Writing an audience and brand voice that gets the best from Wheremore

Wheremore helps you plan and repurpose content using two settings you control: the audience (who you're writing for) and the brand voice (how you sound). The system uses them to suggest topics and subtopics, build outlines, and repurpose your published content into LinkedIn and X posts. It does not write the finished articles — that step is yours — but these two fields shape everything it does suggest, structure, and repurpose. Get them right and the output sounds like you and serves your readers. Get them vague and it produces competent, generic suggestions that could belong to anyone.

This guide explains how to write both fields well. There's a short version for experienced users at the top, and a fuller walkthrough with examples below.

The short version

1. Derive your voice from real examples, don't invent it from adjectives. Paste in three to five of your best existing pieces and pull out the actual patterns. "Professional and engaging" steers nothing; "warm, plain-spoken, short sentences, opens with a concrete example before any theory" steers a lot.
2. Keep the two fields separate. Audience is who reads and what they already know. Voice is how you sound. Don't put tone in the audience box.
3. Be specific, and be concrete. Name the things you do and the things you avoid. Specific instructions beat abstract praise every time.
4. Keep it tight. A focused field works better than an exhaustive one. Long fields get skimmed — by the model as much as by people.
5. The voice controls how it sounds, never whether facts are true. The system's outlines and social posts can carry statistics with real, working source links — but a link that resolves is not proof the source supports the claim. Verify every figure against its source before it becomes a published article. No voice setting fixes this. This is the most important line in this guide.
6. The system already checks new topics against your existing cluster titles to avoid obvious overlap — so topic-level duplication is largely handled. What it can't catch is two differently-titled pieces making the same argument or taking opposite positions; that still needs your eye (see below).

If you do only the first and the fifth of those, you're most of the way there.

Audience: who you're writing for

The audience field tells the system who's on the other end — and, just as importantly, what they already know. This is the part people most often leave too vague.

A weak audience field names a job title and stops: "Business owners." A strong one captures what the reader already understands, so the system doesn't waste the article explaining the basics:

Finance and operations leaders at mid-sized firms who already understand the core concepts and use the standard tools daily. They want help making a specific decision, not an introduction to the subject from scratch.

The difference shows up immediately in output. Tell the system the reader is a beginner and it will define terms and walk through fundamentals. Tell it the reader is an expert and it will assume knowledge and get to the point. Most generic-sounding output comes from an audience field that didn't say which.

What to include:

• Who they are (role, level, context).
• What they already know — so the system pitches at the right level.
• What they want from your content (to decide something, to learn a skill, to be persuaded, to stay current).
• What they're not — "not a beginner," "not looking for a how-to" — if that keeps output on track.

Keep tone out of it. "Write in a friendly, upbeat way" is a voice instruction, not an audience one. Mixing them muddies both.

Brand voice: how you sound

This is the field that makes output recognisably yours rather than generic. It's also the one most often filled in with words that feel right but steer nothing.

Derive it from real examples — this is the single most valuable thing you can do

Don't sit in front of an empty box and brainstorm adjectives. Instead, gather three to five pieces of your existing content that you're proud of — the ones that sound most like you — and read them for patterns:

• How do they open? (A question? A scene? A bold claim? A definition?)
• What's the sentence rhythm? (Long and flowing? Short and punchy? A mix?)
• What's the attitude? (Warm? Sceptical? Authoritative? Playful?)
• What do they consistently avoid? (Hype? Jargon? Hedging? Exclamation marks?)
• Is there a recurring move or structure?

Then write those patterns down as your voice field. You're not inventing a voice; you're describing one you already have. If you don't have existing content, write down concrete do/don't examples instead — even a handful sharpens the output dramatically.

Specific beats abstract, every time

This is the most common mistake, so it's worth seeing side by side:

Words like engaging, compelling, high-quality, and insightful feel like instructions but carry no information the system can act on. Replace every one of them with a concrete description of what it actually means for your content.

Name what you avoid, not just what you want

The system responds well to negative instruction. If your brand never uses exclamation marks, never says "unlock" or "supercharge," never opens with a dictionary definition — say so explicitly. Banning your specific clichés is often more effective than describing your aspirations.

Keep it tight

A long voice field doesn't produce a richer voice; it produces a skimmed one. The system, like a reader, weights a focused set of clear instructions more reliably than an exhaustive list. Aim for the rules that genuinely shape your sound, and cut the rest. If a line could apply to any brand, it isn't doing work.

The rule that matters most: voice is not fact-checking

Your audience and voice fields control how the content sounds and what stance it takes. They do not, and cannot, control whether the facts in it are true.

When a topic invites it, the system produces confident, specific statistics and claims in its outlines — and when it repurposes a piece for social, it can compress or reshape a figure from the original. Often these come with real, retrieved source links. That sounds reassuring, and it's better than invented links — but it is not the same as the claim being correct.

Here's the trap, and it's the one most likely to slip past you: a link that resolves to a real page is not proof that page supports the claim. In testing, the system produced outlines where some statistics were real, correctly stated, and correctly linked — and, sitting right beside them, a real source link attached to a distorted version of its number (the right figure, the wrong meaning). Only reading the source revealed that one of them didn't actually say what the outline claimed. A  link under a wrong claim is the easiest error to wave through and the most damaging to publish, precisely because the link checks out.

So treat every outline and every social post as a draft to verify, never as finished fact. Before anything becomes a published article:

• Open the source and confirm it actually states the figure — not just that the link works. A working link is necessary, not sufficient.
• Check the claim matches the source's meaning. A real number attached to the wrong claim is still wrong. Watch for figures borrowed from one context and applied to another.
• Trace numbers to a primary source, not a blog restating them. A statistic that only lives on content-marketing sites repeating each other hasn't been verified, however many links point to it.
• Check social repurposing didn't distort or strip a caveat. Compression is where a careful claim turns into an overconfident one.

A polished voice makes all of this more important, not less — a well-written statistic is more believable, so a wrong one does more damage. If you take one thing from this guide: the system suggests and structures; a human verifies every fact. Always.

Two things to handle outside the voice field

1. Keeping your content from all sounding the same. If your voice field describes one rigid structure ("always open with a statistic, then three points, then a conclusion"), every outline will copy it, and your library will start to feel formulaic. The system builds each outline on its own — it can't judge that the last five took the same shape.

The fix isn't in the voice field. Vary the approach per piece: some argue against a common assumption, some compare two options, some walk through evidence, some answer a single narrow question, some simply explain something clearly. Rotating the shape deliberately keeps your library varied even though each piece is planned on its own.

2. Argument-level overlap and contradiction. The system already checks new topic and subtopic suggestions against the existing titles in your cluster, so obvious duplication — two near-identical topics, a subtopic restating one already there — is largely caught for you. What a title check can't see is two pieces with different titles that make the same argument, or that take opposite positions on the same question. Those slip through because the titles don't look alike; the collision is in the reasoning, not the wording. When it matters, give new topics and outlines a quick human read against what you've already published, with that specific risk in mind.

Because every account is different

Your audience and voice are yours alone. A consumer recipe site writing for home cooks, a financial-services firm writing for cautious first-time investors, and an agency writing on behalf of a client all need completely different settings — and an agency managing several clients will keep a distinct audience and voice for each.

There's no universal "good" voice to copy. The good voice is the accurate one: the audience that matches your real readers, and the voice derived from your real content. The steps in this guide work for any of them, because they're about describing what's true of your brand precisely — not about reaching for a particular style.

A quick checklist

Before you save your settings:

• [ ] Audience names who the reader is and what they already know.
• [ ] Tone lives in the voice field, not the audience field.
• [ ] Voice is derived from real examples, not brainstormed adjectives.
• [ ] Every vague word (engaging, high-quality, compelling) is replaced with something concrete or cut.
• [ ] You've named what to avoid, not just what you want.
• [ ] The fields are tight — every line earns its place.
• [ ] You understand the system suggests, structures, and repurposes; a human verifies every fact before publishing.