Disclaimer: Independent educational project. Not affiliated with JPMorgan Chase. Built by Kaushal Khodifad. Data from public sources; figures are estimates.return to portfolio
Disclaimer: Independent educational project. Not affiliated with JPMorgan Chase.

The core design

A governed environment for a new card partner

The worked design at the centre of this study. We onboard a new strategic co-brand partner - illustrated as “Partner Aurora,” a national retail/airline-class program - as a first-class data-product domain. Aurora is fictional; the market context that frames it is real and cited.

3.1

The scenario, grounded in the real market

Chase scale
  • Chase is the #1 US card issuer by purchase volume - $1.344T in 2024.
  • Ahead of Amex ($1.168T), Citi ($616B), Capital One ($610B), and Bank of America ($502B).
  • Chase is the largest US co-brand card issuer.
Real co-brand roster
Amazon (largest co-brand portfolio)UnitedSouthwestMarriott BonvoyHyattIHGDisneyBritish AirwaysAer LingusIberiaAeroplan
Benefit partners (not co-brand cards)
DoorDash (DashPass)Instacart (Instacart+)
Deal structure
  • Tri-party deal

    Network (Visa / Mastercard) + issuer (Chase) + brand. The brand gets a cut even on spend elsewhere.

  • Issuer card P&L mix

    The credit/interest function is the large majority of card profitability (~80%), with fees a smaller slice (~15%); the transaction/interchange function is roughly neutral-to-slightly-negative once rewards are netted out (Fed FEDS Note, 2022). Any finer split shown is illustrative.

  • Installed base is the moat

    Deals turn on installed base + prepurchased loyalty currency - huge capital to dislodge an incumbent.

  • Multi-issuer is real

    Marriott deliberately runs multi-issuer (Chase + Amex); partners can split portfolios.

The core design tension

The partner owns the customer interface; the bank owns the account and the regulated data - so neither side can act unilaterally post-issuance. Designing the environment = standing up a governed data-product domain that lets each side see exactly what it is contractually entitled to, in time to drive the program, with PII/entitlement controls and revenue-share settlement built in.

3.2

One source of truth, two entitled views

The heart of the design: a single governed plane forks into Chase-internal and partner-facing lenses - the exact masking the simulator proves.

one governed plane
Single source of truth

One row-level-secured table. No copies, no extracts - lineage by construction.

Chase-internal lensfull · unmasked

Full, PII-governed view across all 7 product domains. Row scope = entire portfolio; masking by role + purpose.

Partner-facing lensmasked · aggregated

Aggregated, de-identified, contractually scoped. Every partner-visible metric traces to a contract clause + a governed aggregate.

The sharing spectrum - from safe aggregates to real-time offers
◀ safest - aggregatesreal-time offers - highest control ▶
01
Contractual aggregate demographics + shopping-behavior (standard)
02
Co-marketing attribution on entitled aggregates
03
Real-time transaction sharing for targeted offers (Amex-Uber pattern)
04Aurora sits here
Clean-room / row-and-column entitlement as the headline control

The partner_analyst_nova role already in the live RLS plane IS this partner-facing lens - row-scoped to partner data, PII masked, partner_reporting purpose. Switch to the Partner Aurora Analyst role in the live simulator and watch the rows scope and PII mask.

Open simulator
3.3

Seven curated data-product domains, instantiated for the partner

Each domain declares which lens it serves. Two are first-class, named, shared products.

DomainContentsConsumersPartner lens
Spend / transactionsAuths, settled txns, MCC/merchant, channel, geo, recurring / card-on-filePartner reporting, marketing, financeAggregate only
Accounts & cardmemberLifecycle, balances, credit line, status, tenureRisk, ops, CRMChase-only
Rewards / loyaltynamedPoints/miles earned-burned, redemption, prepurchased-currency liability, transfersLoyalty finance, partnerAggregate + liability share
Credit & riskUnderwriting, delinquency, charge-off, fraud, exposureRisk, regulatoryChase-only
Partner economics / settlementnamedInterchange share, revenue-share, marketing-fund accruals, contract KPIsPartner finance, deal teamShared, named product
Marketing / acquisitionCampaigns, app funnel, approval, activation, attribution, co-marketingMarketing, partnerAggregate + co-marketing
Servicing / opsDisputes, calls, digital engagementOps, CXAggregate
Rewards-liability and revenue-share / marketing-fund settlement are first-class, named products - that is what makes the deal team and partner-finance trust the environment.
The interchange/transaction function is already roughly neutral after rewards (Fed FEDS Note, 2022) and is being further compressed by regulation → analytics-led value is the new margin lever. That is the 'why now'.
Tokenization + instant issuance and rewards liability (the dominant balance-sheet item) are cross-cutting concerns across every domain.
3.4

Onboarding runbook - time-to-value as the differentiator

A partner domain provisioned in days, not a bespoke quarter-long build.

1
Register the domain

Provision the partner domain; name a Data Owner and a Data Product Manager.

2
Publish the data contract

Schema, freshness SLA, lineage, plus versioned APIs - the contract is the interface.

3
Bind entitlement policies

Via governed tags. The partner lens is de-identified aggregates by construction, not by review.

4
Stand up the reporting mart

Partner reporting mart provisioned against its SLA.

5
Wire settlement + rewards-liability

Instantiate the named settlement and rewards-liability products.

6
CDO sign-off + go-live

Office of the CDO signs off; the domain goes live - in days, not a bespoke quarter-long build.

Card Data Owner (VP)

Accountable for all card data created, provisioned, and consumed - the single point of data accountability for the portfolio.

Card Analytics Data Product Manager

Owns data-product strategy, backlog, and metrics; drives cloud migration and an API-first posture on AWS S3 + Snowflake + Databricks.

Role charters mirror JPMorgan's two publicly-posted Card roles (Data Owner VP; Card Analytics Data PM) - the closest public proxy for the mandate this environment serves. Chase does not publish its internal card data-platform topology; this runbook is a grounded reference design.

3.5

Partner reporting surface

A governed partner portal - program KPIs fed only from entitled aggregates, analogous to how Chase exposes governed reporting to commercial clients via Chase Connect.

Acquisition funnel · spend · engagement · redemption · co-marketing ROI - every tile traces to a governed aggregate, never a cardholder row. The same dashboard charts run filtered to the partner-visible aggregate set.

See the dashboards
Acquisition funnel
Spend per active
Redemption rate
Co-marketing ROI
3.6

Partner-environment success metrics

MetricTargetNote
Time-to-onboard a partner domainDays (illustrative target)The headline differentiator vs a bespoke quarter-long build.
Provable entitlement100% (target)% of partner-visible metrics traceable to a contract clause + governed aggregate.
Acquisition → activation funnelProgram KPISurfaced to the partner from entitled aggregates only.
Spend per active / redemption rateProgram KPIAggregate, contractually scoped.
Co-marketing ROIProgram KPIAttribution on entitled aggregates.
Settlement accuracy / reconciliation latencyOperationalRevenue-share reconciliation built into the settlement product.

Sources & methodology

Real · live Public-sourced · cited Synthetic · illustrative
  • 01
  • 02
  • 03
  • 04
  • 05
Synthetic dataFictional partnerIllustrative metricsGrounded-inference topology

↑ Independent educational project by Kaushal Khodifad. JPMorgan Chase is a real company in the card data & analytics space; this design and the underlying prototype were built by Kaushal as a portfolio study. Not affiliated with, endorsed by, or representative of JPMorgan Chase's actual product. Data is from public sources. Figures are estimates.

Return to portfolioOpen the live demo