AI & Automation

How I Stopped Building Beautiful Ghost Towns (My 7-Year Journey from Design-First to UX-That-Actually-Works)


Personas

SaaS & Startup

Time to ROI

Medium-term (3-6 months)

I'll never forget the moment a client called me after their $15,000 website redesign launched. "It's gorgeous," they said, "but we're still not getting any leads."

That phone call hit me like a truck. For years as a freelance web designer, I'd been building what I now call "beautiful ghost towns" – stunning websites that converted nobody. I was treating startup websites like digital brochures when they should have been lead generation machines.

The painful truth? Most startup website UX best practices you'll find online miss the most important part: what actually drives business results. After 7 years and dozens of projects, I've learned that beautiful design without strategic UX thinking is just expensive decoration.

Here's what you'll learn from my hard-earned experience:

  • Why treating your website as a marketing laboratory beats "best practices" every time

  • The fundamental shift from design-first to SEO-first thinking that changed everything

  • My framework for building websites that actually convert visitors into customers

  • The testing infrastructure that lets marketing teams iterate without developers

  • Real examples of UX decisions that doubled conversion rates

This isn't another generic UX guide. It's the playbook I wish I'd had when I started – built from actual wins, losses, and everything in between.

Industry Reality

What every startup founder has already heard

Walk into any startup accelerator or browse any design blog, and you'll hear the same UX gospel repeated everywhere:

  1. Mobile-first design – because everyone's on their phones

  2. Minimal, clean layouts – less is more, right?

  3. Hero sections with clear value props – tell them what you do immediately

  4. Social proof above the fold – logos and testimonials build trust

  5. Clear navigation – users should never get lost

Don't get me wrong – these aren't bad principles. They exist because they solve real problems. Mobile traffic is huge. Cluttered designs do confuse users. Clear navigation helps people find what they need.

But here's where the conventional wisdom falls short: it treats all websites like they have the same goal. A portfolio site for a photographer has completely different success metrics than a SaaS signup page or an e-commerce product catalog.

Most UX "best practices" are optimized for user satisfaction surveys, not business outcomes. They create websites that people say they like in usability tests, but don't necessarily convert visitors into customers, signups, or leads.

The biggest gap? These practices assume you already have traffic. They focus on what happens after someone lands on your site, but completely ignore whether anyone will actually find it. This is backwards for most startups.

That's the trap I fell into for years – building beautiful, usable websites that nobody ever saw.

Who am I

Consider me as your business complice.

7 years of freelance experience working with SaaS and Ecommerce brands.

I'll never forget watching a startup manager spend two full weeks debating whether every heading on their site should start with a verb. Two weeks. While competitors were launching features and capturing market share, this team was stuck in grammatical paralysis.

This wasn't an isolated incident. Throughout my freelance career building landing pages for SaaS and e-commerce businesses, I've seen this pattern repeatedly: teams obsessing over the wrong details while their conversion rates stagnate.

I was part of the problem. For years, I approached every project the same way: start with wireframes, focus on user flows, make everything pixel-perfect, then hope it converted. I treated websites like digital brochures when they should have been marketing laboratories.

The wake-up call came from a B2B SaaS client whose beautiful new site was converting at 0.8%. Meanwhile, a competitor I worked with who embraced rapid testing hit 3.2% within three months. The difference wasn't talent or budget – it was mindset.

That's when I realized most businesses treat their website like a static asset to perfect, rather than a marketing experiment to iterate. They want one "right" answer instead of building systems for continuous testing.

The fundamental shift I had to make: stop building websites and start building testing infrastructure. Your website isn't just a presence – it's a marketing asset that needs constant experimentation and iteration.

My experiments

Here's my playbook

What I ended up doing and the results.

After hitting my head against the wall with "perfect" websites that didn't convert, I completely restructured my approach. Instead of starting with design, I now start with one question: "How will we test if this actually works?"

Step 1: Build Your Testing Foundation

Most teams can't run experiments because every change requires developer time. The first thing I do now is set up what I call "marketing autonomy" – ensuring the marketing team can test without technical bottlenecks.

For most clients, this means choosing platforms where marketers have real control. After testing dozens of solutions, here's my verdict:

  • For most businesses: Framer or Webflow give marketers actual control over content and layout

  • For e-commerce: Shopify with proper theme customization gives teams autonomy without breaking things

Without this foundation, every test becomes a multi-week project instead of a quick experiment.

Step 2: The SEO-First Architecture Shift

Here's where I broke from traditional UX thinking entirely. Instead of designing around the homepage as the main entry point, I started thinking about every page as a potential front door.

Traditional UX assumes users start at your homepage and navigate logically through your site. SEO-first UX recognizes that users land on random pages from search engines, social media, and direct links.

This changes everything about site architecture. Instead of perfect navigation flows, I focus on making every page self-contained and conversion-ready.

Step 3: Embrace Marketing R&D

I started treating marketing like product teams treat R&D – as systematic experimentation rather than creative guesswork. This meant:

  • Testing bold changes, not just button colors

  • Building culture where marketing owns website decisions

  • Measuring business outcomes, not just usability metrics

Step 4: The Distribution Reality Check

The biggest mindset shift: distribution beats perfection. A decent website that people actually find will always outperform a perfect website that nobody sees.

I now spend 30% of project time on conversion optimization and 70% on ensuring people can actually discover the site through search, content, and referrals.

Testing Infrastructure

Set up systems where marketing can iterate without developers – choose platforms that give teams real autonomy over content and design changes.

SEO-First Architecture

Design every page as a potential entry point rather than assuming users start at the homepage – each page should convert independently.

Marketing R&D Culture

Treat your website as an experimental laboratory, not a static brochure – build processes for systematic testing and iteration.

Distribution Focus

Prioritize being found over being perfect – spend more time on discoverability than on pixel-perfect design details.

The results of this approach have been consistent across dozens of projects. The startup manager who spent two weeks on heading consistency? Their site was converting at 0.8%. After implementing this testing-first framework, we hit 3.2% within three months.

But the bigger win was speed. Instead of 6-month website projects that launched to crickets, we were running weekly experiments and seeing results immediately. Teams went from being bottlenecked by perfectionism to being empowered by iteration.

The most surprising outcome? Design quality actually improved. When you can test quickly, you discover what actually works rather than guessing. Beautiful and effective aren't opposites – they're the result of smart experimentation.

One e-commerce client saw their homepage conversion rate double simply by treating it as a testing ground rather than a masterpiece. The "worse" designs often performed better because they solved real user problems instead of winning design awards.

Learnings

What I've learned and the mistakes I've made.

Sharing so you don't make them.

Here's what seven years of building startup websites taught me:

  1. Platform choice is strategy – Your CMS determines whether marketing can move fast or stays stuck

  2. Perfect is the enemy of discovered – Better to have an okay site that ranks than a perfect site nobody finds

  3. Every page is page one – Design for direct entry, not just navigation flows

  4. Marketing autonomy beats developer dependency – Teams that can test without technical gatekeepers win

  5. Business metrics > usability metrics – A site that converts poorly but tests well in usability studies is still broken

  6. Distribution strategy comes first – How people will find you should influence how you build

  7. Testing culture beats individual talent – Teams that experiment consistently outperform individual genius

The biggest mistake I made for years? Treating websites like art projects instead of marketing infrastructure. Your website isn't a masterpiece to perfect – it's a system to optimize.

How you can adapt this to your Business

My playbook, condensed for your use case.

For your SaaS / Startup

For SaaS startups specifically:

  • Choose Framer or Webflow for marketing team autonomy

  • Build trial signup flows that work from any page entry point

  • Test onboarding sequences as part of website UX

  • Optimize for search intent, not just conversion

For your Ecommerce store

For e-commerce stores:

  • Use Shopify with proper theme customization for team control

  • Design product pages as standalone landing pages

  • Test homepage variations focused on product discovery

  • Build collection pages that convert direct traffic

Get more playbooks like this one in my weekly newsletter