How to build a B2B SaaS brand voice in 30 days

Most brand voice projects produce a PDF that nobody reads after the kickoff meeting. The document says things like 'confident but approachable' and 'expert yet human,' and then every writer on the team continues doing whatever they were doing before. That is not a brand voice. That is a mood board with punctuation.

A real brand voice is operational. It changes what a writer types on Tuesday morning. If yours does not do that, it does not exist yet.

Here is a 30-day process that has worked for early-stage and Series A SaaS companies. It is not glamorous. It involves reading a lot of your own copy and admitting most of it is mediocre.

Week one: audit what you already sound like

Before you decide what you want to sound like, find out what you actually sound like right now.

Pull 20 pieces of copy from across your company: homepage headline, three email subject lines, five sales deck slides, two blog posts, a LinkedIn post from the CEO, a support article, a few social captions. Read them all in one sitting.

You are looking for three things.

First, vocabulary drift. Does your homepage say 'streamline' while your emails say 'simplify' and your sales deck says 'optimize'? These words feel interchangeable but they signal different things to a reader. Drift means your voice is being written by whoever happened to be at their keyboard that day.

Second, sentence rhythm. Are all your sentences the same length? Long, clause-heavy sentences suggest a company that is still figuring out its own product. Short sentences suggest confidence. Most SaaS copy defaults to long.

Third, who is missing. Read the copy and ask: where is the customer in this? A lot of B2B SaaS copy describes the product in loving detail and mentions the buyer as an afterthought. If your homepage says 'we built a platform that' more than it says 'you can,' that is a structural problem, not a voice problem.

Document what you find. Be specific. 'We use passive voice in 14 of 20 samples' is useful. 'Our tone feels a bit off' is not.

Week two: find your voice in the wild

Your brand voice already exists. It is in the emails your best salesperson writes. It is in the Slack message your founder sent when a customer churned. It is in the support reply that got a five-star rating.

Collect these. Seriously. Ask your sales team to forward the three emails that got the fastest replies. Ask support for the tickets where customers said something kind. Pull any public writing from your founders that felt natural rather than polished.

Now look for patterns. What words come up repeatedly? What sentence structures? What does the writer assume the reader already knows? What does the writer never bother explaining because it feels obvious to them?

This is your raw material. You are not inventing a voice. You are excavating one that already exists in the people who built this company.

One practical exercise: take the best email your top salesperson wrote and rewrite your homepage headline in their voice. Not their words. Their rhythm, their assumptions, their level of directness. If the result sounds better than what you have now, you have found something worth keeping.

Week three: write the rules, not the adjectives

Here is where most brand voice projects go wrong. They produce a list of adjectives: 'bold, clear, empathetic, expert.' These are descriptions of an outcome, not instructions for producing one.

Instead, write rules. Rules are actionable. A rule sounds like this:

• Write the customer's problem in the first sentence, not the second.
• Never use 'solution' as a noun. Say what the thing actually does.
• If a sentence has more than 20 words, cut it in half.
• Avoid the word 'seamlessly.' It means nothing and it is in every competitor's copy.
• When describing a feature, follow it immediately with a consequence for the buyer.

Aim for 8 to 12 rules. Fewer than 8 and you have not made enough decisions. More than 12 and writers will not remember them.

Test each rule against your audit from week one. If the rule would have improved 15 of your 20 samples, keep it. If it only applies to edge cases, cut it.

This is also the week to make one hard call: who are you not writing for? A brand voice is as much about exclusion as inclusion. If your product is built for technical buyers, stop softening the language for non-technical readers. Pick a side. Copy that tries to speak to everyone reads like it was written for no one.

Week four: put the rules under pressure

A voice document that has never been tested is a hypothesis. Week four is about falsifying it.

Take three real writing tasks your team has coming up. A new feature announcement. A nurture email. A homepage section rewrite. Assign each one to a different writer and give them only the rules document, not each other's drafts.

When the drafts come back, read them together. Ask two questions.

First: do these sound like they came from the same company? If yes, the rules are working. If no, the rules are either too vague or missing something important.

Second: do these sound like a company you would want to buy from? This is the question most teams skip. They get so focused on consistency that they forget consistency of a bad voice is still a bad voice.

Revise the rules based on what you find. Expect to kill two or three rules that sounded good in theory but produced awkward copy in practice. Expect to add one or two rules you did not anticipate needing.

By the end of week four you should have a document that is under two pages, contains specific rules rather than adjectives, and has been tested against real copy. That is a brand voice guide. The PDF with the mood board is a brand feelings guide. They are not the same thing.

The thing most teams get wrong about maintenance

A brand voice is not finished at day 30. It is started.

The companies that maintain a consistent voice do one thing differently from the companies that do not: they treat voice as a quality check, not a creative exercise. Every piece of copy that goes out gets a ten-second voice pass. Does this follow rule 3? Does this violate rule 7? That is it.

This does not require a brand police. It requires whoever is approving copy to have the rules document open. At most companies, that is one or two people. Get those people aligned on the rules and the rest follows.

The 30-day timeline is tight but it is intentional. A longer timeline produces more debate and less copy. Debate about voice is almost always debate about strategy, and the best way to resolve a strategy debate is to write something and see if it works.

Write something. See if it works. Adjust the rules. Repeat.

That is how a brand voice becomes real instead of remaining a document that lives in a shared drive folder called 'Brand 2024 FINAL v3.'