Brand message analysis: how customers actually understand you

Short answer

To measure how customers understand your brand message, look at what they call you, what they expect from you and which of your promises they repeat back. If you present yourself as a "premium service" and customers write "what's your cheapest option?", the message has not landed. If you promise "fast delivery" and customers never mention it, the promise has not stuck.

Four signals are measured in conversations: confusion (the customer does not understand what you do), wrong name (they confuse you with another brand or category), wrong expectation (they expect something you never promised) and promise echo (they repeat your main message in their own words). This is not campaign analysis but the overall perception of the brand.

Why a survey is not enough

Brand perception is usually measured with a survey: "describe us in one word". Surveys are useful, but artificial: the customer thinks before answering and often writes the expected answer. In a conversation, the customer describes your brand unprompted, at a moment of real need: "do you do installation too?", "do you have a subscription like X?". Those sentences show the real mental model customers have of you.

Signal 1: confusion

The customer does not understand what you sell or how you work. Signs: "what exactly do you do?", "is this software or a service?", "how do I buy from you?", "where is your store?" (if you are online only). A high confusion share means the website, profile description or the ad's main sentence is unclear.

Signal 2: wrong name and category

The customer confuses you with a competitor or another category: "are you company X?", "do you give loans?" (you offer a payment plan, not credit). It is either a similar name or a mispositioned category. Both need extra explanation in the sales conversation and sometimes create a risk of legal misunderstanding.

Signal 3: wrong expectation

The customer expects something you never promised: "you're open 24 hours, right?", "is there a free trial?", "returns are 30 days?". An expectation created by one campaign is a separate topic (campaign message mismatch); here it is a general wrong idea about the brand, tied to no specific ad. It often comes from competitors' standards or older messages.

Signal 4: promise echo

The most positive signal: the customer states your main promise in their own words. "I heard you deliver the same day", "my friend said your warranty is long". It shows the message not only arrived but was passed on. Which promise is echoed most actually shows what your brand is remembered for — which may not be the main message you chose.

Setting up the fields

1. 1Brand signalConfusion / wrong name or category / wrong expectation / promise echo / none.
2. 2Which promiseA closed list of your main brand promises: price, speed, quality, warranty, service, other.
3. 3Topic of wrong expectationOpening hours, returns, trial, delivery area and so on.
4. 4ChannelConversation data — which perception dominates in which channel.

Perception differences by channel

The same brand can be perceived differently on different channels, because customers met you in different places. A customer from Instagram knows you from a visual ad and guesses more about price; a customer who reached WhatsApp from the website has already read the terms. Split the signals by channel: if confusion is high on only one channel, the problem is not the overall brand message but the presentation on that channel — the profile description, the automatic greeting or the ad format. That makes the fix cheaper and faster, because changing one channel's copy is enough.

A feedback loop to marketing

1. 1Monthly summaryThe share of the four signals, comparison with last month, differences by channel.
2. 2Customer sentencesFive real sentences per signal — more persuasive than numbers.
3. 3One questionOne concrete question for marketing: "why do customers think we're open 24 hours?".
4. 4Change and checkThe website, profile or ad copy changes; the signal share is tracked next month.

Testing a new message

When the brand message changes (a new slogan, a new positioning, a new product line), conversation signals are the fastest source showing how it is perceived. Collect a 4-week baseline before the change: the confusion share and which promise is echoed. Track the same measures after the change. If the new message increases confusion, it will show in the first weeks.

Illustrative example

This is an illustrative example. A cleaning company presents itself as a "premium service with eco-friendly products". In conversations there is almost no promise echo, but plenty of "what's the cheapest package?" and "what's the hourly rate?". In the wrong-category signal, customers confuse it with a platform for individual cleaners.

Marketing changes the profile description and the website headline: "package service, fixed price, insured team". Two months later, the wrong-category share and the share of the "hourly rate" question are compared.

Typical mistakes

• Measuring brand perception only with surveys or social media comments.
• Treating a few memorable sentences as the overall perception.
• Reading a wrong expectation as the customer's fault.
• Not collecting a baseline before a message change.

Limits

• Conversations only show the perception of people who write to you; those who do not know you are absent.
• Brand signals are relatively rare — watch for random swings in small numbers.
• Social media reviews and public discussion are a different source, tracked with a separate tool.

Brand signals in Vexvon

In Vexvon the company builds fields such as "brand signal", "which promise" and "topic of wrong expectation" itself; AI picks from the customer's words in each conversation and keeps the sentence as evidence. The panel's AI assistant gives the share of signals, a cross-tab by channel and a before-and-after comparison around a message change; the result can be sent to the marketing team as a PDF. More: Vexvon analytics.

Next step

Read the first two messages of 60 conversations from last month and note which of the four signals appears. The most repeated wrong expectation will be your first question for marketing. We can set up the brand fields on your own conversations in a demo.

Further reading on this topic: customer intent analysis, customer conversation market research.