Lending · Build phase · Validation-to-ship

Validated, built, shipped, sunset, in nine months.

An unbranded public Android prototype that took a Citi D10X concept from a validated demand test through a working product in market for three months. Reduced eventual Citi-branded design effort by 50% and tech effort by 15%.

Challenge.

Sum-It is a Citi D10X intrapreneur concept tackling a single problem: helping young adults pay off student debt faster by routing roundup-style spare-change contributions from credit and debit transactions into accelerated student-loan payments, with a family-contribution mechanism that lets parents pledge alongside the borrower. By October 2018, two demand tests had answered "should we build this?" with strong signal (best-in-D10X CPL, ~10% landing-page conversion). The decision Citi brought us was different: how to de-risk the build before bringing it inside Citi's brand and regulatory perimeter. Historically D10X startups jumped straight from demand test to a full Citi-branded build, expensive, slow, and biased toward existing-customer feedback. Rush wanted to try something the program had not done before.

Approach.

• Unbranded public prototype distributed through Google Play under the neutral sum-it.co identity, acquiring real non-Citi-customer users.
• Yodlee-integrated Android native app with full account-linking, goal-setting, family-invitations, and dashboard. Native chosen over responsive web for Yodlee integration and review-cycle reasons.
• Three-month live experiment program with a 17-experiment register. Eight numbered sprints, two to four weeks each, with one to four experiments per sprint, each carrying hypothesis, test design, result, action, and key learning.
• Pre-launch user testing three weeks before launch (three moderated sessions via UserTesting.com) flagged that users did not naturally grasp the roundup concept, driving onboarding rewrites before ship.
• Clean sunset path planned from day one. The MVP closed cleanly when the program ended, with leads not portable back to Citi-branded due to PII restrictions, a known and named cost.

Outcome.

• Built and shipped in 8-10 weeks, then ran live for three months in the Google Play store.
• 134 real users acquired at $209 blended CAC across $28K of ad spend.
• Design effort cut 50% for the eventual Citi-branded version, tech effort cut 15%, by virtue of having a working product spec and live-tested UX before the Citi-perimeter scoping started.
• Closed cleanly at the end of the live phase, with no user-data portability issues.

Methodology lessons.

• The unbranded public prototype is a real and underrated tool for enterprise innovation arms that need to test product-market fit without committing brand risk and regulatory perimeter.
• The 17-experiment in-flight program is the cadence model for any post-launch retainer. Sum-It's experiment register is the worked example for our Continuous Validation methodology.
• Pre-launch user testing surfaces UVP comprehension issues that quantitative landing-page tests can miss. Three sessions are enough to flag the issue.