How members begin, borrow, and deepen their relationship with Abound

Research summary

During the July 20 session, the conversation moved from how members begin a relationship with Abound to what happens when they apply, borrow, and return for another need. Along the way, five questions came up. This report follows those questions:

1. Is checking still Abound's best entry product?
2. Where do applicants leave the process?
3. Does direct deposit still define a primary relationship?
4. How much fraud risk can Abound accept as it removes friction?
5. Can Abound use known-member data to reduce repeat work?

Together, the questions point to one shared challenge: making it easier for members to begin, complete, and deepen a relationship without weakening the controls that protect them. The research suggests five findings:

• Current external research supports checking as an entry product. Abound's own May 2026 analysis shows a different pattern worth investigating: members in its VISA row added another product within three months more often than members in its Checking row. That is an Abound correlation, not proof that a card causes a deeper relationship.
• Speed matters after a loan application begins. Recent research ties delays in approval and funding to lower satisfaction. A separate industry study shows substantial abandonment during online checking account opening.
• Direct deposit is useful, but it does not describe the whole relationship. Current definitions of “primary” combine direct deposit with card use, bill payment, and other recurring behavior.
• Fraud controls and member ease need to be measured together. The evidence supports tracking both fraud loss and the legitimate activity interrupted by fraud controls. The right balance will differ by product.
• Known-member applications can reuse information Abound already holds. A recent credit-union trial found that one additional requirement sharply reduced loan acceptance. Federal guidance allows lenders to consider deposit-account activity when assessing creditworthiness.

The sections below explain the evidence behind each finding and identify the questions Abound can answer with its own data.

Question 1: Is checking still Abound's best entry product?

The July 20 discussion raised a choice between the traditional checking-led relationship and the stronger product-deepening pattern Abound has seen among VISA holders. The outside evidence still favors checking as an entry point. Abound's own analysis gives the team a reason to look more closely at what happens after each product is opened.

External evidence

J.D. Power's research on customers who open accounts at a second institution found that 54% of new checking accounts become the customer's primary checking account. New credit cards become the customer's primary card 21% of the time. The published page does not disclose a sample size.

Elan reports that 80% of new cardholders already carry at least one card, which makes a new card more likely to be an addition than a first financial product. The article does not disclose the study method.

The 2026 Filene CardFit pilot tested card-led acquisition with seven credit unions. Filene reported strong initial engagement but limited conversion to applications and approvals.

Abound's May 2026 analysis

Abigail and Annabelle's May 2026 Member Trends analysis measured whether members in each product row added another product. Within three months, 19% of members in the Checking row added a product, compared with 41% in the VISA row.

The analysis establishes a correlation inside Abound. It does not establish that opening a VISA account caused the difference. The table does not state the number of members in each group or the period covered, and the groups may differ in age, tenure, direct deposit, credit profile, or other ways.

Questions for Abound's data

• How many members are in the Checking and VISA groups, and what period does the analysis cover?
• Which member differences explain the 19% and 41% results?
• How often does each product or behavior begin a relationship that later includes borrowing?
• Does the pattern hold when Abound compares members of similar age, tenure, and credit profile?

Question 2: Where do applicants leave the process?

Recent evidence focuses on the time between application, approval, and funding. Research on online checking also shows that abandonment can be large enough to measure directly. Abound's own journey data can identify the step where the problem occurs.

Approval and funding speed

J.D. Power's 2026 Consumer Lending Satisfaction Study measured 6,513 US personal-loan customers between March 2025 and March 2026:

• Satisfaction fell 41 points when approval took more than one hour.
• Satisfaction fell 47 points when funding took more than one day after approval.
• Nonbanks funded within one day for 68% of customers, compared with 58% for banks.
• Approval times were similar across lenders. Funding speed separated them.

The study covers people who completed a loan, so it measures satisfaction rather than abandonment.

Online account-opening abandonment

Cornerstone Advisors and Alkami reported 3.36 abandoned online checking applications for every account opened. Their 2026 report uses 2025 data supplied by 89 retail institutions, including 57 credit unions. Alkami commissioned the work and sells digital-banking software.

Cornerstone identified identity verification and switching devices during an application as common sources of friction. The 3.36-to-1 figure applies to online checking account opening, not consumer lending.

Questions for Abound's data

• What share of applicants leave at each step of the loan and account-opening process?
• Which steps create the longest delay between approval and funding?
• How do completion and funding time differ by product, channel, and member status?
• How often does a member begin on one device or channel and finish on another?

Question 3: Does direct deposit still define a primary relationship?

Current research treats “primary” as a bundle of recurring behaviors, not a synonym for direct deposit. The useful question for Abound is which behaviors predict continued use, product depth, and value.

Primacy extends beyond direct deposit

Chime's 2025 SEC filing defines a primary account relationship as either 15 or more card purchases in the previous month or at least one qualifying direct deposit of $200 or more. Chime reported that 67% of its 8.6 million active members met that definition as of March 31, 2025.

The definition matters because it treats card use and direct deposit as alternative signs of an active relationship. Chime also cautions that its definitions may differ from those used by other companies.

A July 2026 PYMNTS Intelligence and Velera study surveyed 14,218 US consumers. Sixty-one percent of credit-union account holders called the credit union their primary financial institution, while 48% of credit-union cardholders put its card at the top of their wallet. The results show that institutional primacy and product preference are related but different measures.

Pinwheel's Consumer Banking Trends 2026 surveyed 500 consumers and defined primary relationships largely through direct deposit and bill payment. Pinwheel sells direct-deposit switching services, so its commercial position is useful context when reading the study.

Ron Shevlin described checking accounts as “paycheck motels” in 2019: places where income arrives before the money moves elsewhere. The phrase is a framing, not a measure.

Questions for Abound's data

• Which behaviors belong in Abound's definition of an active primary relationship?
• Does direct deposit predict retention, product depth, borrowing, or profitability?
• Do card use, recurring payments, or money movement add predictive value beyond direct deposit?
• What share of members hold one product, and how has that share changed since July 2024?

Question 4: How much fraud risk can Abound accept as it removes friction?

The evidence does not support one universal amount of “good friction.” It does show that members distinguish between security checks and delays that do not make sense to them. The tradeoff needs a separate measure for cards, account opening, and lending.

Members accept identity checks in a survey

FICO's 2025 survey of 1,000 US bank customers found that 75% said tougher identity checks would not cause them to abandon an application. Thirty-two percent ranked fraud protection as their top need when opening an account, compared with 25% who ranked ease of use first.

The survey addresses identity checks. It does not address income-document requests, manual review, or delays in funding.

The financial exposure differs by product

Javelin's 2026 Identity Fraud Study, based on more than 115,000 consumers, estimated $7 billion in new-account fraud during 2025, up 13% from the prior year. Javelin attributed the growth in part to weaker identity verification, automated attacks, and attempts to reduce onboarding friction. That is Javelin's interpretation, not a controlled estimate of the effect of removing a specific check.

TransUnion's October 2025 analysis reported average loss per fraud incident of $19,611 for auto loans, $3,427 for unsecured personal loans, and $940 for credit cards. The analysis used US loans originated from March through September 2023 and observed them for two years. TransUnion did not disclose the sample size.

Point Predictive's 2026 auto-lending report estimated $10.4 billion in US auto-fraud exposure. It also reported that up to 70% of early-payment defaults showed evidence of fraud in the original application. That is a correlation; it does not show that faster lending caused the loss.

Questions for Abound's data

• What are gross fraud loss, recoveries, and net fraud loss for each product?
• How many legitimate applications or transactions are declined, delayed, or sent to review?
• How much loss does each control prevent, and how much legitimate activity does it interrupt?
• What risk limit can Abound state as a dollar amount or rate for each product?

The separate Fraud and member value research brings together published measures that can inform Abound's fraud-to-member-value ratio.

Question 5: Can Abound use known-member data to reduce repeat work?

Abound's question is practical: when a current member applies for another product, which information can Abound reuse so the member does not have to begin again? Current evidence supports reducing unnecessary steps while preserving the information needed to make a sound decision.

An added requirement can stop a willing borrower

A 2023 randomized trial published in the Review of Financial Studies, with an accessible NBER working-paper version, followed 1,531 interested members at St. Louis Community Credit Union. One group could open a credit-builder loan immediately. Another was told it first had to complete about 50 minutes of free online financial education.

Loan take-up fell from 30% in the standard group to 12% in the extra-step group. The decline occurred when staff stated the requirement; only six people began the course. The specific product and requirement differ from Abound's application process, but the trial shows how one added step can change behavior.

Deposit activity can inform underwriting

The 2020 Interagency Lending Principles for Offering Responsible Small-Dollar Loans state that lenders may use internal or external data, including deposit-account activity, to assess creditworthiness.

The 2019 Interagency Statement on the Use of Alternative Data in Credit Underwriting notes that cash-flow data may improve the evaluation of borrowers with variable income. FinRegLab's empirical work found that cash-flow variables added predictive value across six nonbank lenders.

Current products show how this can work in practice. Huntington's Standby Cash bases eligibility in part on checking-account history and recurring deposits. Pew documented similar small-dollar products for existing checking customers at six large US banks.

Questions for Abound's data

• Which information does Abound already hold when a current member applies again?
• Which fields does the member have to re-enter or document?
• Could Abound establish standing eligibility for a defined group of current members?
• Which existing account behaviors predict repayment well enough to replace a repeat request?
• How do completion, approval, loss, and member effort change when known data is reused?

Market context

Credit demand remains substantial, and access is uneven. That makes approval, denial, and friction questions relevant to both growth and member impact.

The New York Fed's Credit Access Survey reported a 16.1% rejection rate in June 2026, down from 23.1% in June 2025. The application rate was higher than in any survey wave since October 2021.

The Federal Reserve's 2025 Survey of Household Economics and Decisionmaking found that 33% of adults applied for credit during the year. One-third of applicants were denied or approved for less than they requested.

The CFPB's 2024 Making Ends Meet survey found large differences in credit outcomes by race. Among applicants, 64.0% of Black consumers were denied or received less than requested, compared with 31.5% of non-Hispanic White consumers. Differences remained after the analysis accounted for education, income, and age.

Questions for Abound include:

• How do approval, denial, and funding outcomes differ above and below a 650 credit score?
• What share of denials does each reason account for?
• How do outcomes differ by product, channel, member status, and protected group?
• Which applicants return to Abound after a denial, and what happens next?

Research limits

The evidence combines consumer surveys, institutional benchmarks, regulatory guidance, vendor research, and one randomized credit-union trial. These sources answer different questions and cannot be treated as interchangeable.

The report identifies sources sponsored or published by companies that sell the products being discussed. Sponsorship does not make a result false, but it makes the study design and scope important.