Project Snapshot

Solving Spruce’s $250K Problem
Spruce is a prop-tech/fintech platform that simplifies title and closing services across the U.S..

I led the end-to-end redesign of escrow transactions, eliminating $250k+ in annual inefficiencies, cutting manual work by 37%, and speeding up reconciliation by 94%.
(Zillow acquired Spruce in Sept. 2023)

Friction Identified
The Finance team was drowning in spreadsheets, wasting 9+ hours a day reconciling transactions and relying on developers for manual fixes. The cost was upwards of $250K annually.

We needed to eliminate developer bottlenecks, automate transaction matching, and surface only the high-level exceptions that required human review.

Product Type
Enterprise, SaaS (Desktop)

My Role
Lead Product Designer

Team
Product Manager, Director of Design, FE + BE Engineers

Duration
4 Months

Business Impact
πŸ“‰ 37% Reduction in Manual Operations
πŸš€ 94% Faster Reconciliation Workflows
πŸ›‘οΈ 98.6% Regulatory Compliance Achieved

Phase 1 – Discovery

Service Blueprint Workshop

After the kick-off, I gathered the key stakeholders and held a service blueprint workshop to better understand the end-to-end workflow from the finance team’s perspective. I used my initial hypotheses as a framework to determine the next steps, specifically focusing on:

  • “What happens when a new order is opened?”
  • “What occurs at the end of the lifecycle?”
  • “What are the steps for reconciliation when something goes wrong?”
  • “What does a ‘successful’ fully funded order look like?”
  • “Which external apps or tools are being used outside of the platform?”

This exercise served as a knowledge check. By asking these questions, we were able to map out the business process, uncover hidden inefficiencies, and ensure we could trace issues back to their source if they occurred.

(Fig. 1 β€” Collaborative Service Blueprint Exercise)

Outcome

The workshop was a foundational piece that guided me to continue exploring ways to close gaps in my understanding. It became clear that the current flow and UI were performing suboptimally and weren’t up to industry standards.

With a bird’s-eye view, it became easier to generate and test ideas to determine which changes had a positive impact on users without risking the business.

I identified a need to visualize multiple processes and their sequences within a single document. Ultimately, this enabled us to have better conversations, increase efficiency, and provide proper documentation for the team.

(Fig. 2 β€” Flowchart)

Observing Accountants

During these sessions, I noticed that critical order details were often captured on sticky notes and kept on desks.

To find specific data, accountants relied on “Cmd/Ctrl + F” to scan long, continuous lists of transactions.

Another key insight was the urgent need for collaboration tools to streamline tasks, set reminders, and manage comments within the platform itself.

(Fig. 3 β€” Observation sessions with the accounting team)

Phase 2 – Define

Interpreting The Findings

With competing priorities at play, we leaned toward solving for the business risks we had recently uncovered. After a prioritization session, I conducted a UX audit of the existing UI to map out broken processes. This gave the team a clear roadmap for where to iterate and how to inform future decisions.

Previously, the Real Estate Ops team had to manually input every incoming transaction and outgoing disbursement on a per-order basis. Much of this data was already captured elsewhere in the platform. This repetitive work was a major bottleneck and prone to manual error. After extensive interviews with the Ops team and funders in both New York and Texas, we identified this as the most urgent feature needed to streamline the workflow.

(Fig. 4 β€” UX Audit, Feature Roadmap Prioritization)

Phase 3 – Experimentation

Architecture Redesign

I synced with the engineers after the workshop to align our design goals with the realities of a complex microservices setup.

One drawback of microservices is that the entire system is more complex, and it becomes harder to track down issues without additional tooling; monolithic applications, by contrast, are self-contained and typically easier to debug.

With the new insights from users in mind and alignment with engineering, we (product + design) prioritized work accordingly.

(Fig. 5 β€” UX Audits & Iterations)

Solution 1 – Net New Feature

Transaction Management

The process begins by importing transactions, followed by an auto-match execution, confirmation of suggested matches, and the creation of manual matches.

This tool quickly matches high volumes of transactions and flags discrepancies, allowing accounting staff to focus on complex cases and high-level analysis. Once matching is complete, the platform stores confirmed matches and provides suggested pairings or unmatched transactions for the Finance team to review for reconciliation.

With this automated heavy lifting out of the way, the Finance team can focus their time on investigating only the most challenging exceptions.

(Fig. 6 β€” Escrow Transactions)

(Fig. 7 β€” Escrow Transactions, transaction management)

Solution 2 – Net New Feature

Split Transactions

The Split Transaction feature allows users to divide a single transaction into multiple categories by dollar amount or percentage. This works for most posted transactions, including negatives. The primary goal is to let users accurately allocate specific portions of a transaction to different accounts for better reconciliation.

(Fig. 7 β€” Split Transactions)

Reflect

Takeaways

Shifting timelines, resourcing issues, and reprioritization meant the project scope was constantly evolving. I had to adapt quickly to ensure we delivered high-quality designs under tight deadlines.

Through our research, we identified several core issues that differed from the original problem statement. By presenting this evidence, we led the business team to pivot. Don’t hesitate to propose a change when the data supports it. Thanks to our advocacy, we developed a solution that improved both user efficiency and business outcomes.