SEARCH & REFINE

How I unified 2.3 million daily searches into a scalable platform without breaking revenue.

Strategic Leadership

Platform Thinking

Systems Design

Business Impact

TL:DR

Trade Me was migrating millions of users from its legacy “Classic” platform to a modern responsive architecture. While most areas aimed to preserve familiarity, search (powering ~2.3 million daily queries) had become fragmented after 20 years of independent evolution across Marketplace, Property, Motors, and Jobs.

 

Each business unit initially planned to redesign search independently. I intervened and reframed search as a shared platform capability rather than four experiences, redirecting the rebuild toward a unified system. I defined the refinement model, interaction architecture, and UX framework adopted across web and mobile.

Results

~2.3 million daily searches migrated without metric degradation

Mobile delivered +3% statistically significant CTR uplift

Property added 10 new refinements over 4 years with zero front-end build required.

Core components (chips, sliders) were embedded into the design system and reused across the app ecosystem.

The context

Landscape

Search sat inside four business units, each with its own product leadership and local optimisation priorities. There was no existing platform governance model for shared capabilities like search. I operated purely through influence, not mandate.

  • Four business units operating semi-independently
  • Legacy platform (“Classic”) without reusable component architecture
  • Executive pressure to migrate traffic quickly
  • No centralised search governance model
  • No formal authority held by design at platform level

The Problem

Search is was the primary discovery engine driving listing visibility, engagement, and downstream revenue. Any degradation during migration would have had immediate commercial impact. Success was defined not by novelty, but by protecting performance while building for scale. The problem was that trade me classic as a platform was no longer scaleable. It was built on 1990s and early 2000s technology but more importantly the UX was just as dated. there was the risk of UX fragmentation through an independent BU redesign

Constraints

This was not a greenfield rebuild. The work needed to modernise the experience while respecting current API with and timeline limitations.

  • Minimal API changes permitted
  • Tight migration timelines
  • Mobile rebuild funded primarily by Marketplace
  • Vertical-specific bespoke requirements

The process

1

Framing the Strategic Intent

Search was the commercial backbone of Trade Me, powering ~2.3 million daily queries and directly influencing listing visibility, engagement, and revenue flow. After two decades of incremental change, it had become structurally fragmented across business units. The responsive rebuild created a strategic fork in the road: either allow further divergence or unify search as a scalable platform capability. I reframed search not as a business-unit feature, but as the connective tissue of the platform, a shared capability that required architectural integrity rather than local optimisation.

2

Understanding User and System Complexity

Search needed to support radically different user mindsets. From Impulse‑driven marketplace shopping to well considered vehicle comparison and High‑stakes property and job decisions. To active these different mindsets the underlying system had accrued hundreds of refinements across thousands of categories over two decades. Rather than treating this as irreducible complexity, I focused on pattern discovery. Through auditing existing refinements across all businesses, I identified Seven core refinement types, plus a small number of truly bespoke cases. This insight was foundational. It allowed the problem to be simplified into a system that could be designed, built, and communicated clearly across teams.

3

Exploring and Validating Concepts

With a reduced pattern set, I explored multiple conceptual models for how refinements could be surfaced and interacted with across contexts. To ensure the concepts worked for our users I conducted competitor analysis across e‑commerce, motors, property, and jobs platforms. User tested several end‑to‑end search concepts through in‑person usability interviews to test mental models.

A critical finding in the user testing and reinforced by behavioural data was the enduring importance of Category and Location, dating back to Trade Me’s original 1999 implementation. Users expected these to be immediately accessible. This insight led to a key principle to surface the most relevant refinements by context, rather than exposing everything equally.

Tension

I prioritised cross-platform consistency over vertical-specific optimisation in select cases. There was some tension across the Property Team where I declined a business unit request to remove the universal title pattern. I understood this worked poorly for this business unit (only ~15% of property users relied on keyword search). There were tight timelines and I prioritised changes that had a impacts across all business units. Any individual business unit changes would need to go through the BU’s delivery process.

4

Making the System Explicit

One of the biggest risks in a cross‑business initiative was misalignment. Different teams interpreting the solution differently over time.

To address this, I created a UX framework that made the system explicit. It defined:

  • Refinement types and their behaviours
  • Presentation rules across devices
  • Interaction patterns, edge cases, and fallback states

This framework became the shared language between design, product, and engineering. It allowed teams to debate decisions constructively and ensured consistency as the system evolved.

Search & refine UX framework

I created this framework that allowed thousands of categories adapt to a series of basic rules for user predictably and ease of delivery.

View framework

5

Designing for Scale and Longevity

I worked closely with engineering to align the design with a scalable technical model. The goal was to enable refinments to be added to the API and the without any front end delivery requirments. Most of the business rules to be handled by the API. Presentation rules embedded in the front-end

  • New refinements could be added without additional front‑end design or build work

This shifted search from a brittle UI implementation to a durable platform capability that could grow with the business.

6

Extending the System to Mobile Apps

Following the success of the web implementation, I advocated to extend the framework into Trade Me’s mobile apps. This was the most significant change to mobile search in over a decade. Rather than porting web patterns directly, we took a native‑first approach, fully leveraging touch interactions:

  • New refinement chip behaviours
  • Horizontal scrolling patterns
  • Bottom sheets for progressive disclosure

To further reduce risk, the rollout was staged over eight months, allowing gradual user adaptation and continuous learning. This Delivered +3% statistically significant CTR uplift in the top 50 results.

Impact

Commercial impact

~2.3 million daily searches migrated without metric degradation

No drop in member search activity

Mobile delivered +3% statistically significant CTR uplift

No drop in conversion during migration

Slider experiments showed up to 40% uplift in price refinement usage

Product & Platform Impact

Unified refinement model across 4 business units

7 core patterns supporting thousands of refinements

Reusable architecture embedded in front-end which enabled 10+ new refinements added without a front end rebuild

Search recognised internally as cross-vertical “glue”

Organisational Capability Impact

Created a documented UI framework adopted across teams

Reduced future duplication risk

What This Demonstrates

Strategic Reframing

Redirected a fragmented multi-BU redesign into a unified platform capability without formal authority.

Strategic Leadership

Commercial Risk Management

Protected core performance metrics during high-risk migration of 2.3M daily searches.

Business Impact

Design System Contribution

Defined reusable components adopted across the broader app ecosystem.

Organisational Influence

Systems Abstraction

Reduced thousands of refinements into 7 scalable interaction models.

Systems Design

Platform Thinking

Cross-Platform Architecture

Aligned web and native interaction logic while respecting device-specific patterns.

Platform Thinking

Reflection

If I repeated this today

With hindsight, there are structural improvements I would drive earlier to increase long-term leverage and measurement rigour. While the system proved durable, I see opportunities to strengthen platform alignment and commercial attribution.

If repeated today, I would:

  • Push earlier for deeper backend/API alignment to further reduce legacy constraints
  • Formalise governance mechanisms to prevent long-term system drift
  • Tie CTR uplift more directly to downstream revenue and commercial metrics

Gallery

iOS

Android

Desktop

Mobile

SEARCH & REFINE

How I unified 2.3 million daily searches into a scalable platform without breaking revenue.

Strategic Leadership

Platform Thinking

Systems Design

Business Impact

TL:DR

Trade Me was migrating millions of users from its legacy “Classic” platform to a modern responsive architecture. While most areas aimed to preserve familiarity, search (powering ~2.3 million daily queries) had become fragmented after 20 years of independent evolution across Marketplace, Property, Motors, and Jobs.

 

Each business unit initially planned to redesign search independently. I intervened and reframed search as a shared platform capability rather than four experiences, redirecting the rebuild toward a unified system. I defined the refinement model, interaction architecture, and UX framework adopted across web and mobile.

Results

~2.3 million daily searches migrated without metric degradation

Mobile delivered +3% statistically significant CTR uplift

Property added 10 new refinements over 4 years with zero front-end build required.

Core components (chips, sliders) were embedded into the design system and reused across the app ecosystem.

The context

Landscape

Search sat inside four business units, each with its own product leadership and local optimisation priorities. There was no existing platform governance model for shared capabilities like search. I operated purely through influence, not mandate.

  • Four business units operating semi-independently
  • Legacy platform (“Classic”) without reusable component architecture
  • Executive pressure to migrate traffic quickly
  • No centralised search governance model
  • No formal authority held by design at platform level

The Problem

Search is was the primary discovery engine driving listing visibility, engagement, and downstream revenue. Any degradation during migration would have had immediate commercial impact. Success was defined not by novelty, but by protecting performance while building for scale. The problem was that trade me classic as a platform was no longer scaleable. It was built on 1990s and early 2000s technology but more importantly the UX was just as dated. there was the risk of UX fragmentation through an independent BU redesign

Constraints

This was not a greenfield rebuild. The work needed to modernise the experience while respecting current API with and timeline limitations.

  • Minimal API changes permitted
  • Tight migration timelines
  • Mobile rebuild funded primarily by Marketplace
  • Vertical-specific bespoke requirements

The process

1

Framing the Strategic Intent

Search was the commercial backbone of Trade Me, powering ~2.3 million daily queries and directly influencing listing visibility, engagement, and revenue flow. After two decades of incremental change, it had become structurally fragmented across business units. The responsive rebuild created a strategic fork in the road: either allow further divergence or unify search as a scalable platform capability. I reframed search not as a business-unit feature, but as the connective tissue of the platform, a shared capability that required architectural integrity rather than local optimisation.

2

Understanding User and System Complexity

Search needed to support radically different user mindsets. From Impulse‑driven marketplace shopping to well considered vehicle comparison and High‑stakes property and job decisions. To active these different mindsets the underlying system had accrued hundreds of refinements across thousands of categories over two decades. Rather than treating this as irreducible complexity, I focused on pattern discovery. Through auditing existing refinements across all businesses, I identified Seven core refinement types, plus a small number of truly bespoke cases. This insight was foundational. It allowed the problem to be simplified into a system that could be designed, built, and communicated clearly across teams.

3

Exploring and Validating Concepts

With a reduced pattern set, I explored multiple conceptual models for how refinements could be surfaced and interacted with across contexts. To ensure the concepts worked for our users I conducted competitor analysis across e‑commerce, motors, property, and jobs platforms. User tested several end‑to‑end search concepts through in‑person usability interviews to test mental models.

A critical finding in the user testing and reinforced by behavioural data was the enduring importance of Category and Location, dating back to Trade Me’s original 1999 implementation. Users expected these to be immediately accessible. This insight led to a key principle to surface the most relevant refinements by context, rather than exposing everything equally.

Tension

I prioritised cross-platform consistency over vertical-specific optimisation in select cases. There was some tension across the Property Team where I declined a business unit request to remove the universal title pattern. I understood this worked poorly for this business unit (only ~15% of property users relied on keyword search). There were tight timelines and I prioritised changes that had a impacts across all business units. Any individual business unit changes would need to go through the BU’s delivery process.

4

Making the System Explicit

One of the biggest risks in a cross‑business initiative was misalignment. Different teams interpreting the solution differently over time.

To address this, I created a UX framework that made the system explicit. It defined:

  • Refinement types and their behaviours
  • Presentation rules across devices
  • Interaction patterns, edge cases, and fallback states

This framework became the shared language between design, product, and engineering. It allowed teams to debate decisions constructively and ensured consistency as the system evolved.

Search & refine UX framework

I created this framework that allowed thousands of categories adapt to a series of basic rules for user predictably and ease of delivery.

View framework

5

Designing for Scale and Longevity

I worked closely with engineering to align the design with a scalable technical model. The goal was to enable refinements to be added to the API and the without any front end delivery requirements. To achieve this the business rules are handled by the API. Presentation rules embedded in the front-end. This shifted search from a brittle UI implementation to a durable platform capability that could grow with the business.

6

Extending the System to Mobile Apps

Following the success of the web implementation, I advocated to extend the framework into Trade Me’s mobile apps. This was the most significant change to mobile search in over a decade. Rather than porting web patterns directly, we took a native‑first approach, fully leveraging touch interactions:

  • New refinement chip behaviours
  • Horizontal scrolling patterns
  • Bottom sheets for progressive disclosure

To further reduce risk, the rollout was staged over eight months, allowing gradual user adaptation and continuous learning. This Delivered +3% statistically significant CTR uplift in the top 50 results.

Impact

Commercial impact

~2.3 million daily searches migrated without metric degradation

No drop in member search activity

Mobile delivered +3% statistically significant CTR uplift

No drop in conversion during migration

Slider experiments showed up to 40% uplift in price refinement usage

Product & Platform Impact

Unified refinement model across 4 business units

7 core patterns supporting thousands of refinements

Reusable architecture embedded in front-end which enabled 10+ new refinements added without a front end rebuild

Search recognised internally as cross-vertical “glue”

Organisational Capability Impact

Created a documented UI framework adopted across teams

Reduced future duplication risk

What This Demonstrates

Strategic Reframing

Redirected a fragmented multi-BU redesign into a unified platform capability without formal authority.

Strategic Leadership

Commercial Risk Management

Protected core performance metrics during high-risk migration of 2.3M daily searches.

Business Impact

Design System Contribution

Defined reusable components adopted across the broader app ecosystem.

Organisational Influence

Systems Abstraction

Reduced thousands of refinements into 7 scalable interaction models.

Systems Design

Platform Thinking

Cross-Platform Architecture

Aligned web and native interaction logic while respecting device-specific patterns.

Platform Thinking

Reflection

If I repeated this today

With hindsight, there are structural improvements I would drive earlier to increase long-term leverage and measurement rigour. While the system proved durable, I see opportunities to strengthen platform alignment and commercial attribution.

If repeated today, I would:

  • Push earlier for deeper backend/API alignment to further reduce legacy constraints
  • Formalise governance mechanisms to prevent long-term system drift
  • Tie CTR uplift more directly to downstream revenue and commercial metrics

Gallery

iOS

Android

Desktop web

Mobile web

SEARCH & REFINE

How I unified 2.3 million daily searches into a scalable platform without breaking revenue.

Strategic Leadership

Platform Thinking

Systems Design

Business Impact

TL:DR

Trade Me was migrating millions of users from its legacy “Classic” platform to a modern responsive architecture. While most areas aimed to preserve familiarity, search (powering ~2.3 million daily queries) had become fragmented after 20 years of independent evolution across Marketplace, Property, Motors, and Jobs.

 

Each business unit initially planned to redesign search independently. I intervened and reframed search as a shared platform capability rather than four experiences, redirecting the rebuild toward a unified system. I defined the refinement model, interaction architecture, and UX framework adopted across web and mobile.

Results

~2.3 million daily searches migrated without metric degradation

Mobile delivered +3% statistically significant CTR uplift

Property added 10 new refinements over 4 years with zero front-end build required.

Core components (chips, sliders) were embedded into the design system and reused across the app ecosystem.

The context

Landscape

Search sat inside four business units, each with its own product leadership and local optimisation priorities. There was no existing platform governance model for shared capabilities like search. I operated purely through influence, not mandate.

  • Four business units operating semi-independently
  • Legacy platform (“Classic”) without reusable component architecture
  • Executive pressure to migrate traffic quickly
  • No centralised search governance model
  • No formal authority held by design at platform level

The Problem

Search is was the primary discovery engine driving listing visibility, engagement, and downstream revenue. Any degradation during migration would have had immediate commercial impact. Success was defined not by novelty, but by protecting performance while building for scale. The problem was that trade me classic as a platform was no longer scaleable. It was built on 1990s and early 2000s technology but more importantly the UX was just as dated. there was the risk of UX fragmentation through an independent BU redesign

Constraints

This was not a greenfield rebuild. The work needed to modernise the experience while respecting current API with and timeline limitations.

  • Minimal API changes permitted
  • Tight migration timelines
  • Mobile rebuild funded primarily by Marketplace
  • Vertical-specific bespoke requirements

The process

1

Framing the Strategic Intent

Search was the commercial backbone of Trade Me, powering ~2.3 million daily queries and directly influencing listing visibility, engagement, and revenue flow. After two decades of incremental change, it had become structurally fragmented across business units. The responsive rebuild created a strategic fork in the road: either allow further divergence or unify search as a scalable platform capability. I reframed search not as a business-unit feature, but as the connective tissue of the platform, a shared capability that required architectural integrity rather than local optimisation.

2

Understanding User and System Complexity

Search needed to support radically different user mindsets. From Impulse‑driven marketplace shopping to well considered vehicle comparison and high‑stakes property and job decisions. To active these different mindsets the underlying system had accrued hundreds of refinements across thousands of categories over two decades. Rather than treating this as irreducible complexity, I focused on pattern discovery. Through auditing existing refinements across all businesses, I identified Seven core refinement types, plus a small number of truly bespoke cases. This insight was foundational. It allowed the problem to be simplified into a system that could be designed, built, and communicated clearly across teams.

3

Exploring and Validating Concepts

With a reduced pattern set, I explored multiple conceptual models for how refinements could be surfaced and interacted with across contexts. To ensure the concepts worked for our users I conducted competitor analysis across e‑commerce, motors, property, and jobs platforms. User tested several end‑to‑end search concepts through in‑person usability interviews to test mental models.

A critical finding in the user testing and reinforced by behavioural data was the enduring importance of Category and Location, dating back to Trade Me’s original 1999 implementation. Users expected these to be immediately accessible. This insight led to a key principle to surface the most relevant refinements by context, rather than exposing everything equally.

Tension

I prioritised cross-platform consistency over vertical-specific optimisation in select cases. There was some tension across the Property Team where I declined a business unit request to remove the universal title pattern. I understood this worked poorly for this business unit (only ~15% of property users relied on keyword search). There were tight timelines and I prioritised changes that had a impacts across all business units. Any individual business unit changes would need to go through the BU’s delivery process.

4

Making the System Explicit

One of the biggest risks in a cross‑business initiative was misalignment. Different teams interpreting the solution differently over time.

To address this, I created a UX framework that made the system explicit. It defined:

  • Refinement types and their behaviours
  • Presentation rules across devices
  • Interaction patterns, edge cases, and fallback states

This framework became the shared language between design, product, and engineering. It allowed teams to debate decisions constructively and ensured consistency as the system evolved.

Search & refine UX framework

I created this framework that allowed thousands of categories adapt to a series of basic rules for user predictably and ease of delivery.

View framework

5

Designing for Scale and Longevity

I worked closely with engineering to align the design with a scalable technical model. The goal was to enable refinements to be added to the API and the without any front end delivery requirements. To achieve this the business rules are handled by the API. Presentation rules embedded in the front-end. This shifted search from a brittle UI implementation to a durable platform capability that could grow with the business.

6

Extending the System to Mobile Apps

Following the success of the web implementation, I advocated to extend the framework into Trade Me’s mobile apps. This was the most significant change to mobile search in over a decade. Rather than porting web patterns directly, we took a native‑first approach, fully leveraging touch interactions:

  • New refinement chip behaviours
  • Horizontal scrolling patterns
  • Bottom sheets for progressive disclosure

To further reduce risk, the rollout was staged over eight months, allowing gradual user adaptation and continuous learning. This Delivered +3% statistically significant CTR uplift in the top 50 results.

Impact

Commercial impact

~2.3 million daily searches migrated without metric degradation

No drop in member search activity

Mobile delivered +3% statistically significant CTR uplift

No drop in conversion during migration

Slider experiments showed up to 40% uplift in price refinement usage

Product & Platform Impact

Unified refinement model across 4 business units

7 core patterns supporting thousands of refinements

Reusable architecture embedded in front-end which enabled 10+ new refinements added without a front end rebuild

Search recognised internally as cross-vertical “glue”

Organisational Capability Impact

Created a documented UI framework adopted across teams

Reduced future duplication risk

What This Demonstrates

Strategic Reframing

Redirected a fragmented multi-BU redesign into a unified platform capability without formal authority.

Strategic Leadership

Commercial Risk Management

Protected core performance metrics during high-risk migration of 2.3M daily searches.

Business Impact

Design System Contribution

Defined reusable components adopted across the broader app ecosystem.

Organisational Influence

Systems Abstraction

Reduced thousands of refinements into 7 scalable interaction models.

Systems Design

Platform Thinking

Cross-Platform Architecture

Aligned web and native interaction logic while respecting device-specific patterns.

Platform Thinking

Reflection

If I repeated this today

With hindsight, there are structural improvements I would drive earlier to increase long-term leverage and measurement rigour. While the system proved durable, I see opportunities to strengthen platform alignment and commercial attribution.

If repeated today, I would:

  • Push earlier for deeper backend/API alignment to further reduce legacy constraints
  • Formalise governance mechanisms to prevent long-term system drift
  • Tie CTR uplift more directly to downstream revenue and commercial metrics

Gallery

iOS

Android

Desktop web

Mobile web