At N26, our customer onboarding completion rate has been declining. What would you do?

Question

Accepted Answer

Confirm the objective. Before I dive in, let me confirm I understand the objective. You've mentioned that N26's customer onboarding completion rate is declining, so customers who start the process are not finishing it. My goal is to diagnose where and why the drop-off is happening, identify root causes, and propose solutions that are feasible, creative, and cover both short and long term. Is that the right framing, and are we focused on retail customers or business customers?. The interviewer says, Yes, that's correct. Retail customers. You can proceed. Clarifying questions, five batches. Spend the most time here. Extract the funnel, the segmentation, and the data before structuring anything.. Batch 1, scale and timing. What is the current onboarding completion rate and what was it before the decline? And is this a sudden drop or a gradual trend over months?. The interviewer says, Dropped from 72% to 54% over 4 months. Gradual decline.. Batch 2, the funnel. Could you walk me through each step end to end, for example app download, registration, KYC, document upload, activation? And do you have data on which step has the biggest drop-off?. The interviewer says, Funnel: App Download, Registration, KYC and identity verification, Document Upload, Account Activated. The biggest drop-off is at KYC, about 35% of users who reach it do not complete it.. Batch 3, segmentation. One, is the KYC drop-off uniform across all geographies or concentrated in specific markets? Two, is there a difference between mobile and desktop users? Three, has the KYC process itself changed in the last 4 months, a new vendor, new document requirements, or a regulatory update?. The interviewer says, Higher drop-off in newer markets, Eastern Europe and LatAm. Mobile and desktop similar. A new KYC vendor was introduced 5 months ago for these markets, with stricter document requirements.. Batch 4, qualitative signal. Do we have data on why users abandon KYC, support tickets, exit surveys, app reviews? What is the average time users spend on KYC before abandoning? And how many of those who abandon come back within 7 days?. The interviewer says, Support tickets show users don't know which documents are accepted and are submitting the wrong types. Average time before abandonment: 4 minutes. Only 15% return within 7 days.. Batch 5, impact and constraints. One, how many users start onboarding each month, and what would recovering to 72% mean in additional activated users? Two, are there regulatory constraints on simplifying KYC, or is the UX flexible? Three, is there an existing squad working on onboarding?. The interviewer says, 50,000 users start monthly. Restoring to 72% is about 9,000 additional activations a month. We cannot remove verification steps but the UX is flexible. The existing onboarding squad is currently occupied. Structure the problem, MECE. Let me lay out my structure. Onboarding is a process, so I'll use a value-chain split. Bucket A, Pre-KYC: acquisition quality and registration friction. Bucket B, the KYC process itself, the step with 35% abandonment, my primary suspect. Bucket C, Post-KYC recovery: what happens to users who abandon. I'll spend most time on Bucket B since the data points there strongly. A. Pre-KYC (low risk) Acquisition quality, are the right users entering the funnel? Registration friction, steps, load time, UX B. KYC process (primary suspect) Document guidance, do users know what is accepted? Vendor performance, rejection UX, processing time Error messaging, are rejections clearly explained? C. Post-KYC recovery (amplifier) Re-engagement, push and email after abandonment Save-and-resume, can users pick up where they left off? Analysis and root cause. Bucket A, Pre-KYC: 50,000 users a month are still entering the funnel, so acquisition volume has not dropped and registration is not the issue. Low risk, moving on.. Bucket B, KYC: this is where the problem lives. 35% abandonment, and users spend 4 minutes before quitting, so they are trying, not bouncing. Support tickets confirm they are submitting the wrong document types. The new vendor arrived 5 months ago, the decline started 4 months ago, and that one-month lag is the smoking gun: the vendor brought stricter rules that were never communicated to users in the product.. Bucket C, recovery: 85% of abandoners never return, and there is almost no recovery mechanism today. This amplifies Bucket B, every abandoned user is permanently lost. Solutions, feasible and creative. Bucket B is the priority, so I sequence the fixes from cheapest and fastest outward.. Immediate, 0 to 2 weeks, near-zero cost: country-specific document guide. Add a dynamic checklist at KYC entry showing exactly which documents are accepted per country, with inline examples. No backend work, a content and UI change only.. Short-term, 2 to 8 weeks, low cost: KYC recovery flow plus save-and-resume. Trigger a push and email one hour after abandonment, "You're almost there, here's exactly what you need," and let users resume instead of restarting. Recovers an estimated 20% of the 9,000 lost users a month.. Long-term, 2 to 6 months, medium cost: vendor SLA plus alternative verification. Renegotiate the vendor SLA to include clear rejection-messaging standards, and pilot video-selfie verification for lower-risk markets as an alternative to document upload.. One creative solution: a proactive KYC assist trigger. If a user is on the KYC screen for more than 90 seconds without completing, surface "Having trouble? Our team can help right now." It prevents abandonment before it happens, needs no regulatory change, uses existing support infrastructure, and can be A/B tested in two weeks. Measure success. I'd measure in layers: Leading indicator, week 2: wrong-document submission rate. Target down 30%. Tells you the guidance fix is working before the completion rate moves. Primary KPI, month 3: KYC completion rate. 54% back to 68% in 90 days, about 7,000 additional activations a month. Long-term health, month 6: new-market onboarding parity. Eastern Europe and LatAm within 5 points of established markets. Guardrail, ongoing: KYC fraud rate. Must not increase, this confirms we have not compromised regulatory integrity. Bottom line. N26's onboarding decline is not a volume problem, 50,000 users still enter monthly. It is process friction at KYC, driven by poor document guidance and a new vendor whose rejection UX causes 85% permanent loss. Add country-specific document guidance this week at near-zero cost, build a recovery flow in month one, renegotiate the vendor SLA by month six. This recovers about 7,000 activations a month and builds a scalable onboarding model for every new market N26 enters.

Root cause	Evidence	Priority
Document guidance gap in the UI	Support tickets, 4-minute abandonment, wrong uploads	Highest
KYC vendor UX friction	New vendor 5 months ago, stricter rules, poor rejection messaging	Medium
No recovery mechanism	85% permanent abandonment, no re-engagement	Medium