Capital One: How far should the card portfolio travel?
Download case PDF (4 pages)Request teacher’s pack
Strategy and finance case
Decision date: October 2, 2026, 4:00 p.m. ET
Company: Capital One Financial Corporation (NYSE: COF)
A network choice inside a larger integration
Capital One’s acquisition of Discover created an opportunity to operate a large card issuer and a payments network under one corporate roof. It also created two migration streams that are easy to confuse. One moves Discover-originated accounts onto Capital One technology. The other routes Capital One-branded cards over Discover Network. This case concerns only the second stream.
By the cutoff, the public record showed that at least some named Capital One product variants were represented on Discover Network: an official benefits guide effective February 1, 2026 lists eligible Venture, VentureOne, Savor, SavorOne, Quicksilver, and QuicksilverOne variants “on the Discover Network” (P11). The guide does not say that every account in these product families moved, how many accounts or transactions were involved, whether these were the call’s test cohorts, or how they performed.
On Capital One’s July 21 second-quarter call, CEO Richard Fairbank described two tests: new Capital One-branded credit-card originations on Discover Network and conversion of existing Capital One accounts to the network. He said the company would choose the amount of volume and timing after testing. For existing accounts, he raised international travel, card numbers stored with merchants, the friction of changing those numbers, and possible alignment with card expiration as considerations (P02). The tests’ size, outcomes, route performance, customer response, costs, and connection to the products listed in the guide were not publicly disclosed.
At the same time, the separate Discover-to-Capital One technology workstream was progressing. At a September 16 conference, Fairbank said Discover-branded new originations were entirely on Capital One technology beginning in September, while Discover back-book conversion was proceeding in waves toward a final wave planned for January 2027 (P06–P07). Those are Discover accounts moving onto Capital One systems; they do not answer whether or how to extend Capital One card routing over Discover Network.
What is known—and what is not
Capital One had completed conversion of its debit portfolio to Discover Network. Management said Q2 included a full quarterly run-rate of related debit revenue synergies (P01–P02). That progress may make network integration appear more predictable, but debit results do not establish credit-card acceptance, customer response, or economics.
Q2 Global Payment Network volume was $189.612 billion, compared with $174.332 billion in Q1 2026; first-half 2026 volume was $363.944 billion. The prior-year comparison is heavily affected by Discover’s acquisition and migration. These combined totals do not isolate Capital One credit-card routing, the specific guide-listed products, or the two tests (P01, P04).
The broader domestic card portfolio provides operating context, not test evidence. August period-end loans held for investment were $260.642 billion; the annualized net charge-off rate was 4.16%, and the 30+ day performing delinquency rate was 3.57% (P05). Q2 net income was $3.0 billion, or $4.73 per diluted share; adjusted diluted EPS was $5.81, a company-defined non-GAAP measure. Q2 Discover integration expense was $298 million, but the filing does not identify that amount as the cost of this routing decision (P01, P03–P04).
Capital One’s original 2024 transaction materials described converting all of its debit portfolio and selected credit portfolios. A Form 425 described more than $175 billion in combined debit-plus-credit purchase volume to Discover Network by 2027, alongside a $1.2 billion network revenue-synergy estimate (P10). These are historical management estimates, not a credit-only target or achieved result.
The Q3 quarter had ended, but Q3 results were scheduled for October 20. As of this decision date, no public Q3 results or test outcomes were available (P08). The Federal Reserve and OCC had imposed Discover-related remediation and action-plan requirements; they are relevant to integration controls, not a blanket public prohibition on routing Capital One cards over Discover Network (P09).
Decision for the strategy team
Assume the integration steering group asks whether to authorize additional cohorts beyond the publicly evidenced eligibility set. Public evidence does not reveal actual live routing scope, so do not infer which accounts are already routed or recommend reversing an undisclosed rollout. The alternatives below are constructed for discussion; they are not disclosed board proposals.
- Hold the evidenced set. Do not authorize additional cohorts beyond the publicly evidenced eligibility set while collecting test and service evidence. This option does not presume the live account scope; public evidence cannot support reversing an undisclosed rollout.
- Expand in gated cohorts. Extend selectively to additional products or existing-account segments after technical, acceptance, customer-service, and stored-card criteria clear thresholds. Use travel and expiration timing as segmentation factors where supported by evidence.
- Broaden faster after readiness gates. Extend coverage more quickly once route reliability and servicing readiness meet pre-set standards, accepting greater short-run exposure to customer friction in exchange for faster network scale.
Prepare a short recommendation that states the next scope decision, the evidence required to make it, and the stop/go conditions you would use. You may recommend a different sequence, but do not assume account counts, migration volumes, test results, savings, revenue synergies, or customer effects that Capital One did not disclose. Make clear which proposed thresholds are your own decision rules.
Exhibits
Exhibit 1. Two integration streams
| Stream | Direction | Status at cutoff | Decision relevance |
|---|---|---|---|
| Capital One card routing | Capital One-branded credit cards → Discover Network | July call described tests of new originations and existing accounts; Feb. 1 guide lists eligible variants of six products “on the Discover Network.” Volumes, product-to-test mapping, and results are undisclosed. (P02, P11) | Whether and how to extend beyond the selectively evidenced product set. |
| Discover technology conversion | Discover-originated accounts → Capital One technology | Discover new originations were on COF technology beginning in September; back-book conversion in waves, final wave planned January 2027. (P06–P07) | Context only; not the routing decision in this case. |
| Debit network conversion | Capital One debit → Discover Network | Completed; management reported full quarterly run-rate of associated debit revenue synergies in Q2. (P01–P02) | Context only; not evidence of credit outcomes. |
Exhibit 2. Decision evidence and missing measures
| Evidence available | What it does not establish |
|---|---|
| Management named international travel, stored-card friction, changing card numbers, and expiration alignment as considerations for existing accounts. (P02) | Actual acceptance, authorization/decline rates, service contacts, complaints, retention, or measured friction by cohort. |
| Q2 Global Payment Network volume: $189.612bn; H1: $363.944bn. (P04) | Volume attributable to Capital One credit routing, the six named product variants, or the test cohorts. |
| August portfolio credit metrics: NCO 4.16%; 30+ day performing delinquency 3.57%. (P05) | Test-specific loss, credit quality, customer behavior, or network routing performance. |
| Original management estimate: >$175bn combined debit-plus-credit volume to Discover Network by 2027; $1.2bn network revenue synergies. (P10) | Current credit-only target, realized synergies, or a forecast for the tested/guide-listed population. |
Exhibit 3. Questions to resolve before expanding
- Which products/accounts are included in the current eligible guide, and which are in each test?
- What are authorization and transaction-success rates across domestic and international use cases?
- How do stored-card exposure, expiry timing, and customer-service contacts vary by cohort?
- What is the incremental network economics after incentives, servicing, and conversion costs?
- What controls and escalation paths satisfy integration and remediation obligations?
Source note
Source IDs refer to the cutoff-controlled research packet: P01–P11 in the source register. Company metrics and dates are as reported in issuer or regulatory materials; estimates are labeled as management estimates. Transcript text is a third-party retrospective locator for public live remarks, not an issuer-authored, verbatim-verified transcript. The benefit guide identifies eligible variants but is not an aggregate migration disclosure.