Product Design Evaluation

Real-Time PnL Dashboard for Commodities Trading

Replaced fragmented spreadsheets and delayed end-of-day systems with real-time PnL and position visibility across 10+ global trading desks.

Role Technical Product Manager
Status Completed
Year 2024
Trading Systems Real-Time Data Fintech Product Strategy
~75% Reconciliation time saved (4 hrs → 1 hr)
~40% Reduction in reporting discrepancies
10+ Global trading desks adopted

Problem

Traders relied on fragmented spreadsheets and delayed end-of-day systems to estimate intraday PnL. This led to inconsistent numbers, delayed decisions, and recurring reconciliation conflicts with finance teams — particularly during volatile trading periods when accuracy matters most.

Opportunity

Enable real-time visibility into PnL and position discrepancies to improve trading decisions and reduce operational friction between trading and finance. The goal was not perfect accuracy on day one — it was directional clarity, immediately.

Design Decisions

Hybrid estimation model

Built a hybrid system that combined trader-input estimates with system-calculated PnL for real-time comparison. This gave traders an immediate reference point while the system-calculated leg caught up — avoiding a "wait for perfect data" delay that would have killed adoption.

Discrepancy highlighting over data accuracy

Prioritized surfacing discrepancies visually over claiming perfect accuracy at launch. Traders needed to know where the numbers diverged, not just what the final number was. This framing also built trust — the system showed its work.

Unified aggregation layer

Integrated multiple upstream data sources through a single aggregation layer rather than building point-to-point connections. This added upfront architecture complexity but made adding new data sources significantly cheaper over time.

Trade-offs

What we gained

  • Fast to ship — launched before end of quarter
  • High trader adoption from day one
  • Improved decision speed during volatile sessions
  • Clear reconciliation reference for finance

What we gave up

  • Partial data accuracy at launch
  • Required disclaimers on estimated figures
  • Manual override workflows for edge cases

Opportunity Cost Evaluation

Delaying launch to achieve full data reconciliation would have pushed delivery past peak volatility periods — the exact moments where the product created the most value. The cost of waiting was higher than the cost of shipping with partial accuracy and clear labeling.

Key trade-off

Full reconciliation before launch → vs. → hybrid system with clear disclaimers, shipped during a volatile quarter. We chose speed. Reconciliation accuracy improved iteratively in v2 and v3.

Success Metrics

What's Next