Value Discovery Strategy
The Problem
PowerSeller Desktop App customers make daily decisions about investor selection, delivery timing, servicing strategies, and pool allocation. They use InfoMaker reports and DataWindow grids — flat, one-dimensional views that answer one question at a time. There is no drill-down, no multi-dimensional analysis, no way to see what they missed.
These customers are likely leaving significant money on the table and don't know it, because the tools they have can't show them.
The Strategy
Two-Level Approach
Level 1: Inspire the users (people in the chair)
Deploy interactive Superset dashboards against an anonymized copy of the customer's own data. Let secondary marketing desk operators see their portfolio from angles they've never had:
- Profit by investor, by product, by time period — interactive, drill-down
- Pipeline aging with visual alerts
- BestEx scenario comparison
- Risk position visualization
- Trading patterns and trends
This creates the "I need this" moment. The user has never seen their data this way. They become the internal champion.
Level 2: Convince the executives (money on the table)
Run PSSaaS BestEx against the customer's historical data and compare optimal vs. actual decisions. Produce a financial analysis report quantifying how much profit was left on the table.
The executive artifact is a one-page summary with a dollar amount at the top:
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
POWERSELLER VALUE DISCOVERY ANALYSIS
[Customer Name] — Trailing 12 Months
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
Total Identified Opportunity: $XXX,XXX
Investor Selection: $XXX,XXX
Delivery Timing: $XX,XXX
Servicing Optimization: $XX,XXX
Lock Management: $XX,XXX
Pool Allocation: $XX,XXX
PowerSeller SaaS Annual Cost: $XX,XXX
ROI: X.Xx
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
That's not a product demo — it's a business case that pays for itself.
Opportunity Categories
1. Sub-Optimal Investor Selection
Analysis: Compare actual delivery investor vs. BestEx-recommended investor for every loan in the trailing period.
Data needed: loan_shipped (actual deliveries), rmcat_bestex_analysis (if BestEx was run), or re-run BestEx against historical prices from rmarc_prices.
Finding example: "127 loans delivered to Investor A at 100.75 when Investor B would have paid 101.25 — 50bps × $127M volume = $635K opportunity."
2. Delivery Timing Drag
Analysis: Measure days between loan closing and investor delivery. Compare actual hold time vs. optimal delivery window.
Data needed: loan.close_date, loan_shipped.mkt_shipped_date, settlement dates.
Finding example: "Average 18 days to deliver; reducing to 12 days earns 6 additional days of carry income and reduces warehouse financing cost."
3. Servicing Strategy Gaps
Analysis: Evaluate whether buy-up/buy-down/excess servicing strategies were optimized for each delivery.
Data needed: BestEx results showing servicing type, actual servicing decisions in trade data.
Finding example: "34% of loans used default servicing; optimization across buy-up and excess servicing would have yielded $95K in additional value."
4. Lock Management Losses
Analysis: Identify loans where rate locks expired and were re-locked at worse rates.
Data needed: loan_history (fallout records with "Expired" status), re-lock rates vs. original rates.
Finding example: "43 loans re-locked after expiration at an average 12bps worse — $67K cost that better pipeline monitoring would have prevented."
5. Pool Allocation Inefficiency
Analysis: Compare actual pool fill rates vs. optimal allocation. Identify pair-off costs from under-filled pools.
Data needed: pscat_pools, pscat_trades_pools_relation, pair-off records.
Finding example: "3 pools shipped under-filled below tolerance; better allocation eliminates $28K in pair-off fees."
Technical Enablement
Data Pipeline
Anonymization
PII columns are stripped or replaced. Analytical data is preserved:
| Action | Columns |
|---|---|
| Replace with synthetic | borr_first_name, borr_last_name, borr_mid_name, coborr_* names |
| Hash | borr_social_security, coborr_social_security, all SSN fields |
| Generalize | property_street_address → remove (keep city/state/zip) |
| Preserve as-is | loan_amount, note_rate, ltv, cltv, instrument_name, pool_name, all dates, all pricing, all trade data, all risk data |
The anonymized dataset is analytically identical to the original. Every calculation, every comparison, every insight works the same.
Dashboard Design (Greg + Lisa)
Recommended starter dashboards for the value discovery engagement:
| Dashboard | Audience | Key Metrics |
|---|---|---|
| Executive Summary | C-suite | Total volume, weighted avg profit, trend over time, top/bottom investors |
| BestEx Scorecard | Secondary marketing VP | Optimal vs. actual by investor, gap analysis, money left on table |
| Pipeline Health | Desk manager | Loans by status, aging alerts, lock expirations approaching, delivery backlog |
| Risk Position | Risk manager (Greg) | Net position by instrument, hedge coverage, rate sensitivity, SFAS compliance |
| Trading Performance | Trader | Trade P&L, settlement variance, pair-off frequency, counterparty analysis |
Greg validates the risk and BestEx dashboards. Lisa validates the executive and trading dashboards. Both contribute domain knowledge for what "good" looks like vs. "needs attention."
Go-to-Market Sequence
- Select a pilot customer — ideally a HostedPS customer (Joe has access to their SQL Server backup)
- Get permission — "We'd like to analyze your anonymized trading data to show you insights your current tools can't provide. No cost, no commitment."
- Joe exports and anonymizes — database backup → anonymization script → restore to SQL MI
- Greg and Lisa design dashboards — 3-5 Superset dashboards tailored to the customer's profile
- PSSaaS runs historical BestEx — compare optimal vs. actual for trailing 12 months
- Produce the executive artifact — one-page summary with the dollar amount
- Present to the desk team (Level 1) — "here's what your data looks like in a modern BI tool"
- Present to the executive (Level 2) — "here's how much you left on the table, and here's what it costs to capture it"
Pricing Implication
The value discovery analysis naturally anchors the PSSaaS pricing conversation. If the analysis shows $400K in annual opportunity and PSSaaS costs $48K/year, the ROI conversation is trivial. The product pays for itself in the first month.
This positions PSSaaS not as a software cost but as a profit recovery tool. The executive isn't buying software — they're capturing revenue they're currently losing.
Dependencies
| Dependency | Owner | Status |
|---|---|---|
| Anonymization script | AI agent / Joe | Not started |
| Superset dashboards | Greg + Lisa + AI agent | Not started (Superset already runs in PSX stack) |
| Historical BestEx capability | PSSaaS BestEx engine | Engine built; needs historical price data support |
| Pilot customer selection | Lisa / Jay | Not started |
| Executive report template | Lisa + AI agent | Not started |