Turning an agency site people trusted but didn't act on into one that closes

A site that informed people without ever convincing them

Alan + Co. had a working website. Navigation was fine, services were listed, the contact page existed. But leads weren't coming from it, and nobody internally could say why.

Looking at the pages, the pattern was hard to miss. Every section used the same layout, the same weight of type, and the same wall of paragraph text, so nothing led the eye anywhere. Services were grouped into broad buckets — "Integrated Marketing," "Graphic & Web Design" — that described categories instead of outcomes. Case studies existed but were buried, unlabeled, and thin on results. There was no visual hierarchy and no clear path from "this looks fine" to "I should contact them."

• No hierarchySame section pattern repeated down the page — headline, two paragraphs, stock photo.
• Vague positioningService lists that could belong to any agency in the country.
• Hidden proofCase studies with no measurable outcomes and no obvious entry point.
• No clear next stepContact was a nav item, not a moment the page built toward.

I tested the site we had against what people expect from any agency

I ran two studies side by side. An unmoderated usability test in Maze with 11 participants doing real tasks on the live site, and a UX preference survey with 8 participants about what they look for in an agency website generally. One told me what people could do; the other told me what they wanted.

Task success was high — 100% found services, 91% found contact information. But only 64% found the case studies, the one thing that proves an agency can do the work. And the perception scores dropped hard where the decision actually happens.

"Basic and generic." "Nothing unique." "Cookie cutter." "Too much information."

• 3.9 / 5Navigation clarity — the site was usable.
• 3.7 / 5Overall design — acceptable, not compelling.
• 2.3 / 5Ease of finding information — below average.
• 7.1 / 10Trustworthiness — enough to browse, not enough to contact.

Usability wasn't the problem. Persuasion was.

45% of participants said they'd consider reaching out. After actually reading the site, 55% decided against it. That reversal was the whole finding: people weren't getting lost, they were getting unconvinced.

I framed it as a gap between what the site delivered and what people came looking for, and reduced it to four problems the redesign had to solve. That gave the team something to build against instead of a list of opinions about the homepage.

• HierarchyFunctional navigation, but no visual order telling people what mattered.
• PositioningClean but generic — indistinguishable from competitors.
• ProofText-heavy sections where people expected results and real client context.
• ClarityToo much information, not enough of a story.

Alan + Co. became Running Point Media

"Generic" started with the name. "+ Co." is what every small agency is called, and the research showed people couldn't tell the company apart from its competitors before they'd read a single line. The rename to Running Point Media borrowed a sports phrase the owners already used — the person who runs the point sets up the play — which gave the brand a story local business owners repeat back.

I built the identity around it: the RP mark, a deep ink base with brick red as the primary and brass as the accent, pine green carried over from the logo, and a Bricolage Grotesque and Manrope pairing that gives headlines real weight against quiet body copy. Every token got defined once so the site couldn't drift back toward beige.

Restructured around outcomes, then built as real code

I mapped the new architecture before designing anything: a homepage that states who the agency serves above the fold, services broken into four named practices with outcomes attached, and case studies pulled up into their own labeled section with measurable results. Mockups and user flows covered the paths that mattered — homepage to service to case study to discovery call — including where someone lands from a Google search on a phone.

Instead of handing off static screens, I built the site as a coded prototype in Lovable — React with TanStack Start, brand tokens defined in one stylesheet, scroll-reveal animation and a real contact flow. The team clicked through the actual thing at review speed, which meant feedback was about hierarchy and copy rather than what a rectangle in Figma might feel like. That prototype became the production site.

A site that leads somewhere

The new homepage opens with a type-led statement of who the agency works with, then breaks into four numbered practices, a services ticker, proof-driven case cards, and one obvious next step repeated down the page. The four problems from the research each have a section answering them.

It's live, and because it's code rather than a template, every future case study and service page drops into the same system without redesigning anything.