Designing and building my own agency's site, end to end
TANGERINE · AGENCY · WEB DESIGN · 2025
Lead designer and Webflow developer. Strategy, visual design, copywriting direction and build. Everything except the copy.
Problem
The site listed services without positioning and showed credentials without confidence. It read as designed to avoid controversy rather than to attract the right clients.
What I did
Audited competitors, interviewed past clients, reviewed analytics, then rebuilt the information architecture around the three services that actually drove qualified inbound, and built all 15+ pages in Webflow.
Result
15+ pages shipped solo. Seven services collapsed into three. After launch, inbound shifted toward better fit clients and away from small one off requests.
The call
Collapsing seven services into three met real internal resistance. The research is what made the decision hold. I could show which services drove qualified leads instead of arguing taste.
CONTEXT
Your own agency is the hardest client of all.
Tangerine's existing website did not communicate the quality of the work behind it. It listed services without positioning, showed credentials without confidence, and felt built to avoid objections rather than attract the right clients.
I led the project end to end: strategy, visual design, copywriting direction, and the Webflow build across all 15+ pages. Everything except the copy itself was mine.
PROBLEM
The site did not communicate confidence or focus.
The agency offered seven loosely defined services across too many industries. To a client evaluating agencies side by side, that reads as unfocused. The visual system was inconsistent, hierarchy was not directing attention, and there was no clear story of who Tangerine was for.
Being an internal project does not make stakeholder management easier. It makes it harder. Everyone has an opinion, nobody has a decision making mandate, and the feedback loops never close.
So the first problem to solve was not the homepage. It was how to make a decision stick.
FINDINGS
Research made the case that opinions could not.
Analytics were bluntQualified inbound came from three services: Product, Brand and Growth. The other four generated volume, not fit. That was not an opinion, it was the log.
Clients bought peoplePast client interviews kept returning to the team rather than the portfolio. Agency buyers are hiring humans, and the old site barely showed any.
Founders decide in secondsAlex, a startup founder, judges fit almost immediately. If the first screen does not say who this is for, there is no second screen.
Evaluators need proofPriya, a marketing lead, is looking for real work, real roles and real outcomes. Credentials without a story do not survive a shortlist.
DESIGN PROCESS
Positioning first, pixels second.
I opened with research rather than moodboards, because the problem was positioning and not aesthetics. I audited competitor agencies, interviewed past clients about why they chose Tangerine, and reviewed analytics to find which services actually drove qualified inbound.
With three services established, I rebuilt the information architecture around them, designed the system in Figma, and developed everything in Webflow, iterating page by page with the team.
The homepage now leads with who Tangerine is for and the three things it does best, instead of a catalogue of everything it could conceivably do.
OUTCOME
Fewer services, better clients.
The site shipped as 15+ pages designed and built solo, on one component system, with a CMS the team can run without a designer on call.
After launch, inbound shifted toward better fit clients and away from small one off requests. The measurable win was not traffic. It was the shape of the conversations that started.
Internal projects need a single ownerWithout one person holding actual decision making authority, every design decision becomes a negotiation with no deadline. I would establish a project owner before a single pixel moves.
Focus is a design decisionThe biggest impact did not come from visuals. It came from information architecture and positioning, which is the part of the work that looks like nothing in a portfolio screenshot.
Bring data to a taste fightArguing taste with colleagues goes nowhere. Arguing with the analytics ends the conversation in one meeting.
NEXT CASE
Good credentials, buried by bad structure WEBITMD · 2023 →