[ Case study ]
The firm's practice-area pages sat inside a slow legacy site; the intake form asked for a life story before any trust was built, and analytics showed most visitors leaving before the form's second section.
CLIENT a family-law practice — FOCUS Earn the form progressively
Representative examplesEvery case study in this library is an illustrative composite of the kind of engagement we deliver — written to show our method and standards, not to name clients.
Family-law clients arrive frightened, usually at night, usually on a phone — and this four-attorney practice has served one metro area long enough to know it. Consultations are genuinely scarce; the partners take a fixed number each week, and the intake coordinator screens every request by hand. The firm's practice-area pages sat inside a legacy site a vendor has maintained since it was built, describing the firm well and taking their time about it. The intake form beneath those pages asked for a life story before it asked for a name.
The firm's practice-area pages sat inside a slow legacy site; the intake form asked for a life story before any trust was built, and analytics showed most visitors leaving before the form's second section.
Earn the form progressively: the intake splits into a three-question first step anyone can answer, with the full history requested only after the visitor has already responded once — commitment follows contact. The practice page itself is hand-coded against the firm's identity at a fraction of the weight, with the bar-required disclaimers kept verbatim and untouchable. Form-step events are instrumented so future changes argue from where anxious visitors actually leave, not from anyone's guess about which question scares people off.
Just as important is what we ruled out, and why:
The intake split into a three-question first step with the full history requested after the first response — commitment follows contact.
The practice page was hand-coded against the firm's identity at a fraction of the weight, with the disclaimer and bar-required language kept verbatim.
Form-step events were instrumented so future changes argue from data about where anxious visitors actually leave.
Delivered by the experience pod — frontend engineer over 3 weeks, with working increments reviewed with the client every week.
Obstacle
The required disclaimer wording differed between the firm's handbook and its website footer, and neither source could say which version was current.
Handled: We paused the launch, asked the managing partner to confirm the governing version in writing, and shipped against that text with the discrepancy flagged for cleanup.
Obstacle
Launch-week step data showed the form's second step carried a question the coordinator never uses, and visitors experienced it as required.
Handled: Her confirmation plus the drop-off numbers got it removed within days; the form now asks nothing the screening call will not actually use.
Obstacle
The legacy vendor controlled the DNS, and coordinating the page's routing consumed more calendar time than building the page did.
Handled: We wrote the exact routing instructions their tech needed and joined one call; the change took twenty minutes once someone finally asked precisely.
The headline: p75 lcp on practice-area pages, 28 days post-launch — 1.8s → 0.9s, read from CrUX field data. A second check: completed consult requests versus the prior quarter at +37%.
The form started asking the way the coordinator screens: gently, in order, one thing at a time. Completed requests arrive with enough context that her first call is a conversation instead of an interrogation, and the ambivalent visitors who used to vanish at the second section now stay through it. The partners noticed consult quality held while volume rose — which was the entire point of qualifying before booking. And the firm stopped treating its website as a brochure it pays someone to update.
The result was read from CrUX field data against the pre-engagement baseline over the stated window, with a guardrail check on completed consult requests versus the prior quarter. Where platform-reported numbers and business outcomes differ, this record says which layer it is quoting.
What we would do differently
We would have written the second-step questions with the firm's intake coordinator — one question she never uses was removed after launch week.
[ Related service ]
[ Related builds ]
2.9% 7.8%Replay-to-enrollment conversion across the next two cohorts versus the two prior
4.1% 9.3%Visit-to-demo-request conversion across all ad groups, 8 weeks after launch versus 8 weeks before
[ Next step ]
Next case study