Sentiment

Sentiment shows how your design makes people feel in the moment. It captures emotional reactions like delight, confusion, frustration, or curiosity before users start explaining or analyzing what happened.Use this metric early, when first impressions are critical. It works well in concept testing, visual reviews, and after key interactions like onboarding or checkout. You can also use it to compare design directions or track how emotional responses shift over time.Sentiment gives you a fast and honest signal. It helps you catch how users truly feel, especially when the experience looks fine on the surface but doesn’t leave the right impression.Interpreting the ResultsUse this key to understand what your Sentiment score means and how to interpret that for your product experience:How to Calculate SentimentThe Sentiment metric helps quantify how positively or negatively users feel about a design by analyzing the impressions they select in response to visual or content stimuli.Set up questionsTo collect sentiment data, use a multiple choice question with no selection limit. Participants are shown an image or screen and asked to select all the words that describe their impression.This question can be included in a visual design or branding test and distributed to your target audience using a remote survey platform.Collect dataIn this example, participants viewed brand imagery from a university site and selected from a list of adjectives.Some participants chose positive words like Successful, Relevant, and Diverse, while others selected negative impressions such as Confusing, Unprofessional, or Boring.Plug data into formulaTo turn these impressions into a measurable score, you’ll use the Sentiment formula:The total pool of impressions includes both positive and negative word selections.
 This step ensures that the score reflects the balance of positive feedback in relation to all impressions provided.Calculate the Sentiment ScoreThis score represents the proportion of positive impressions out of the total impressions.In this example, the 92% Sentiment score is Very Good on a scale from Very Poor to Very Good, meaning participants overwhelmingly associated the design with positive qualities.When to Use Sentiment MetricsSentiment UX metrics provide insights into how users feel about your product designs, helping you measure emotional responses, brand perception, and overall satisfaction. By evaluating user sentiment, you can pinpoint areas where designs resonate or fall short, ultimately improving user trust, engagement, and loyalty.Navigation UpdatesSentiment metrics can reveal how users feel about changes to a website's navigation, such as a redesigned menu or updated search functionality. Positive sentiment indicates users find the navigation intuitive and efficient, while negative sentiment might point to confusion or difficulty finding key information. Tracking this data ensures that navigation updates enhance user experience rather than creating frustration.Visual Design OverhaulsWhen refreshing the visual design of a website—such as introducing a new color palette, font styles, or layout changes—sentiment metrics help measure user reactions. A positive response indicates that the changes resonate with users and align with brand perception, while negative sentiment might signal that the new look feels inconsistent, unappealing, or less functional. This feedback ensures design updates strengthen the overall user experience.AnimationsSubtle design features like hover effects, loading animations, or progress indicators can greatly influence how users feel about a website. Sentiment metrics can gauge whether these microinteractions enhance the user experience by making it feel polished and engaging or whether they frustrate users by being distracting or unnecessary. This feedback allows designers to create interactions that delight users without compromising functionality.How Indiana University Used Sentiment Testing to Validate Brand VisualsAs part of a major website rebrand, Indiana University needed to ensure its new brand imagery reflected the right emotional tone and visual identity. The design team wanted to move beyond subjective creative reviews and instead ground their decisions in how real users actually felt when encountering the visuals. Using the Sentiment UX metric, they tested three distinct image styles to find out which direction best aligned with the university’s values.The SetupParticipants were shown three versions of potential brand imagery that could be used across Indiana University’s updated website. Each set was presented as part of a simple layout mockup, and users were asked to select how the visuals made them feel using eight impression options—four positive (Diverse, Determined, Relevant, Successful) and four negative (Confusing, Unprofessional, Unrealistic, Boring). The Sentiment score was calculated as the percentage of positive impressions out of all selected impressions.The ResultsThe responses from participants produced the following data set:The data from the multiple choice question above was plugged into the Sentiment formula to reveal the score:One image set scored a Sentiment rating of 92%, far surpassing the team’s expectationsThe most frequently selected impressions were Successful (60%), Relevant (56%), and Diverse (43%)Less than 10% of impressions were negative, with only small percentages marking the imagery as “Boring” (7%), “Unprofessional” (4%), or “Confusing” (3%)Participants noted that the selected imagery “felt authentic and inspiring” and “showed students in real, empowering moments”The ImpactThe creative team moved forward with the winning image direction as a cornerstone of the rebrand. These visuals now appear across high-visibility pages like the homepage and admissions section, reinforcing key themes of academic success and campus inclusivity. By backing their creative decision with sentiment data, the team was able to align brand perception with audience emotion—and do so with confidence.SourceHelio SurveyCSVHow to Use AI to Measure SentimentUsing the multiple choice question outlined in the How to Calculate section above, gather responses on a survey from an audience of at least 100 respondents. We find that 100 responses is statistically significant in most markets. Once the responses are collected, download the CSV file of your data report and upload it into an AI platform along with the prompt below.Copy this AI prompt to calculate your own Sentiment score, and check out the type of output it would produce:Technicals for Measuring SentimentWe’ve built out a UX Metric framework that we’re using in our Helio platform. Here, we’ve laid out what we’ve done, and how other developers can use this. You can also become a Glare Code Contributor on ourUX Metric framework.How to UseSentiment Data ParserWith our framework, you can use theGlare::UxMetric::Sentiment::Parsermodule to parse Sentiment data.Here are the steps:First, we have to require the module to use it.Then, we need to pass in a valid data structure into the choices parameter.When callingGlare::UxMetric::Sentiment::Parser.new, you get aParserinstance. This grants access to the parse and valid? methods.Use parse to obtain a score result, known asGlare::UxMetric::Result.Templates & Presentation MaterialsCreate effective presentation slides, document design concepts, and implement UX Metrics with templates and resources.We've done the work to provide professional layouts that communicate to your stakeholders. UX Metric cards clearly communicate the totals, allow space for breakdowns, and styled to allow for your own brand.Visit Findings for TemplatesResourcesThe Resources section provides a collection of articles, case studies, methods, and blog posts to support your work within the UX metrics framework. These materials offer insights into best practices, research methodologies, and practical applications for improving design comprehension and usability. Whether you're refining your design process or conducting user research, these resources will help guide you towards data-informed, user-centered decisions.ArticlesSentiment Analysis: Uncovering Emotions in Text with a Real-World Case Studyby, Data OverloadUsing Sentiment Score to Assess Customer Service Qualityby,Shuai ShaoTailor the model to address challenges like skewed negative data and multilingual inputs, and aggregated raw sentiment scores using methods calibrated to align well with NPS while offering faster, more actionable insights.Helio MethodsVideo Testingby HelioInteraction Matrixby HelioHelio Case studiesHelloFresh Membership Offer Effectivenessby HelioValidated Banking Site Landing Page Concepts, by HelioHelio blog postsMastering Copy Testing: Your UTake This Further with the UX Metrics AI SkillsSentiment measures the overall tone of how users feel about your product using survey questions turned into a single number score. TheUX Metrics AI Skillsis a package you load into your LLM so you can ask questions and get expert answers anytime.Write survey questions that capture positive and negative sentimentUnderstand what is shifting sentiment across your productCompare sentiment scores before and after changesConnect sentiment data to prioritization decisionsDrop it into your LLM and start asking questions right away.