This portfolio is password-protected.
Case Study
I spent three years building, documenting, and championing an evidence base for financial institution insights at Plaid. The research outlived its initial rejection and became the foundation of Plaid Bank Intelligence — now an org-level priority with 39 beta partners.
In early 2023, I was researching a dashboard concept that would show financial institutions which apps their customers used. During discovery, the conversation kept drifting from the dashboard to something more fundamental.
The unmet need
FIs wanted to segment and analyze usage patterns — to identify their primary customers, the apps those customers used, and any apps competing with the FI's own products. They assumed aggregators already had this capability.
Despite being a clear, recurring priority, the PMs were hesitant. The visibility would help institutions on Plaid's free API standard but could strain relationships with paying customers. The finding was acknowledged and set aside.
JTBD synthesis: 70+ interviews, two top jobs
Understand My Consumers' App Usage
Evidence level: Highest (Top Job). FIs wanted to know which third-party apps their customers connected to, how frequently, and what it revealed about financial behavior.
Manipulate Aggregator Data for Consumer Segmentation
Evidence level: Very High. FIs wanted to cross-reference app usage with customer segments to identify cash cows and competitive threats.
Every research touchpoint over the next two years reinforced the same signal. Meanwhile, external forces were removing the objections that had blocked the idea.
Early 2023
Initial discovery. Dashboard research surfaces segmentation as the real need. PMs decline — risk to paying customer relationships.
Mid 2023
JTBD synthesis. 70+ interviews confirm two top jobs. I document findings in year-end reports and champion them through ongoing conversations.
Late 2023
CFPB 1033 ruling. Open banking established as a consumer right. Market softens toward data portability — reducing the competitive-risk objection.
2024
Ongoing research reinforces. FI stakeholders consistently say segmentation is crucial for Plaid API adoption. Smaller institutions see insights as leveling the playing field. Competitors already offering basic versions.
Mid 2024
The PRD moment. While I'm on parental leave, the PM who initially rejected the idea digs up my JTBD synthesis and inserts it as key justification in a PRD for an enhanced FI insights product.
October 2025
Plaid Bank Intelligence launches. Retention Score enters beta — the first product in the suite.
May 2026
Suite expands at Plaid Effects. Fraud Insights + Loyalty Insights announced. Four new capabilities. Product language maps 1:1 to JTBDs.
What happened
The PM who had initially balked at the insights product dug up my JTBD artifacts and year-end reports and used them as key justification for a new product. The research didn't persuade in the moment — it was available when the moment arrived.
The research survived because it was preserved in multiple formats: JTBD synthesis decks, year-end reports, and interview documentation that a PM could find and pull up independently. Research that is documented, repeated, and contextualized outlasts short-term roadmap pressure.
What launched maps almost 1:1 onto the JTBDs I surfaced years earlier.
Primacy Score
User-level 0–100 score built on share of network, competitive standing, and engagement quality.
Research origin: the "competitive-context segmentation" banks asked for in 2023
Retention Score
Flags direct deposit churn risk weeks or months before internal FI systems would detect it.
Research origin: "understand where you stand in your customer's financial life"
Connection Risk Score
ML-driven fraud detection at the third-party app connection event.
Research origin: fraud prevention needs surfaced in FI interviews
Network Alerts
Network-wide account takeover signals pushed to banks in real time.
Research origin: cross-institution fraud patterns FIs said they lacked visibility into
"With Bank Intelligence, our customers now have a suite of tools that give them insights into their relationship with their users as well as the safeguards they need to protect that relationship... It's also a meaningful opportunity to convert data partner-only relationships into Plaid customers."
— Internal launch announcement, May 2026
"By 2028, Bank Intelligence is the substrate for our DP relationships — DPs rely on our insights for better business outcomes on user engagement and fraud reduction, and being integrated with Plaid is value-accretive enough that DPs seek deeper participation."
— 2026 NEA two-year strategy
"Share of network," "competitive standing," "engagement quality" — these constructs started as bullet points in a 2023 interview synthesis.
Org-level NEA goal for 2026
Bank Intelligence is one of three NEA organizational goals, with explicit key results: land 5 paying DP/platform customers by EoY, inbound fraud signals from 5 DPs. H1 commitment: 4 betas + MVP launch of bi-directional fraud signals.
Named partners in active betas
Retention: Step, Current, Chime, GreenFi (active churn data contributions). Fraud: Q2 (committed H1 2026 go-live for ATO signals), Fiserv, Alkami, Jack Henry Banno.
Net-new pipeline from GTM outreach
March 2026 outreach produced first meetings with direct FIs (Veridian CU, Elevations CU, Found) and platforms (Jack Henry, Bankjoy, Mahalo). Found committed to internal Primacy Score analysis. Target audience expanded by 70+ DPs.
First commercial amendments signed
Bank Intelligence-specific contract amendments with Chime (signed Feb 2026), Alkami, and Jack Henry/Banno — granting 3-month evaluation period in exchange for structured feedback.
Codified as long-term DP value exchange
The 2026 two-year strategy designates Bank Intelligence as the data partner value exchange: "By 2028, Bank Intelligence is the substrate for our DP relationships." The original insight is now the organizing principle for an entire side of Plaid's business.
Adjacent research validates baseline
In a follow-on DPD Insights study I ran in late 2025, participants said app usage and ecosystem insights were "already largely satisfied by Bank Intelligence" — letting me redirect discovery toward newer unmet needs.
Documentation survives churn
Research that is documented, repeated, and contextualized survives leadership churn and short-term roadmap pressure. The finding was rejected when first presented — but because it was preserved in JTBD artifacts, year-end reports, and interview documents, a PM could pull it up independently a year later.
Long-horizon work compounds
The same interview corpus that informed Bank Intelligence is still feeding adjacent decisions: the 1033 Compliance research, the 1PN/Passport recommendations, and the 2-Way Value product research plan. One body of evidence, multiple product outcomes.
What I'd change
I should have been more persistent in following the product through development, not just the research phase. I checked in occasionally, but frequent, deliberate check-ins would have given me more influence on product direction — and more importantly, let me advocate for measuring outcomes like user satisfaction and adoption quality rather than just shipping features. The research got the product built; staying closer to the build would have made it better.
The bigger lesson
A single research project is rarely enough to change a roadmap. A continuous research program — refreshed, re-synthesized, and re-championed across cycles — eventually does. Market shifts create the moment, but only if the evidence is already in place.