UX-writing and brand platform: how meanings get into the interface

Why does a UX-editor need a document, not just a feel for language

Let’s take a specific situation: the product is growing, and second editors, marketing, support, and outsourced contractors start contributing to the texts. Each has their own taste, and after six months, the interface hosts three ways to address the user — sometimes 'you', sometimes 'you' (informal), sometimes impersonal constructions. Without a platform, the team has no argument except 'it seems to me that this is better.' With a platform, there is a document arbiter: any disputed decision is checked not by taste, but by the brand’s principles.

A strong editor can maintain the brand's voice in their head without a document — but this works only until their departure or until the first scaling to a new channel.

Four layers through which meaning reaches the button

Texts that 'seem to follow the guidelines but don’t sound like the brand' — a typical symptom that one of the layers between strategy and screen is missing.

• Brand principles — basic settings, for example, 'transparency', 'user control over decisions', 'care without fluff'.
• Tone of voice — scales that translate principles into sound: formal ↔ informal, serious ↔ playful, distant ↔ warm. The principle 'care without fluff' on the scale gives a specific point: warm but not playful tone.
• Editorial policy — rules of writing and ready formulations: how to format numbers, use quotation marks, use an exclamation mark, what to write in a modal confirmation window.
• Microtexts — final words on the screen: field label, error text, push notification.

The editorial policy itself is also structured into three parts: spelling and punctuation (the letter 'yo', declensions, number formatting), style dictionary (the degree of emotionality, addressing the user, permitted and forbidden words), and web components — templates for typical situations: error 404, empty list state, payment confirmation. Such structure relieves the editor of the task of inventing a tone from scratch for each button — they refer to a ready-made solution.

A telling example of flexibility within one voice — Monobank. In critical scenarios (card blocking, suspicion of fraud), the tone becomes serious and precise, without a single joke. In non-critical cases — friendly, with a mascot and light irony. The brand remains recognizable in both cases: the register changes, not the identity. This is the difference between a 'living voice' and a rigid template of one phrase for all situations.

Six steps that synchronize the platform and the interface

Developing a tone of voice, aligned with the brand platform, is a sequential process with a verifiable result at each step, not a brainstorming session with posters on the wall.

1. Extracting brand principles from the platform and translating each into the voice scale with a specific point on it.
2. Audit of existing interface texts: what already matches the desired voice, what contradicts it, and on which screens.
3. Voice dictionary — a list of words and constructions with 'yes/no' examples: for example, 'money will arrive tomorrow' instead of 'funds will be transferred within one operational day'.
4. Assembling the editorial policy from three layers: spelling, style, and component templates.
5. Testing the editorial policy on real interface scenarios — on live error and confirmation texts, not on abstract examples from the brief.
6. Implementation in the UI kit: texts become part of the component library next to color and typography, not a separate Word document that nobody opens.

For a company with dozens of channels — an app, a website, email, push, a chatbot, a call center — one policy document is not enough. A single brand dictionary is needed, to which the product team, marketing, and support team all refer simultaneously. Without it, the same product sounds different in the app and in a support email, and the user feels a disconnect, even if they can't name it.

How to distinguish a well-written text from a text that speaks in the brand's voice

Literacy and fitting into the brand's voice are different measures. A text can be stylistically flawless and yet completely fall out of the company's image. The check is based on three levels: qualitative compliance with the dictionary and policy, behavioral metrics — conversion, time on task, number of support requests due to unclear phrasing — and A/B testing, but only with a hypothesis based on user observation, not random synonym replacement.

Numbers here are more telling than arguments. According to Nielsen Norman Group, replacing a single word or pronoun in a button can increase clickability by 90%, and concise and objective text increases website usability by 124%. A classic example is the Booking.com test, where the button «I’ll book» beat «Book» not because of grammar, but because of psychology: the first-person formulation creates a sense of control at the moment of a stressful decision, and reducing anxiety was stated as the brand's principle.

There is also the reverse side — the cost of not having a system. According to Lucidpress and Demand Metric, consistent voice across all contact points raises revenue by 23–33%, but about 60% of marketing materials miss brand standards, and 81% of companies admit they struggle with deviations from the tone of voice. This is not a problem of individual editors — it is a symptom of the lack of a working platform to which one can appeal in a disputed situation.

Where meanings break down: the text says one thing, but the product behaves differently

UX is the direct carrier of the brand's promise, and the gap between the brand's declaration and the product's behavior happens precisely here, more often than in the formulations themselves.

• The brand claims transparency — but an error message sounds like «something went wrong», without a reason or next step.
• The brand promises efficiency — but the path to results consists of unnecessary confirmation screens that protect nothing and only slow down the user.
• The brand speaks of reliability — but in the event of a failure, the person is left without feedback: it's unclear whether the data has been lost and whether the action needs to be repeated.

The mistake of companies in such cases is not that the editor 'writes error texts poorly.' The problem is deeper: the semantic layer and product logic are developed by different teams without a common arbiter document. The editor writes a polite phrase that corresponds to the tone of voice, but cannot change the architecture of the scenario — and thus a correct message is obtained that the system still behaves in a non-transparent way.

In such a situation, the user does not make a concession for the complexity of the product and does not 'forgive' the brand for the mismatch between words and behavior. He quietly reduces the value of the product in his head — without complaints, but also without a second chance. Therefore, the brand platform should be a working document not only for copywriters and designers, but also for product managers who design the logic of scenarios, not just their formulations.