Auditing a brand’s vocabulary

Hospitality inherits a lot of language from real estate. Real estate had already been forced to reckon with some of it. Consolidating two portfolios into one brand was the one moment when the underlying vocabulary could be changed deliberately rather than under pressure later.

I wrote the recommendation — not a sensitivity memo, a specification, with substitutions, reasoning, and a rollout argument, written to be adopted into the copy standards.

I organized it into four categories, because the fixes are not all the same kind of fix.

1. Terms with a clean replacement. Main bedroom for master bedroom. Dual vanities for “his and her” sinks. Two closets. Partner instead of husband or wife. They and their where a couple’s gender had been assumed.

2. Terms with no acceptable use. “Plantation” in every context, including as an architectural descriptor — with a redirect rather than a synonym: describe the Southern hospitality the word was reaching for and drop the reference.

3. Terms that exclude by assumption. The category most brands miss. “Walking distance” and “steps away” both assume the reader walks. In a category with a meaningful share of older and mobility-limited owners, that quietly tells some customers this is not for them. The fix is more informative anyway: nearby, adjacent, on-site, a few hundred yards away.

4. Terms that depend on context. “Colonial” is fine for a specific period and wrong as atmosphere — specify Georgian, Neo-classical, Greek Revival. Idioms with uncomfortable origins flagged for removal; idioms that simply fail in translation flagged separately, for a different reason, because the brand sells internationally.

Alt text was being written inconsistently across a large content operation. The problem was not bad writing — nobody had written down what good looked like.

So I wrote the standard. It opens with the three separate readers who need the text and why one rule cannot serve all of them: a person using a screen reader, a search engine, and a browser that failed to load the image. Then the rules: describe the image’s purpose, not its contents; account for surrounding copy; punctuate, because it is read aloud; keep it to 125 characters; skip photo credits, skip “image of,” skip anything the caption already says.

The character ceiling did most of the work. A hard constraint forces the writer to decide what the image is for. Rules that say “be concise” do not change behavior. Rules with a number in them do.

This is the work I would point at first for a role that involves content governance, brand guidelines, or reviewing generated copy at scale. Auditing a vocabulary for exclusion, ambiguity, and translation failure is the same discipline — done upstream, on the source language, before anything gets published.

Sanitized writing samples demonstrating both frameworks are available on request.