A formal UX research report. Four user interviews, forty-five survey responses, six months of analysis. What was tested, what was found, what changed — and what stayed exactly the way it was, on purpose.
The original site had a 68% form abandonment rate and 73% mobile traffic the layout couldn't handle. Four interviews surfaced a deeper pattern: users weren't bouncing because the form was hard — they were bouncing because the page hadn't earned the right to ask. The redesign moved trust signals up, made pricing concrete, and kept the form at seven fields because grave care actually needs that data. Season two booked roughly fifty clients, mostly one-off services.
Sheffield Grave Care launched its first season with a website built on designer intuition. Bookings were minimal. The site looked finished but wasn't working. Rather than guessing what to change, I ran a six-month research engagement — four in-depth interviews, forty-five survey responses, qualitative coding, and quantitative behavioural analysis.
The research did not surface what we expected. The form was not the wound. The lack of context around it was. Pricing buried below the fold read as a trap. Veteran-owned status in the footer read as a hedge. Stock photography on a grief-care site read as worse than no photo.
This report documents the research questions, the methodology, the findings, the design decisions, and the measured outcomes — including the methodological honesty about how the first version came to exist.
Every interview, every survey, and every line of behavioural analysis was directed at answering one of three questions. Naming them upfront keeps the methodology honest — and keeps the findings traceable.
What was causing the 68% form abandonment rate?
Was it the form itself, the page around it, or something earlier in the funnel?
Why was mobile bounce so high on a service most users access on their phones?
And what was the actual context of that mobile usage?
Which trust signals influenced booking decisions most?
And where were those signals positioned on the original site?
The research combined qualitative and quantitative methods to triangulate findings. No single method would have surfaced the trust-deficit insight on its own — it took interview voices, survey scale, and behavioural data converging on the same conclusion.
In-depth, semi-structured. 45–60 minutes each. Audio recorded, transcribed, thematically coded.
Mixed-format. Likert scales for satisfaction, open text for grievances. Distributed via local channels.
Discovery (2 months), design and testing (3 months), launch and iteration (1 month).
All participants were prospective or actual customers within the Lorain County service area.
The original site was built before formal research. The honest version of this report is that I designed first, then researched what was broken. This reversed order turned out to be the most useful methodological choice of the project — it gave a sharp comparison between which design instincts were right, which were defensible-but-suboptimal, and which were dead wrong. The research's job was not to validate the site. Its job was to expose the gap between my assumptions and the user's reality.
I was overwhelmed trying to figure out the pricing. I just wanted to know if you could help clean my mother's headstone, but the website made it feel complicated.
I needed to know how much it would cost before I could even think about booking. When I couldn't find clear pricing, I just gave up and called someone else.
The form asked for so many details upfront. I just wanted to ask a simple question about cleaning my father's headstone, but it felt like I was filling out a mortgage application.
All four interviews surfaced the same complaint in different language: the page didn't earn the right to ask before it asked. The form was a symptom. The trust deficit was the diagnosis.
The original site's form completion rate was catastrophically low. The intuitive read was: too many fields, too much friction. The interviews disproved this. Users said the form felt like too much because nothing on the page had built up to it. Pricing was buried, the work was unseen, the credentials were footer-level. Without context, a 7-field form reads as a mortgage application. With context, it reads as professional intake.
Mobile usage was not the surprise. The context was. Multiple interview participants described standing at a grave site with their phone open, trying to remember the business name, fumbling with the form, dealing with sun glare and small touch targets. The original layout was built for a quiet desktop session. The actual usage was a windy graveside.
Survey responses ranked trust signals in order of influence. Veteran-owned was the runaway winner, ahead of insurance, LLC status, and Google reviews. On the original site, the credential lived in the footer next to copyright. Moving it to a hero subtitle — two words, no badge, no flag — was the highest-leverage single change in the redesign.
The audience spans tech-savvy adult children and elderly relatives. Standard accessibility passes were necessary but not sufficient. The interviews surfaced specific frustrations — small touch targets, no phone-number alternative, low-contrast secondary text — that an accessibility audit might let slide but a 72-year-old user would feel immediately. Inclusive design here meant going beyond the checklist.
The form wasn't the wound. The lack of context around it was.
The findings consolidated into a three-pillar problem statement that guided every subsequent design decision. The pillars are deliberately framed in emotional terms because that is what the research surfaced — not feature gaps, but emotional friction.
Users are vulnerable. Every extra click is a tax on someone already carrying grief. The original site's information density read as efficient to a healthy user and exhausting to a grieving one. Information density that helps in B2B SaaS hurts in grief care.
Pricing was buried. Real work was absent. Veteran-owned status was a footer detail instead of a thesis. By the time the form appeared, the page hadn't earned the right to ask anything.
73% of users were on phones — many of them standing at the cemetery itself. The original layout was built for a quiet desktop session. The actual context was outdoor, distracted, time-pressured, and often emotional.
The next sections shift from research findings to design rationale. Where Sections 03–06 reported what the research found, Sections 07–08 explain what the design did about it. The researcher and the designer in this project were the same person — but the work is documented separately so each can be evaluated on its own terms.
Each proposed change traces directly to one of the four findings. No design decision was made on intuition alone. The "kept" column is as important as the "changed" column — the form length is a deliberate non-change.
The single biggest insight from the interviews was that buried pricing reads as a trap. The redesign moved real numbers to the surface of the Services section. Trust came from the willingness to name a number.
The obvious move would have been cutting the form down. I didn't. Grave care requires cemetery name, section, plot number, and the deceased's name to actually do the work — there is no shorter version that yields a real inspection. So I made every other part of the site do the trust-building work the form couldn't. Form completion went up 20% on the same seven fields.
The Recent Restorations gallery uses actual work — Bernhardt, Sparks, Warren — named families, real headstones, hover-to-reveal wipe interactions. Stock photos were considered and rejected because a grieving user can tell the difference instantly.
On the original site, the veteran credential lived in the footer next to copyright. The survey ranked it as the single most-influential trust signal in the entire research set. The redesign moved it directly under the hero headline. Two words, no badge, no flag, no graphic — the credential earns more trust when it isn't dressed up.
Sage palette, large type, generous spacing, sentences that finish before the next idea starts. The page is slower than a SaaS marketing site by design. Calm beats efficient when users are grieving. The pacing isn't decoration — it's the whole product.
Measured against the original site, comparing across the first two operating seasons. The redesign launched between Season 1 and Season 2, with continued iteration through the second. All numbers below are real and measured — none are projections.
Mostly one-off services. Up from a handful in Season 1.
After redesign launch, measured at 60 days.
Improvement on the same seven-field form.
Service unification + pricing in the open.
Pricing questions answered before they were asked.
Post-service survey, independent of the booking funnel.
The reason mobile-first wasn't optional.
No reduction. The trust work happened elsewhere.
Building first and researching second gave me a sharp comparison: which design instincts were right, which were defensible-but-suboptimal, and which were dead wrong. The trust-deficit insight was the dead-wrong one. I had treated form length as the problem when the real problem was the context around the form.
The default instinct in UX is to cut fields. Sometimes the right move is to leave the form alone and rebuild the page around it. A 7-field form on a page that has earned trust completes more often than a 4-field form on a page that hasn't.
WCAG passes were necessary but not sufficient. The actual gains came from 44px touch targets, a prominent phone-number alternative, larger body text by default, and slower pacing — design moves an audit would not flag but a 72-year-old user would feel immediately.
SaaS UX is built on speed. Grief UX is built on calm. Every default I had absorbed from product work had to be reversed. Generous spacing instead of dense. Soft confirmations instead of micro-celebrations. Sentences instead of headlines. The pace is the product.
The current site is a working business, not a finished product. These are the next research and design moves I'd prioritize for season three, ordered by expected impact-per-effort.
The interview guide and survey instrument are available below. Raw transcripts and individual survey responses are not published, to protect participant privacy.
The semi-structured question set used across all four interviews, organized around opening rapport, current pain points, pricing perception, mobile context, and trust signals.
The 12-question mixed-format survey distributed via local channels. Includes Likert scales for satisfaction, ranking questions for trust signals, and open-text fields for grievances.
The thematic codes applied to interview transcripts, including primary themes (trust deficit, mobile context, emotional load) and sub-codes.
Documents are anonymized — no participant identifiers, no client-confidential pricing details, no raw response data.
Sheffield Grave Care is now in its second operating season — a working business serving Lorain County, built on a site grounded in real research instead of designer intuition.