U.S. Bancorp — Where Should the Next Customer Dollar Go?
Download case PDF (5 pages)Request teacher’s pack
Decision date and information cutoff: August 4, 2026, 11:59 p.m. U.S. Eastern Time
Decision scope: Consumer and small-business channel investment
Case convention: This is a hypothetical management decision exercise. The assignment and alternatives are not a reported U.S. Bancorp budget conflict or plan.
The decision
On July 16, 2026, U.S. Bancorp reported second-quarter Consumer and Business Banking (CBB) revenue of $2.373 billion, essentially unchanged from a year earlier, and segment net income of $589 million, down 4.4%. The bank’s broader results were stronger: consolidated net revenue was $7.712 billion and net income attributable to U.S. Bancorp was $2.177 billion. What those results did not show was whether a branch, banker, mobile feature, or credit tool created the most value for a customer or market. [S1]
U.S. Bancorp already served customers through branches, ATMs, telephone, online and mobile services, mortgage loan officers, and partners. At year-end 2025 it operated 2,075 branches across 26 states, primarily in the Midwest and West, and 4,428 ATMs. The company said a “significant percentage” of consumer transactions occurred online or through its app, but disclosed no percentage in that discussion. [S2]
For this classroom exercise, you are advising a hypothetical strategy committee on the next wave of consumer and small-business channel investment. No fixed branch-versus-digital budget dispute has been disclosed. The question is where the bank should put its next increment of discretionary effort and capital—and what evidence should determine whether to expand, change, or stop it.
Assignment: Recommend one of the following approaches, or a clearly specified combination:
- Digital-led: prioritize mobile self-service, service automation, and digital credit workflows, with human escalation for exceptions.
- Local relationship-led: protect or add banker and branch capacity in selected markets where advice, cash access, or small-business relationships may matter most.
- Market- and task-segmented: run a limited portfolio of digital and physical-channel pilots, differentiated by customer need and geography, then scale against common evidence gates.
For your recommendation, identify the customer tasks and markets to prioritize, the strongest objection to your choice, and the evidence gates that should govern further investment. Do not assume that the case’s research estimates are U.S. Bancorp results.
A multichannel franchise, with an incomplete scorecard
U.S. Bancorp described convenience, technology, service, experience, relationships, product breadth, price, and reputation as dimensions of competition. Its disclosed delivery model already spans both self-service and human channels. The strategic question is not whether the bank should “go digital” or “keep branches” in the abstract; it is which activities and customers benefit from each channel, and how channel choices affect the total relationship. [S2]
The second-quarter CBB numbers offer context but no channel attribution. Revenue was flat year over year; noninterest expense declined slightly; income before provision for credit losses and taxes rose modestly; segment net income fell. Other parts of the bank grew faster. The results could reflect prices, rates, product mix, expenses, credit provisions, or other factors. The public release does not isolate a branch or digital effect. [S1]
Management attributed consolidated performance to loan growth, record consumer deposits, broad-based fee income, and productivity. That is management’s explanation of the quarter, not evidence that a particular channel caused the outcome. Nor does growth in another business unit establish which consumer-channel investments have better economics. [S1]
What the research can—and cannot—tell the committee
Research points in more than one direction. A U.S. study of county-level 3G rollout found fewer branches and more lending after coverage expanded, with different strategies among banks that relied more or less on branches. Its model also found that older depositors who continued to value branches could bear a disproportionate share of welfare losses. The study concerns historical connectivity and county-level effects, not today’s U.S. Bancorp investment returns. [A1]
A February 2025 Census working paper examines bank AI use and small-business lending. In its specification, a one-standard-deviation increase in bank AI use attenuated the negative relationship between distance and loan growth by 46% for loans below $100,000 and 64% for loans from $100,000 to $1 million. Those percentages describe a weaker negative distance association under the paper’s specification—not increases in total lending and not results for U.S. Bancorp. The working paper is a pre-cutoff version; a later revision is outside the information set. [A2]
Research on Italian bank-firm lending suggests that AI scoring and relationship lending can interact: relationship information mattered around the COVID-era credit shock, while AI use changed some pricing and lending patterns. Its country, period, and bank-level measures limit direct transfer. Together, the studies support hypotheses about reach, human relationships, and complementarity; they do not determine the optimal U.S. Bancorp channel mix. [A3]
Exhibits
Exhibit 1. U.S. Bancorp’s disclosed channels and footprint
| Measure | Disclosed information |
|---|---|
| Branches at December 31, 2025 | 2,075 across 26 states, primarily Midwest and West |
| ATMs at December 31, 2025 | 4,428 |
| Consumer transactions online or via app | “Significant percentage”; no numeric share stated in this passage |
| CBB delivery modes | Banking offices, telephone, online, direct mail, ATMs, mobile, distributed mortgage officers, and intermediaries/partners |
Source: 2025 Form 10-K, Business Description and Business Segments. Branch and ATM counts describe footprint, not market-level profitability, overlap, or utilization. [S2]
Exhibit 2. Selected second-quarter results
| U.S. Bancorp measure | Q2 2026 | Q2 2025 | Year-over-year change |
|---|---|---|---|
| Consolidated net revenue | $7.712B | — | — |
| Net income attributable to U.S. Bancorp | $2.177B | — | — |
| CBB total net revenue | $2.373B | $2.376B | −0.1% |
| CBB noninterest expense | $1.510B | $1.517B | −0.5% |
| CBB income before provision and taxes | $863M | $859M | +0.5% |
| CBB net income attributable to U.S. Bancorp | $589M | $616M | −4.4% |
| Wealth, Corporate, Commercial and Institutional Banking net revenue | $3.771B | — | +17.1% |
Source: Q2 2026 earnings release, July 16, 2026. Dashes indicate comparison not reproduced here. Segment results are not channel-level results; the faster-growing segment includes a BTIG acquisition contribution. [S1]
Exhibit 3. Research evidence card
| Study | Setting and unit | Finding relevant to the decision | Boundary |
|---|---|---|---|
| Jiang, Yu & Zhang (2026) | U.S. bank-county panel; staggered 3G rollout, 2007–2018 | Full county 3G coverage was associated with 1.5% fewer branches and 1.2% more lending; strategy and modeled welfare effects varied by bank/customer characteristics. | Historical connectivity; county estimates and model-based welfare, not U.S. Bancorp forecasts. |
| Piao, Wang & Weng (Feb. 2025 version) | U.S. bank-county small-business lending and SBA loans; primarily 2017–2019 | AI use attenuated the paper’s negative distance–loan-growth relationship by 46% and 64% in two loan-size groups under a specified model. | Not an increase in total lending; not a company result or branch-return estimate; working paper with identification limits. |
| Gambacorta, Sabatini & Schiaffi (2025) | Italian bank-firm credit relationships, 2019–2020 | AI scoring and relationship lending interacted in lending and pricing around the COVID shock. | Different country and exceptional period; cannot calibrate U.S. Bancorp economics. |
Sources and locators: [A1]–[A3]. Findings are not directly comparable because the interventions, units, and samples differ.
Decision questions
- Which customer tasks should be primarily digital, primarily human, or deliberately multichannel? How should the answer differ for routine consumer activity, advice, service exceptions, cash access, and small-business credit?
- Which market or customer cohorts should receive the first investment? What evidence would distinguish genuine local relationship value from historical footprint or customer self-selection?
- What common scorecard and comparison design should govern pilots? Include economics, customer outcomes, access, risk, and implementation feasibility.
- What result would cause you to expand, redesign, or stop each investment? How will the bank avoid treating an academic estimate, segment result, or raw before-and-after change as a USB-specific causal return?
Sources and cutoff
[S1] U.S. Bancorp, Q2 2026 earnings release, July 16, 2026, Exhibit 99.1 to Form 8-K, “Income Statement Highlights,” “Business Segment Financial Performance,” CBB table, and CEO remarks. SEC exhibit.
[S2] U.S. Bancorp, 2025 Form 10-K, filed February 24, 2026, Item 1, “General Business Description,” “Business Segments,” and “Competition.” SEC filing.
[A1] Jiang, Erica Xuewei, Gloria Yang Yu, and Jinyuan Zhang, “Bank Competition Amid Digital Disruption: Implications for Financial Inclusion,” The Journal of Finance, first published online May 17, 2026, §§I, III–VI; Tables III and VIII. Article.
[A2] Piao, Jeffery, K. Philip Wang, and Diana L. Weng, “U.S. Banks’ Artificial Intelligence and Small Business Lending,” U.S. Census Bureau CES Working Paper CES 25-07, February 2025 version, §4.3, Table 3 columns 3–4 and footnote 8 (PDF pp. 18–19; Table 3 p. 40). Pre-cutoff PDF.
[A3] Leonardo Gambacorta, Fabiana Sabatini, and Stefano Schiaffi, “Artificial intelligence and relationship lending,” BIS Working Papers No. 1244, February 2025, §§1, 3–6; Tables 6–9. BIS paper.
The information cutoff is August 4, 2026, 11:59 p.m. U.S. Eastern Time. The July 16 earnings release is included. The Form 10-Q filed August 6, 2026 is excluded. No subsequent results or decisions inform this case.