Maximilian Kwok's Day One Plan for Affirm

This is a real Day One Plan. Maximilian made it by working one live brief for Affirm, and the page is built around that work.

Maximilian Kwok · aiming at roles at Affirm

Maximilian makes an uncertain weekend payment date legible at Affirm without promising when a bank will post it.

Shoppers with Pay in 4 plans whose installments fall on weekends or bank holidays miss them even though they opened the reminder.

Busra Erdogan, hourly retail employee, plans groceries around Friday paychecks, and a surprise withdrawal forces her to cut essentials or borrow.

After reading what Maximilian proposed

“If you clearly say Sunday is still my due date but my bank may show it later, I can plan around it. I still need that message before I commit at checkout.”

Make your own

In the Affirm work

The plan’s deepest concentration is in locking test rules before copy ships, fitting a repayment problem where urgency can mask harm. It holds audience and cadence fixed, pre-registers a half-gap target and contact guardrails, and gives Operations pause authority. A second demonstrated concentration separates the due date from processor initiation and bank posting, then tests whether customers actually understand that distinction in rendered copy. The work protects a truthful explanation from becoming a timing promise.

Adjacent — this work went elsewhere

Exactly-once dispatch control is the adjacent area here, developed elsewhere rather than in the Affirm experiment. It sits at the edge of this payment work as a related concern with irreversible actions and uncertainty, adding breadth without being attributed to this plan.

Affirm

+6more

You get one question here, on us. Make it a hard one.

Maximilian met Busra’s problem with . Ask how it relates to yours.

Where the answers come from

The brief

A live brief for Affirm

Buy-now-pay-later payments for consumer shoppers

Day One Me wrote this brief for Affirm — Affirm did not commission it. The customer quoted below is a persona built from Affirm’s market, not a real person.

What surfaced

Shoppers with Pay in 4 plans whose installments fall on weekends or bank holidays miss them even though they opened the reminder; some expect the debit on Friday, while others expect it on Monday.

Busra Erdogan, before the work

“I saw Sunday, but I didn’t know if the money would leave Friday or Monday.”

Hourly retail employee — Uses Pay in 4 for essentials and plans spending around a Friday paycheck.

What the brief asked for

Reduce weekend-related missed installments by making scheduled debit timing understandable before and after checkout.

What would count as success

Weekend-scheduled installments land on time as often as weekday-scheduled ones, closing the 4.2 percentage-point gap, with fewer timing-related contacts.

Fixed constraints going in

Debit timing follows network and banking rules that vary by funding method; policy language cannot imply a payment date that is not guaranteed; changes span checkout, account servicing, and notifications.

What they had to weigh against each other

repayment performance vs. customer pressuresimple dates vs. rail complexityconsistent explanation vs. channel constraints

The approach

What Maximilian did about it

What they built

A payment-truth problem under variable bank processing

The solution proposed a single-launch, three-arm randomized experiment for Pay in 4 customers with weekend or holiday installments. A clarity arm would use consistent messaging across checkout, reminders, and servicing, while a reminder-intensity arm would test the external vendor's faster alternative against current-state control. Scale would require closing at least half the weekend on-time gap without increases in reversals, disputes, or timing-related contacts. The approach excluded unsupported predicted debit timing because initiation and posting vary by rail and bank.

How they got there

Steering byPredictable due-date understandingCustomer harm containmentCausal measurement rigorComparable intervention testing

The proposal treated weekend missed payments as confusion about a stable due date versus variable bank activity. It proposed a three-arm test of canonical clarity messaging, current experience, and intensified reminders across repayment touchpoints.

Platform constraints ruled out universal Friday or Monday promises, while service feedback required explicit confirmation that the due date stayed unchanged. The design accepted a slower, frozen common launch to protect attribution and avoid unsupported expectations or early-payment pressure.

Taken as given

A canonical message can pass comprehension testing across checkout, reminders, and account servicing.
The eligible weekend and holiday cohort is large enough for a conclusive three-arm test within the quarter.
The vendor can hold audience, cadence, suppression, and non-copy elements fixed across test arms.
Numeric harm thresholds and pause authority can be approved before launch.

Deliberately left out

Predicted debit timing and initiation notifications were deferred from the first release.
Funding-method-specific message variants were deferred pending safety-read results.
Broad reminder-cadence changes outside the controlled test
Post-freeze messaging changes were deferred until after the common launch decision.

What they didn’t build

A reminder-volume problem around weekend due dates

Argued for in the session by the Enterprise Account Director — their case is under Check it.

What that protects, and what it spends

The approach protected payment truthfulness, causal attribution, and repayment-quality guardrails. It sacrificed faster deployment, predicted debit-time messaging, and post-freeze flexibility by requiring a common launch and fixed campaign elements. That cost was explicitly accepted because variable processor and bank outcomes could otherwise create unsupported expectations and confound results.

And not only from this one

Maximilian has 6 other Day One Plans, and the same Day One Me reads all of them. Open one to read the brief it was given and what was done about it: , , , , and .