Repositioning and rebuilding an IT consultancy’s web presence, end to end

Why the site had to change

My employer, a Spanish IT consultancy in its startup years, was competing for attention against firms many times its size. The existing website was a handful of succinct, outdated pages that undersold what the company had grown into, and gave prospects nothing to build trust on. I was asked to fix it. “Fixing it” turned out to mean strategy, information architecture, a design system, every page of content in two languages, the build itself, and the security and performance work that followed launch.

Research: what were the bigger players doing right?

Before opening any design tool, I spent a week benchmarking competitors, from global consultancies down to Spanish mid-size players. Patterns emerged fast: conversion runs on social proof and authority (client logos, big reference numbers, success stories); navy blue plus one accent colour dominates because it signals trust; services are framed as solutions on the homepage, not listed; layouts share block-and-card structures, generous white space, and motion.

Just as useful was the inverse question: what could a small challenger credibly claim that the big names couldn’t? Locality and cultural closeness, direct access to founders, agility, transparency, hunger. That list became the raw material for the brand direction.

Strategy: three archetypes, one hard decision

I distilled the research into three brand archetypes the company could credibly own: a technology-driven AI specialist, a human-centric digital partner, and an agile growth accelerator. Each came anchored to real competitor references, a visual mockup, and the success metric it would optimise for. I brought them to leadership with a structured set of decision questions rather than a recommendation, because the choice was theirs to live with.

The signals came back conflicting. Our CTO leaned toward the AI-specialist direction; the references our COO kept sharing pointed somewhere else, toward competence and prestige. Rather than escalate the disagreement or quietly pick a side, I synthesized: an AI-forward offering wrapped in an authority-driven presentation, using the competitor research as the referee for which elements of each archetype earned their place. That synthesis became the brief I designed against.

From concept to direction

I translated the chosen direction into an information architecture: what earns a place on the homepage, what lives behind the menu, how services become solutions. An early concept mockup explored a light, corporate look. It didn’t survive long. The references leadership responded to kept leaning toward darker themes, and the light corporate territory already belonged to the big legacy consultancies the strategy said not to imitate. A dark, bolder direction felt more current and carried the confidence the research said a small firm needs to project. That became the final look.

A design system, not just a design

To keep a fast-moving site consistent, and usable by people who aren’t designers, I built the visual language as a documented token system: a type scale with real usage examples, a colour system with semantic and state colours across light and dark surfaces, component specs with states, iconography, and breakpoints. It opens with a “how to read this file” section so anyone at the company can apply it, and it’s versioned: v2 shipped with a changelog, because a design system that can’t absorb change is just a style page.

The build: six weeks, two languages, one pair of hands

With the strategy set, I built the site in WordPress/Elementor in a focused six-week push: ten pages, every one of them written from scratch in both English and Spanish, with the backend managed solo after a single walkthrough from our CTO. The bilingual content authoring was the heaviest single load in the project. It doubled the writing and forced every structural decision to work in two languages at once.

Launch didn’t end the content work. Three Spanish-only legal pages were finished a few weeks after going live, once the actual legal materials arrived, and I set up the site’s cookie management to match what the site really tracks and what those legal pages promise. Compliance turned out to be a significant design-and-build task of its own, not a checkbox.

Hardening what shipped

After launch, the work moved below the surface. Two authorized security audits over the following months drove layered hardening: authentication and login protection, an application-layer firewall, security headers, closing user-enumeration vectors, and email authentication, with the second audit verifying that the first round of fixes held. I documented the whole setup as a replication playbook, including a “pitfalls learned the hard way” section, so the security posture survives any single person’s memory, mine included. Performance tuning brought the desktop PageSpeed score to 99/100.

Beyond the website

The brand system didn’t stop at the site. I extended it into a maintenance-service document for clients and sales one-pagers in two deliberate variants, a dark theme matching the site and a light alternative, so the company meets prospects with the same visual language on paper as on screen.

What shipped

A bilingual production site on a new brand direction, live since March 2026: ten pages in English and Spanish plus three Spanish legal pages, with cookie management matched to what they promise. A versioned design-token system in daily use. A collateral suite on the same system. A documented, audit-verified security baseline. Desktop PageSpeed 99/100. Delivered as a team of one, with a six-week core build.

What I’d do differently

Lock the content before the build. Writing two languages of copy while building was the main drag on those six weeks. Next time: content architecture first, then the tool.

Use the people around me earlier. I absorbed most decisions alone and treated asking for input as an interruption to progress. The archetype synthesis worked, but it would have been cheaper as a workshop than as a solo balancing act.

Defend the trade-offs you believe in. The site’s video hero costs mobile Lighthouse points (61, against 99 on desktop) because mobile scoring simulates lower-tier hardware and connection speeds and penalises video backgrounds hard. I would still ship it: the video is a genuine differentiator that elevates the site, and the benchmark’s simulated device is weaker than what most of the site’s actual audience carries. Real-device user testing would settle the question; it was out of scope for this project’s goals, with other projects already queuing, and it’s the first thing I’d add given more room.