a gap.

The verification checklist

When the doc is done, it should pass five tests:

1. Every persuasion surface on the live site maps to a BrandScript section.
2. Each surface has a PASS or FAIL score against the SB7 rubric.
3. Architectural decisions that the code locked implicitly are codified in writing.
4. A brand vocabulary table exists — preferred phrasings and what to avoid — so future copy edits don't drift on word choice.
5. Stale items and future work are flagged honestly, not hidden.

If you can check all five, the brand and the implementation are aligned, and you have the artifact that keeps them aligned.

What I'd change if I did it again

Write the rewrites doc first, even when it feels like overhead. The retroactive version works, but the forward version is faster because you're writing copy you can paste in, not reverse-engineering copy that already shipped.

But — and this is the part most rebrand processes skip — if you've already shipped without it, do the retroactive pass anyway. The audit is the value, not the order.

The script is the thing your future copy edits answer to. Write it. Audit it. Even after.