As an operational researcher working in performance marketing, I look at a website as an instrument. It needs to be fast, measurable, and predictable. Our previous setup made that harder than it should be, so we rebuilt it.
What we wanted to fix
The old stack added variance we could not explain:
- Performance noise: slow or inconsistent load times made it harder to trust landing page results.
- Measurement friction: tagging and data-layer work was brittle and slow to ship.
- Content overhead: small updates took too long and relied on too many plugins.
Our working hypotheses
We treated the rebuild as a controlled change, not a redesign for its own sake.
- Lower technical variance improves decision quality. If the platform is stable, test results are easier to interpret.
- Simpler delivery reduces lead time for experiments and content updates.
- Clearer structure improves crawlability and user comprehension.
Requirements we set upfront
- Static-first output with minimal runtime JavaScript
- Clear content model for pages and posts
- Easy metadata control (titles, descriptions, OG tags)
- Straightforward deployment and rollback
- Separation between site code and analytics configuration
The stack we chose
- Astro for static rendering and fast pages
- Tailwind for consistent, low-friction styling
- Netlify for deployment and previews
- GTM for analytics and consent tooling
This keeps the site simple, fast, and easy to maintain, while allowing the measurement layer to evolve independently.
What this enables
- Faster iteration on copy and landing page structure
- Fewer variables in performance analysis
- A site that is easier to audit, explain, and improve
If your current site introduces noise into your marketing data, we can help you isolate the constraints and fix them. Get in touch for a rapid evidence review of your tracking and your accounts.