How to Audit a Product's Voice for Emotional Congruence With Its Brand Promise

A product's voice is every word a user reads from the moment they open the app to the moment they close it. Button labels, error messages, empty states, confirmation screens, push notifications, all of these carry emotional weight, and all of them either reinforce or undercut what the brand promised before the user arrived. Most audits focus on visual consistency or tone-of-voice guidelines applied to marketing copy. The product itself gets a different standard, or no standard at all, and that gap is where trust breaks down.

We worked on a health and wellness genetics app where this gap was the entire problem. The brand's printed packaging and online presence were luxurious and aspirational, built around transformation and reaching your full potential. The app users opened to view their results was cold, clinical, and purely functional. Every word inside the product contradicted what every word outside it had promised. Users had moved through ordering a kit, doing a test, and building genuine anticipation, and then the product voice stopped the journey entirely. That kind of misalignment is an emotional congruence issue, and fixing it requires a proper audit.

A product's voice must carry the emotional promise forward, not treat it as something that ends at the download screen.

This article walks through how to run that audit, how to define what the product is actually promising emotionally, how to score individual touchpoints against that promise, and what to do when the product and the brand were built by different people with different priorities.

Define the Emotional Promise Before You Touch the Product

An audit without a benchmark is just an opinion. Before you review a single line of copy, you need a precise statement of what the product promises to make people feel, not what it claims to do functionally. These are different things, and conflating them is how audits produce vague findings nobody acts on.

One framework we use for this is what we call the dinner party exercise. We ask clients to picture the product as a person at a dinner party. How do they dress? How do they talk? Are they the loudest voice in the room or the quiet, attentive listener in the corner? Once that personality is defined concretely, it becomes a usable benchmark. Applying abstract brand adjectives like "warm" or "aspirational" to individual design decisions is genuinely difficult. Asking whether a specific person would actually phrase something that way is not.

The emotional promise usually lives somewhere between the brand's stated values and the emotional problem the product exists to solve. A genetics wellness app promising transformation and personal discovery carries a very different emotional obligation than a productivity tool promising clarity and control. Get the promise written down as a single sentence before the audit begins. Something like: "This product makes users feel they are gaining a real, personal understanding of how their body works." That sentence becomes the standard every touchpoint is measured against.

Map the Touchpoints You Are Auditing

A voice audit can cover a lot of ground, and trying to audit everything at once without a structure means you will miss things or conflate issues that belong in different categories. Before scoring anything, list every touchpoint where the product speaks to the user and group them by type.

The five categories worth auditing separately are microcopy, error states, notifications and empty states, onboarding, and confirmation or success messages. Each category carries a different emotional function and a different risk of misalignment. Onboarding copy sets expectations and can over-promise relative to what the product actually delivers. Error states are where anxious, frustrated users encounter the brand at its most vulnerable moment. Empty states speak to users who have just arrived and do not yet know what to do.

Mapping these before you begin also helps you scope the project honestly. A full audit of a complex product might involve several hundred individual copy instances. Prioritise the categories where the emotional stakes are highest given the brand promise. For a product built around calm and reassurance, error states and onboarding carry more risk than confirmation screens. For a product built around achievement and progress, empty states and notifications matter most.

List every screen type before you start scoring. Auditing onboarding, error states, and notifications with the same rubric but separate scorecards keeps findings actionable and prevents them from blurring together in the final report.

Build Your Scoring Rubric

Once you have the emotional promise defined and the touchpoints mapped, you need a rubric that lets you score each instance consistently. Without one, the audit produces a list of opinions. With one, it produces a prioritised list of specific problems with clear reasoning.

We score each touchpoint on three dimensions.

Every line of product copy is either carrying the brand promise forward or quietly dismantling it.

The first dimension is emotional alignment. Does this copy make the user feel something consistent with the product's emotional promise? The second is clarity. Does the user know exactly what is happening and what to do next, without having to work for it? The third is voice consistency. Would the dinner party character actually say this, or does it feel like a different person wrote it?

Any touchpoint scoring a 1 on emotional alignment is a priority fix, regardless of how it scores on the other dimensions. A clear message that contradicts the brand promise is doing more damage than an unclear one, because users read it and believe it.

Auditing Microcopy Against the Brand Voice

Microcopy is the smallest unit of voice in a product: button labels, placeholder text, tooltip copy, field labels, helper text. Because each instance is short, teams often write it quickly and without reference to brand guidelines. The result is copy that is technically functional but emotionally disconnected from the product it sits inside.

The dinner party test is particularly useful here, because short copy is exactly where the benchmark earns its value. "Submit" is what a form does, not what a person says. "Continue your journey" is what a character with warmth and investment in the user might say. Neither is universally right, the right choice depends entirely on who the product is as a person. A clinical precision tool should probably say "Submit." A wellness app should probably not.

Work through each microcopy instance and ask three questions. Does this match the voice of the dinner party character? Does it carry any emotional signal at all, or is it purely transactional? And would a user in the emotional state they are most likely to be in at this point in the journey read this positively?

Read button labels and field labels aloud as if you are the dinner party character speaking them. If they feel out of character, they probably are. The ear catches register mismatches that the eye skips over.

Auditing Error States: Reframe Before You Redesign

Error states are the highest-risk touchpoints in any product, and they are almost always the most neglected. Users who hit an error are already frustrated or anxious. The copy they read at that moment either absorbs the emotional pressure or amplifies it. Most error messages amplify it, because they were written by someone solving a technical problem rather than an emotional one.

The most powerful intervention we made on the genetics wellness app was reframing error states before they became errors in the user's mind. When users saw unexpected genetic data, the default experience let them encounter the result cold and interpret it themselves. Many read variation as a mistake. We rewrote the surrounding copy to pre-empt that interpretation: this is to be expected, everyone is different. The error state effectively stopped occurring, not because we redesigned the data output, but because we changed the frame the user brought to it.

That principle applies across error types. Before redesigning an error message, ask whether the error state could be prevented through earlier framing. If a user is likely to misunderstand a result or a system state, the fix often belongs one screen earlier, not on the error screen itself. Where errors are unavoidable, the copy should name what happened plainly, explain what to do next, and never make the user feel they caused the problem through carelessness.

The Three-Part Error Message

A useful structure for error copy is: what happened, why it matters to the user's goal, and what to do next. All three in plain language, all three written in the product's voice. "Something went wrong. Please try again." fails on all three counts. It is vague about what happened, silent about why it matters, and dismissive of the user's next step.

Auditing Notifications and Empty States

Notifications and empty states occupy opposite ends of the same emotional challenge. Notifications interrupt users who are elsewhere. Empty states greet users who have arrived and found nothing waiting for them. Both moments carry real risk of emotional misalignment, and both are frequently written as afterthoughts.

For notifications, the audit question is whether each message is timely, relevant, and written in the product's voice. A notification that arrives at the wrong moment, says something generic, or uses a different register from the rest of the product creates friction and erodes trust. Teams often send notifications at the frequency that suits the product's metrics rather than the frequency that serves the user. Part of the audit is evaluating whether the notification strategy as a whole is working with users or against them.

Empty, and they are typically written last. A new user who opens a fitness tracker and sees "No workouts logged yet" has been told nothing useful and felt nothing positive. The same screen could say something that reflects the product's character, acknowledges where the user is, and points clearly toward the first action. The emotional promise applies here as much as anywhere else in the product. Check whether empty state copy is written in the product's voice and whether it treats the user as someone at the beginning of something worth starting, rather than someone who has failed to fill a database table.

When the Product and the Promise Are Built by Different People

A consistent finding across voice audits is that misalignment between brand promise and product voice almost always traces back to a structural problem rather than a craft problem. The marketing team defined the emotional promise. The development team built the product. And the two groups had different briefs, different timelines, and different definitions of what done looks like.

On the genetics wellness app, this was precisely what had happened. The brand promise was actually well-constructed, the narrative of offering a real inside look at how your body works and how it is ageing was compelling and credible. But the product had been built primarily by a developer team, and in that process the storytelling and narrative had been stripped out in favour of functional delivery. The result was a product that worked technically but left users unable to connect with their data personally. Retention suffered not because the content was wrong but because it was presented without a human frame.

When an audit reveals this pattern, the recommendation is rarely to rewrite the brand promise. The promise is usually sound; the problem is that it never made it into the product. The fix involves bringing the narrative back in and giving users a sense of personal identity within the product experience, rather than a data readout that could belong to anyone.

If the audit reveals a structural gap between who built the brand and who built the product, flag it explicitly in the findings. Fixing individual copy instances without naming the root cause means the problem returns with the next feature release.

Worked Example: A Health and Wellness Genetics App

The genetics wellness app project illustrates most of the audit principles in sequence. We were brought in because the brand experience and the product experience were in direct emotional conflict. The packaging and website created genuine aspiration. The app, which users opened to see deeply personal health data, delivered it in a register that belonged to a technical report rather than a personal discovery.

We ran the dinner party exercise with the client team to establish the product's intended character before touching any copy. The character that emerged was knowledgeable and warm, like a trusted expert who explains complex information in a way that makes you feel capable rather than overwhelmed. Then we mapped every touchpoint in the app against that character and scored each one on the three-dimension rubric.

Focus groups and five-second tests measured comprehension and emotional response before and after the rewrite. Initially, comprehension scores were low, partly because the volume of content and the cold framing of unexpected data results left users unsure what they were reading and whether something had gone wrong. After rewriting to create a more narrative experience and pre-framing potential issues, a second round of user testing showed meaningful improvements in clarity, sense of purpose, and retention. The changes were not structural. The data, the features, and the visual design stayed largely the same. What changed was the voice carrying users through their results, and that shift alone moved the product from a functional report into something that felt personal.

What to Do With Your Audit Results

An audit produces findings, and findings without a clear action plan stay findings. The output of a voice audit should be a prioritised list of specific changes, not a general observation that the product needs a warmer tone.

Organise findings into three tiers. Tier one covers instances where the copy actively contradicts the emotional promise, such as an onboarding screen that promises calm while immediately pressuring users to complete a long goal-setting exercise, or an error message that places blame on the user. These are the highest priority and need rewriting before anything else. Tier two covers instances that are emotionally neutral, where the copy is functional but carries no signal and no personality. Tier three covers instances that are broadly aligned but off in register or phrasing, the small adjustments that move the overall voice from approximately right to precisely right.

1. Fix tier one first. Active contradiction does the most damage.
2. Rewrite tier two with the dinner party character in mind for each instance.
3. Adjust tier three last, and use it as a final consistency pass.
4. Document the reasoning for each change so future copy decisions have a precedent.
5. Build the dinner party character and the emotional promise into the product's ongoing design process, so new features start aligned rather than needing retroactive audits.

Emotionally connected customers deliver between 25% and 100% greater value in revenue and profitability compared to those who are merely satisfied, according to Harvard Business Review, 2015. That range is wide, but the direction is not in doubt. A product voice that sustains the emotional promise is a retention and revenue decision.

Conclusion

A voice audit is a structured exercise in honesty. It asks whether the product is keeping the promise the brand made before the user arrived. Frequently, the answer is partial at best and contradictory at worst, not because anyone made bad decisions, but because the people who defined the emotional promise and the people who built the product were working from different briefs.

The dinner party exercise, the three-dimension rubric, and the tier system for prioritising fixes are all tools for closing that gap. They work because they give teams a concrete, human benchmark rather than an abstract one. "Would this person actually say that?" is a question anyone on a product team can answer in ten seconds. "Does this copy achieve emotional resonance?" is a question that produces a two-hour meeting and no decisions.

The genetics wellness app work showed what is possible when the audit findings are acted on fully rather than partially. Narrative and personal identity restored through copy changes, with no structural redesign, moved the product from one that worked to one that retained. The data was always there. The features were always there. What was missing was a voice that made users feel the product was speaking to them.

If your product's voice and your brand's promise feel like they belong to different products, the audit is the right place to start. Let's talk about your product's voice.

Frequently Asked Questions