| 2024
Search Configuration Tool for E-Commerce Merchandisers
How I helped identify and define the foundations for customer-facing search through UX research and early AI exploration.
CONTEXT
Elastic Path powers enterprise e-commerce through API-first tools that allow brands to customize and scale storefronts.
At the time, Elastic Path storefronts had no customer-facing search or tools for configuring search. Both prospective and existing clients wanted this to drive product discoverability and sales. They had to otherwise pay for additional services to help them do this, and we were losing customers to competitors with native search. Leadership prioritized this blue-sky project to be a differentiator.
ROLE
UX Researcher, UX Designer
TEAM
1 PM, 1 Tech Lead
SKILLS
Generative Research
Competitor Research
Object Modelling
0-1 Research and Design
IMPACT
0-1
blueprint for bringing search configuration to Elastic Path.
Validated
demand for AI-supported, data-driven tools.
Future-Ready
Established a scalable foundation for future AI features and search analytics.
PROBLEM
Customers wanted search, but Elastic Path had no infrastructure, and no workflows for merchandisers to set up search for their storefronts.
Both prospective and existing clients wanted this to drive product discoverability and sales. They had to otherwise pay for additional services to help them do this, and we were losing customers to competitors with native search.
MY RESEARCH APPROACH
I partnered with tech leads and PMs to understand technical possibilities. None of us had prior search expertise, so I led foundational research, in collaboration with my tech lead.
Main research goals.
01
Identify the ideal shopper experience for our clients' customers.
02
Identify ideal merchandiser experience.
03
Identify what our clients what to achieve through search.
DEFINING THE IDEAL SHOPPER EXPERIENCE
To understand what shopper experience merchandisers might be setting up search for, I first mapped out how search typically works in commerce — from indexing to attribute configuration to user experience.
I also identified a general e-commerce search journey for a shopper throuhg interviews of shoppers.
USER INTERVIEWS TO IDENTIFY THE IDEAL MERCHANDISER EXPERIENCE
I interviewed 8 merchandisers who had various kinds of products like furniture with 10 different upholstery options, merchandisers running a rental designer bag business and merchandisers on teams of businesses such as Ugg Boots and Breville coffee.
User Goals
Search as a sales tool.
Merchandisers wanted precise control over which attributes were searchable (SKU, product name, size, color) because they viewed search as a way to drive sales outcome.
For example, making sure seasonal attributes like "color" could be promoted during certain campaigns.
Flexible search configurations.
They wanted flexibility to create different search configurations for different catalogs or product categories.
For example, Apparel needed attributes like size and color searchable; Electronics might prioritize brand and model number.
Search to solve business needs.
They wanted search configurations to align with business priorities, not technical defaults.
For example, making certain attributes more searchable to help move overstock inventory.
They were unsure which attributes mattered most for customers.
Interviews also helped me identify the ideal user journey for setting up search as described by merchandisers.
COMPETITOR RESEARCH
I analyzed competitors like Algolia, Salesforce Commerce Cloud, and Shopify to understand best practices and gaps.
Feature
Manual search rules by merchant, basic AI features .
Inspiration
AI-guided suggestions.
Most Relevant for MVP
Feature
Attribute-based search configuration, AI tuning
Inspiration
AI-guided suggestions.
Feature
Relevance rules, category-based search controls.
Inspiration
AI-guided suggestions.
Feature
Predictive modeling of search performance
Inspiration
Simulating keywords as user feedback.
MVP SCOPING AND FINAL SOLUTION
Through research, we identified Search Profiles as the solution:
A reusable, named configuration that defines which attributes customers can search. Flexible enough for complex catalogs, simple enough for daily workflows.
I shared research findings through multiple working sessions with PMs, engineers, and architects. This helped identify: What was achievable for MVP and what could be phased or outsourced. These decisions were mostly constrained by technical resources and budget.
Search Profiles
Relevance based autocomplete and partial search synonym matching.
Attribute Selection
Auto-recommend attributes based on historical search logs and customer behavior.
Search Result Boosting Rules
A tool that lets merchandisers boost search results by weighting product attributes and setting business-driven rules to prioritize visibility.
Profile Management
Create, edit, delete, and apply profiles across catalogs.
FUTURE ENHANCEMENTS, SOME AI-POWERED
The MVP laid the groundwork for future AI features that were identified as differentiators against competitors.
Autocomplete
Relevance based autocomplete and partial search synonym matching.
AI-Attribute suggestion
Auto-recommend attributes based on historical search logs and customer behavior.
Search Performance Analytics
Track impact of search configurations on key metrics over time. Predict how changes to search profiles would impact discoverability, clickthrough rates, and conversion.
