Context and problem

The company had decades of valuable data locked across disconnected systems. Title policies, lender packets, parcel data, and supporting records were stored in inconsistent formats and required manual lookup. Researchers and analysts spent hours per request and still had gaps or conflicting data. The business needed a centralized product that could handle large scale searhes and provide the data in a way that was actually useful.

Spectr was built to consolidate over 25 million owner and lender title policies and enrich them with additional property data covering more than 140 million parcels across the United States. The goal was to make the research faster, more reliable, and less dependent on specialized knowledge.

Key question: How do you centralize vast historical data at a national scale without creating a confusing or fragile research interface?

Discovery and research

  • Analytics review — Confirmed top dropoff points in the flow and average time-on-task.
  • Stakeholder interviews — Agents requested searches for Address, APN, Signle Match, Multiple Match, Grantor/Grantee, and Legal Search.
  • User interviews — Agents said they wanted specific search fields and smarter defaults.
Page 1 of excel spreadsheet showing requested Searches
Page 2 of excel spreadsheet showing requested Searches
Search Fields: Outline for searches for Address, APN, Signle Match, Multiple Match, Grantor/Grantee, and Legal Search.

User journey and information architecture

Example user goal: find the correct address for a given property quickly and record the address for the closing file.

Top IA decision: consolidate search, filters, and address details into a single landing surface to reduce context switches.

Low fidelity design and wireframes

Wireframes focused on reducing steps. I removed redundant fields, prioritized defaults, and source the information shown in context.

Wireframe: search and results layout
Wireframe: initial search layout with notes on default filters.
Wireframe: rate detail drawer
Wireframe: search flow.
Wireframe: mobile layout
Wireframe: express search layout.

High fidelity design and design system

I translated wireframes into a small, pragmatic design system that prioritized consistency and developer handoff simplicity.

Figma components page showing buttons, form fields, and cards
Component library excerpt. Each component uses the MUI Design System as a base with other customized components.
High fidelity screen showing search results
Final high fidelity screen. Note consistent spacing and clear visual hierarchy.
High fidelity screen showing search results
Final high fidelity screen. Note consistent spacing and clear visual hierarchy.
High fidelity screen showing parcel detials
Final high fidelity screen. Note consistent spacing and clear visual hierarchy.

Prototype and interaction

Prototype fidelity: medium. Clickable flows covered search to address selection and the results table. Microinteractions prioritized clarity of state and provenance visibility.

Usability testing and iteration

Method: remote moderated testing. Sample: 8 agents. Tasks: find address, record address, explain provenance. Top 3 findings below.

  1. Default filters were invisible. Fix: show active filter pill.
  2. Result rows needed clearer affordance to view provenance. Fix: add a badge and drawer.
  3. Mobile action placement caused accidental taps. Fix: moved primary action to bottom sticky bar.

Developer handoff

Handoff included: component library in Figma, tokens mapped to CSS variables, and a single-file PDF with redlines and accessible annotations.

Engineer checklist
  • Import Figma tokens to shared CSS variables
  • Verify color contrast for all interactive states
  • Implement keyboard support for inputs and tables

Results and metrics

Measurement method: A/B pilot with 500 agents over 6 weeks. Metrics below show conservative, validated impact.

  • Adoption: +80% active agents using new flow
  • Time saved: ~2 hours per agent per week on average, measured via time-on-task logs
  • Error reduction: 18% fewer address lookup errors recorded in QA logs

Qualitative: stakeholders reported fewer support tickets and agents called the new search "faster and less stressful."

What I would do next

  1. Ship analytics hooks to measure provenance clicks and refine defaults.
  2. Expand component coverage to support localization and new product types.
  3. Run a follow-up usability test focused on enterprise workflows.