The Data Was Right But the Bank Still Said "No." Rebuilding Trust in a Nordic BNPL Partnership.
By turning a “high-risk” cohort into a quantified edge case, partner trust was restored—shifting from restriction to flexible lending (<0.5% losses).
The bank's risk team had flagged a serious concern: users who were declined and later approved through a recovery path might represent a hidden risk pool. But the real problem wasn't the data — the data was fine.
The problem was that nobody trusted the team presenting it. Decisions were stalling, policy was hardening, and the BNPL team had already lost credibility with the partner.
BNPL
Team
data
Bank
decision
Policy tightening
Trust gap — evidence not crossing
The Nordic BNPL team had a trust deficit with their banking partner. The bank's appetite for risk was low — and when concerns about the recovery cohort surfaced, they weren't looking for analysis. They were looking for a reason to restrict policy further.
Three things had broken down before the analysis even started:
A credibility gap.
The BNPL team had lost the bank's trust before the analysis even started. The data existed — but no one believed the team presenting it.
Two languages, no translator.
Data existed in one language, banking decisions required another. No one was bridging the gap between what the numbers said and what the bank needed to hear.
A stalling relationship.
Blocked decisions, hardening policy, a partner growing more defensive by the week — and no clear path to reset the conversation.
Strong analysis alone wasn't going to fix this. The bank didn't need more data. They needed someone who could make the data land.
I stepped into the partner meetings as the translation layer — between the data team, the BNPL team, and the banking partner.
Define the cohort precisely.
Identified the exact population of users who had been declined and subsequently approved through the recovery path. No ambiguity about who was being evaluated.
Calculate realized loss, not modeled.
Pulled actual performance data — not projections. Compared recovery cohort losses against the total portfolio loss base and assessed volume materiality.
Answer the bank's actual question.
Structured the findings around the one thing the bank cared about: Is this a systemic risk or a manageable exception? Walked through each figure in terms of their exposure, not ours.
Present without jargon.
No model outputs without explanation. No dashboards without narrative. Just: here is the risk, here is how it compares, here is why this doesn't require a structural policy change.
Post-decline losses below 0.5% — an immaterial figure relative to total portfolio exposure
Bank reversed its position — shifted from restriction to alignment after seeing the numbers
More flexible lending terms approved for the recovery cohort — unnecessary policy tightening avoided
Reusable framework established — define cohort, calculate realized loss, assess materiality, present in business terms
The conversation shifted from combative to collaborative. What had been a stalling relationship became an ongoing partnership.
Case Studies
This is just one example.
I have many more case studies with full technical walkthroughs, data, and methodology.
Work with me