Booking.com

Scaling copy experimentation into a repeatable system

Project Overview

Role: UX Writer / Experimentation Lead

Tools: Documentation, experiment planning templates, Figma, analytics readouts, workflow trackers

Collaborated with: UX Writers, Product Managers, Designers, Engineers, Data Analysts, and leadership

Industry: Travel Tech

Problem / Goals:

Copy experiments had potential, but the system around them was not strong enough.

Some surfaces were difficult to test cleanly. Measurement setups were inconsistent across platforms. Leadership was not always convinced that copy alone could drive meaningful impact. At the same time, many writers wanted to experiment but did not yet have the confidence, templates, or support to do it independently.

The goal was to make copy experimentation easier, more reliable, and more visible so content decisions could influence product outcomes at scale.

How I helped:

Proved the value of pure copy experiments through early high-impact tests.
Framed copy experimentation as a low-cost way to de-risk product decisions.
Partnered with Engineering to improve the technical foundations for cleaner testing.
Created reusable standards, templates, and guidance for experiment planning.
Coached UX Writers on hypotheses, variants, QA, launch readiness, and storytelling.
Increased visibility of experiment learnings through leadership updates, forums, and shared documentation.

Project Requirements

The system needed to:

Allow copy changes to be tested independently where possible.
Connect experiments to product goals and customer problems.
Improve confidence in results across web and mobile surfaces.
Reduce setup friction for writers and product teams.
Make learnings reusable beyond a single test or team.

What I Did

Proved copy could move product outcomes

I started with the testable surfaces we already had and ran focused copy experiments with clear hypotheses, clean variants, and measurable outcomes.

These early wins helped change the conversation. Copy experimentation was no longer an abstract craft preference. It became a practical way to learn faster, reduce risk, and improve product performance.

Reframed experimentation for leadership

I presented copy experimentation as more than a way to optimise words. I framed it as a decision system:

A way to test customer understanding.
A way to validate product assumptions.
A way to improve commercial outcomes without major build effort.
A way to create reusable learnings for future work.

That framing helped secure leadership support for experimentation velocity and for the technical cleanup needed to make tests more reliable.

Strengthened the technical foundations

I partnered with engineers across platforms to identify where copy experiments were difficult to isolate or measure.

Together, we broke the work into smaller, trackable improvements so teams could gradually make more surfaces testable without treating the infrastructure work as one giant project.

The result was a stronger foundation for cleaner copy tests across web and mobile experiences.

Built standards and playbooks

Once the foundations were improving, I focused on consistency.

I created practical guidance for:

Writing stronger hypotheses.
Choosing useful success metrics.
Creating variants that tested one clear idea at a time.
QA'ing experiments across platforms.
Reading results without overclaiming.
Turning results into reusable content principles.

The goal was not to make every writer follow a rigid process. It was to give them enough structure to move faster with more confidence.

Mentored writers and scaled the skill

I coached writers across multiple product areas on how to turn product questions into experiments.

That included leading workshops geared towards hands-on support for experiment setup, hypothesis framing, stakeholder alignment, and result storytelling.

I also helped run focused experimentation sessions where writers could leave with a stronger hypothesis, a clearer plan, or a test ready to move forward.

Results

Helped make copy experimentation a more trusted and repeatable capability.
Improved the reliability of copy testing across web and mobile surfaces.
Increased leadership visibility of content impact and experiment learnings.
Enabled more writers to design, launch, and interpret experiments independently.
Shifted the perception of UX Writing from "wording support" to a strategic partner in product learning and decision quality.
Created a foundation for a broader experimentation program across the content design craft.

What I learned

Scaling craft impact requires more than doing excellent individual work.

The bigger unlock is building the conditions that let other people do excellent work too: clearer standards, stronger systems, better evidence, and enough psychological safety for people to try, learn, and improve.

What I Would Do Differently Now

I would build the learning library earlier. I created visibility through updates and forums, but a searchable repository of experiment hypotheses, variants, outcomes, and principles would have made the learnings easier to reuse across teams.