# Message test — https://ux.snu.ac.kr/ After reading your page, only 0 of 15 personas could quickly tell what problem it solves and who it is for. - **Page tested:** https://ux.snu.ac.kr/ - **Audience tested against:** 대학 학부 졸업 예정인 서울대학교 언어학, 정보문화학 전공 학생. 20대 중반 남성. 관심사는 본전공과 유관한 NLP 관련 HCI 연구. 연구실의 교수님을 실제로 뵌 적은 없고, 이런 hci 연구실이 있다는 얘기만 들었음. 다음 학기 대학원 지원을 위해 관심 연구실을 물색 중. - **Personas:** 15 simulated - **Report:** https://grader.wynter.com/r/home-snu-ux-lab-_BneSwE > These answers are generated by AI, scored on Wynter's B2B Message > Layers framework using behaviorally-diverse simulated personas. The > methodology is real and the critique is directional. What a simulated > persona cannot have is a live budget, a renewal coming up, or a boss > asking about this quarter. --- ## 01 · The scores Every persona answered all four questions. These are four independent proportions of the same panel, not stages of a funnel. | Layer | Question | Cleared the bar | Strength | Of those who passed | | --- | --- | --- | --- | --- | | 1. Clarity | Do they understand what you do? | 7/15 | 90% | 12 without hesitation, 2 with reservations | | 2. Relevance | Can they tell what it solves, and who it's for? | 0/15 | 4% | — | | 3. Value | Do they actually want it? | 1/15 | 19% | with reservations | | 4. Differentiation | Is there a reason to pick you over the alternatives? | 0/15 | 16% | — | **Brand alignment** (a side metric, not one of the four layers) — 6/15, 52% strength (all without hesitation). Does the page read like the company you actually are? **Fix first: Clarity.** Earliest failing layer, walking the sequence in order — not simply the lowest score. ### What they thought you sell 8 of the personas who named a category got it wrong: - 3× “Academic UX research lab” - 1× “Academic UX research lab / consultancy” - 1× “Academic UX research lab / consulting” - 1× “Academic UX research lab / consulting projects” - 1× “Academic UX research lab / research consultancy” - 1× “University UX research lab” --- ## 02 · What to change, layer by layer Ordered worst-first. Specific edits, not a restatement of the score. ### Relevance **Add a short problem statement before the project list describing what clients come with.** The page jumps straight to project summaries without naming the problem a buyer already feels. State it plainly, such as teams shipping AI assistants with no reliable way to test whether the conversation actually works. *effort medium · impact high · tested against Problem before solution* **Add a line above 'Recent Project' naming who the lab works with.** Nowhere does the page say who it is for, so a product or quality lead has to guess whether it applies to them. Name the roles and teams, such as conversational AI product and quality teams launching voice or chat assistants. *effort low · impact high · tested against Name the audience* **Add a contact or project inquiry link in the header and after the project list.** There is no visible way to start a conversation, so an interested reader has nowhere to go. Add one inquiry link with a clear label about commissioning a project. *effort low · impact high · tested against One clear next action* ### Differentiation **Expand one LG project entry into a full case study with method and results.** Named clients give trust but nothing behind them, so a buyer has to take the roster on faith. Build out one project with the question asked, the method used, and what the client did with it. *effort high · impact high · tested against Proof next to the claim* **Add a line stating what the lab's evaluation methods do that standard usability testing does not.** Nothing on the page says why this lab rather than an agency or an in-house team. Name the specific edge, such as validated evaluation criteria for voice and chat agents built from repeated enterprise studies. *effort medium · impact high · tested against Give a reason to choose you* **Add a client logo or name strip near the top listing LG전자 and 삼성전자.** The enterprise client names are buried inside project metadata where a scanning reader misses them. Pull the names up so the track record is visible before any body copy. *effort low · impact high · tested against Proof next to the claim* ### Value **Add an outcome line to each project entry stating what the client received.** The LG and Samsung entries describe background context but never say what was delivered or what changed. Add one line per project naming the deliverable, such as an evaluation framework, test protocol, or benchmark scores. *effort medium · impact high · tested against Tie the feature to the outcome* **Add a short section describing how an evaluation project runs and how long it takes.** A buyer cannot tell what working with the lab involves, what methods are used, or how long it takes. Describe the steps of a typical evaluation and a typical timeframe. *effort medium · impact high · tested against Answer the live objection* **Rewrite project summaries to lead with the finding, not the market background.** Each summary opens with industry context about LG's AI initiative before reaching anything the lab did. Put the result or method first and the background after. *effort medium · impact medium · tested against Conclusion first* ### Clarity **Add a one-line statement under the SNU UX LAB title naming what the lab does for companies.** The page opens with a name and a project list, so a reader cannot tell whether this is a research archive or a service they can hire. Write a line saying the lab evaluates and validates conversational AI user experience for companies. *effort low · impact high · tested against Lead with the use case* **Replace '[종료]' tags on project titles with the year and the client outcome.** Every project is stamped closed, making the work read as finished case files rather than capabilities available now. Show the delivery year and what the client got instead of a closed marker. *effort medium · impact medium · tested against Concrete over abstract* **Rewrite the 'Recent Project' heading to state the kind of work shown.** 'Recent Project' tells a scanner nothing about what kind of work the lab does. Change it to name the work, such as voice and chat UX evaluation projects for enterprise clients. *effort low · impact medium · tested against Front-load the meaning* ### Brand alignment (side metric) **Move the Lab Stories intern posts to a separate careers or lab life page.** Intern diary posts about dinner and bookshelves sit directly under enterprise project work, making the site read as recruiting material. Keep the homepage on client work and link to the stories from a recruiting page. *effort medium · impact high · tested against Name the audience* **Rename 'Lab Stories' to label it as student and recruiting content.** The heading gives no clue that what follows is internal culture writing, so a corporate reader reads it as the lab's professional voice. Label it clearly as life in the lab for prospective students. *effort low · impact medium · tested against Headings stand alone* --- ## 03 · What is working ### Named client work for LG and Samsung is the one concrete asset respondents believed Four respondents cited the client roster — LG ThinQ and Samsung conversational AI UX work — as evidence of enterprise capability and prior voice/chat validation experience. One called it the only concrete differentiator on the page. > The one thing that would actually pull me toward shortlisting them is the named client list with specific project titles — "LG전자 고객품질연구소" on the ThinQ On voice agent work and "삼성전자 경험기획그룹" on the Clothing Care pilot. > > — HCI Researcher, Higher Education, 51-200 > they've done paid work for LG전자 and 삼성전자 on exactly my domain (voice/chat agent UX evaluation for ThinQ On and ChatThinQ) > > — Graduate Research Assistant, Technology, 51-200 > The one thing that'd tip me toward shortlisting them over a generic UX consultancy is the client roster embedded in the project titles—LG전자 (twice, for ThinQ On voice agent and ChatThinQ), 삼성전자, 콘텐츠진흥원 > > — HCI Researcher, Technology, 51-200 > The one concrete thing that would pull me toward actually contacting them is the named client work — "LG전자 고객품질연구소" on the ThinQ On voice-agent UX evaluation > > — User Experience Specialist, Technology, 51-200 --- ## 04 · What the personas said ### The page reads as a university lab portfolio, not a commercial offering Seven respondents identified the site as an academic research archive or project portfolio rather than a product, service, or vendor offering. Several noted project titles read as internal research reports with no value proposition attached. > This isn't a product at all — it's a university research lab page. SNU UX LAB (Seoul National University) is an academic HCI/UX research group > > — Graduate Research Assistant, Technology, 51-200 > The project titles themselves are the blocker — things like '음성 중심 대화 경험의 UX 평가 요소와 검증 방법 연구' read as dense academic-paper language, not a pitch > > — User Experience Specialist, Higher Education, 51-200 > This is a university research lab, not a product at all — it's Seoul National University's UX Lab, doing applied research projects with corporate partners like LG and Samsung > > — User Experience Specialist, Higher Education, 51-200 > This isn't a product at all — it's a university research lab's site. > > — UX Researcher, Research and Development, 51-200 > things like '[종료] Voice 중심 대화 경험의 UX 평가 요소와 검증 방법 연구' read as internal research-report headlines, not pitch copy, so there's no verb like 'we help' or 'we offer' anywhere > > — HCI Researcher, Technology, 51-200 > University UX research lab doing client projects for LG, Samsung etc. Not a product at all. > > — Interaction Designer, Higher Education, 51-200 > This isn't a product at all — it's a university research lab's site, specifically Seoul National University's UX LAB. > > — HCI Researcher, Research and Development, 51-200 ### Closed-project tagging and missing deliverables make the work feel archived and… Respondents flagged that '[종료]' closed tags frame all research as finished case files, and that no measurable outcomes or deliverables from past work are stated. > The project blurbs literally cut off mid-sentence—'기능적 성능과 사용 효능감(UX)을...'—so I never get the actual claim or outcome > > — UX Researcher, Higher Education, 51-200 ### Audience and problem statement must be reverse-engineered Seven respondents said no target audience, buyer problem, or market use case is stated anywhere, forcing them to infer purpose from context. One noted there is no visible engagement path such as a contact form or commissioning button. > I'd need a line near the top like 'we help product teams validate conversational AI UX before launch' plus a visible way to engage — a contact form, a rate card, or 'commission a study' button > > — User Experience Specialist, Higher Education, 51-200 > There's no problem statement or target-reader line anywhere on this page because it's not selling anything > > — Graduate Research Assistant, Technology, 51-200 > Honestly, there's no "problem solved" or "who it's for" stated anywhere because this isn't selling anything — it's a lab homepage listing research projects > > — User Experience Specialist, Higher Education, 51-200 > It wasn't spelled out as a "problem/solution" at all — this reads as a lab homepage, not a pitch, so there's no stated pain point to solve. > > — UX Researcher, Research and Development, 51-200 > this is just a university lab's project portfolio, not a tool solving a problem for a defined buyer > > — Interaction Designer, Higher Education, 51-200 > there's no stated problem or audience, just a list of project titles and client names like "LG전자 고객품질연구소" or "삼성전자 경험기획그룹" with a sentence or two of context each > > — Graduate Research Assistant, Research and Development, 51-200 > the 'who it's for' (students wanting to intern, or companies wanting to commission UX studies) is never stated outright — I pieced it together from context > > — User Experience Specialist, Technology, 51-200 ### No methodology, metrics, pricing, or engagement model is given Six respondents said the page omits methodology details, outcome metrics, pricing, timelines, capabilities statements, and benchmarking against current evaluation methods, leaving them unable to assess value or buy. > there's no pricing, engagement model, timeline, or methodology detail on the page, so I can't tell what it would cost my team in time or budget, or how an engagement would actually run. > > — HCI Researcher, Higher Education, 51-200 > There's no stated methodology, no comparison to how I'd evaluate conversational-AI UX today, no metric like "reduces eval cycle time" > > — Interaction Designer, Research and Development, 51-200 > there's no methodology writeup, no published findings, no sense of turnaround time or cost, just client names and a sentence of project description > > — Graduate Research Assistant, Research and Development, 51-200 > there's nothing on this page telling me what that methodology actually is, what the evaluation criteria were, or any outcome data from the LG or Samsung engagements > > — HCI Researcher, Research and Development, 51-200 > it's not a tool or service with a value prop, it's a lab's project archive and student blog > > — Graduate Research Assistant, Higher Education, 51-200 ### No product feature list, demo, or comparison signals exist for vendor evaluation One respondent noted the absence of any feature list, demo link, or vendor comparison cues that would let a buyer evaluate the lab against alternatives. > It's not a vendor page with a product, pricing, or deployment claim; it's SNU's lab page listing project case studies like "LG전자 고객품질연구소" and "삼성전자 경험기획그룹" > > — User Experience Specialist, Higher Education, 51-200 ### The tone targets students and recruits, not corporate buyers Five respondents said the intern blog and Lab Stories content is internal culture and recruiting material aimed at prospective students. The mix of dense academic summaries with casual intern diary posts was called misaligned with professional buyers. > half the page is intern blog diary entries about packing for a conference or chopping clams, which is charming for students browsing the lab but tells me nothing about engagement model, cost, or methodology. > > — HCI Researcher, Higher Education, 51-200 > The tone definitely isn't written for someone like me — it swings between dry academic project summaries ("Voice 중심 대화 경험의 UX 평가 요소와 검증 방법 연구") and chatty, diary-style intern stories about book shelves and plastic bags > > — User Experience Specialist, Higher Education, 51-200 > writes for grad students/interns, not buyers like me > > — Interaction Designer, Higher Education, 51-200 > half the page is intern diary posts about book shelves and plastic bags over someone's face during a study, which reads like it's written for prospective students or the lab's own community, not for a company deciding on a research partnership > > — Graduate Research Assistant, Research and Development, 51-200 ### The client roster is credibility without evidence behind it Four respondents said named clients signal trust but lack full case studies, complete project descriptions, or measurable outcomes, and would require independent confirmation to verify. > The only semi-concrete signal is the client roster — LG전자 고객품질연구소 and 삼성전자 — which at least tells me this lab has done work with large Korean corporates > > — Interaction Designer, Research and Development, 51-200 > The one thing that would actually tip me toward this lab over a competitor is the specific client roster tied to specific named products — "LG전자 고객품질연구소" on the 'LG 씽큐 온' voice agent project > > — Graduate Research Assistant, Research and Development, 51-200 > seeing "LG전자 고객품질연구소" and "삼성전자 경험기획그룹" actually commissioned repeat work (two separate LG projects, in 2024 and 2025) tells me real companies trust them enough to pay twice > > — UX Researcher, Higher Education, 51-200 > the "Lab Stories" intern blog posts actually undercut confidence by showing this is largely a training ground for students, not a seasoned consultancy bench. > > — UX Researcher, Research and Development, 51-200 --- ## 05 · The hardest read An adversarial pass over the findings. Every claim below was checked against the panel's own answers; unsupported ones were dropped. - **The page does not read as a commercial offering at all, so no amount of message tuning will help until it is reframed as a service.** *(high)* Six respondents read the site as an academic archive or project portfolio rather than a product or vendor offering, with project titles functioning as internal research reports carrying no value proposition. - **Buyers are being asked to do the selling work themselves, and most will simply leave instead.** *(high)* Six respondents had to reverse-engineer audience, buyer problem, and use case from context, and one found no contact form or commissioning path to act on even after inferring purpose. - **The page cannot survive a procurement conversation because every input a buyer needs is absent.** *(high)* Six respondents cited missing methodology, outcome metrics, pricing, timelines, and capability statements; a further respondent found no feature list, demo, or comparison cues for vendor evaluation. - **The one asset that works is being wasted.** *(high)* Four respondents named the LG ThinQ and Samsung work as the only concrete differentiator, yet four also said it sits without case studies, project detail, or measurable outcomes and would need independent verification. - **The page serves two audiences and converts neither.** *(medium)* Five respondents read the intern blog and Lab Stories as recruiting content for prospective students, and called the mix of dense academic summaries with casual intern diaries misaligned with professional buyers. - **The site signals the lab is not currently taking work.** *(medium)* Respondents flagged that '[종료]' closed tags frame all research as finished case files with no stated deliverables, compounding the absence of any visible engagement or commissioning path. --- ## 06 · Who answered | # | Role | Industry | Company size | | --- | --- | --- | --- | | 1 | HCI Researcher | Higher Education | 51-200 | | 2 | Interaction Designer | Research and Development | 51-200 | | 3 | Graduate Research Assistant | Technology | 51-200 | | 4 | User Experience Specialist | Higher Education | 51-200 | | 5 | UX Researcher | Research and Development | 51-200 | | 6 | HCI Researcher | Technology | 51-200 | | 7 | Interaction Designer | Higher Education | 51-200 | | 8 | Graduate Research Assistant | Research and Development | 51-200 | | 9 | User Experience Specialist | Technology | 51-200 | | 10 | UX Researcher | Higher Education | 51-200 | | 11 | HCI Researcher | Research and Development | 51-200 | | 12 | Interaction Designer | Technology | 51-200 | | 13 | Graduate Research Assistant | Higher Education | 51-200 | | 14 | User Experience Specialist | Research and Development | 51-200 | | 15 | UX Researcher | Technology | 51-200 | --- ## 07 · Before you act on this The methodology is real, and the critique is directional. What a simulated persona cannot have is a live budget, a renewal coming up, or a boss asking about this quarter. **Validate anything you're betting on with real ICPs who are actually in-market.** Being wrong is more expensive than you think. Finding out is cheaper than you'd guess. Wynter runs message testing with verified B2B professionals — trusted by HubSpot, RingCentral, Shopify, Cognism, Paddle, Veeam, Rippling and Miro. This report is kept for 60 days from 2026-10-19, then deleted along with the personas and their answers.