PANA · GROWTH & ONBOARDING

Is the onboarding funnel healthy against the rest of fintech?

Signup → KYC → Bank Approval, benchmarked stage by stage against published fintech and neobank onboarding data. Funding/activation is out of scope for this pass.

Scope Pana Global only — US is a separate, discontinued product Window 2026 YTD (Jan–Sep 8); trend vs. H2 2025 where noted Pulled Sep 8, 2026

Below industry range — once you count from where users actually start.

Correction from the previous version of this deck: "Started KYC" was measured from Persona's own inquiry-creation event — but Pana's own product analytics show the funnel really starts earlier, when a user opens the in-app Bridge onboarding wizard (71.3% of signups reach that, not 46.1%). Measured from that true starting point, KYC completion — Wizard opened → Approved — runs 41.6%, below the industry's 55–65% range. Persona's own intake-to-approval step (64.3%) looks fine in isolation, but that's not where users actually begin: ~35% of wizard opens never even become a Persona application.

Signup → Bank approved
(Global, 2026 YTD)

29.6%

2025 full year: 31.1%

-1.5pp YoY

Signup → Wizard opened
(true KYC start)

71.3%

Amplitude, Bridge onboarding wizard

No benchmark

KYC completion
Wizard opened → Approved

41.6%

Industry range 55–65%

Below range

Wizard opened → Persona
application created

64.7%

~35% never reach Persona at all

Biggest internal leak

Where the volume actually goes

Pana Global, 2026 YTD. Absolute users at each stage — the width is the true drop, not a rescaled one.

Signups

user.createdAt

108,290

100%

baseline

Wizard opened

Amplitude: Bridge onboarding wizard

77,158

71.3%

of signups

Persona application created

personainquiry record exists

49,901

46.1%

of signups

Bank approved

bankapplication.status = Approved

32,101

29.6%

of signups

Stage conversion vs. industry benchmark

Bar = Pana Global's actual conversion rate for that stage. Shaded band = the range reported across fintech/neobank onboarding benchmark sources.

Pana Global, 2026 YTD

Industry benchmark range

No comparable external metric

Signup → Wizard opened

71.3% · no external benchmark

71.3%

The true entry point to KYC — measured from Amplitude, not from Persona's backend. No public benchmark isolates this exact step, so this is unscored.

Wizard opened → Persona application created

64.7% · no external benchmark

64.7%

~35% of users who open the wizard never generate a Persona application at all — they drop mid-form, before the backend even logs an attempt. This is Pana-specific plumbing, not something any industry report isolates either, but it's the single biggest lever inside KYC.

Wizard opened → Approved (KYC completion)

41.6% · industry 55–65%

41.6%

This is the metric that maps to what the industry calls "KYC completion" — measured from where users actually start. Pana runs below the 55–65% range. (Measured only from Persona's own intake step instead, completion looks like 64.3% — in line with the range, but that's not where users actually begin.)

KYC completion, over time

Wizard opened → Approved, Pana Global. Amplitude wizard-open tracking only goes back to Jul 2025, so this compares H2 2025 to 2026 YTD — same definition both periods, different lengths.

47.8%

H2 2025

41.6%

2026 YTD

-6.2pp — already below the 55–65% industry range in H2 2025, and the gap has widened since. Not a one-quarter blip.

Sourced benchmark comparison

Stage	Definition	Pana Global 2026 YTD	Industry benchmark	Verdict	Source
Signup → Wizard opened	User opens the in-app Bridge onboarding wizard (Amplitude) — the true start of KYC	71.3%	—	No benchmark	No source isolates this exact step
Wizard opened → Persona application created	Backend creates a `personainquiry` record — the first point Persona's own system knows the user exists	64.7%	—	No benchmark	Pana-specific backend step; no vendor publishes this either
Wizard opened → Approved (KYC completion)	Identity verification completed and approved, measured from where the user actually starts	41.6%	~55–65%	Below range	Perspective AI, 2026
Persona application created → Approved	Same "completion" measured only from Persona's own intake step, not from the wizard	64.3%	~55–65%	Not the industry metric	Included for continuity with the prior version of this deck
Overall onboarding abandonment	Share who never complete onboarding, any stage	context only	~68% abandon	Context	Perspective AI (Signicat-sourced)

Definitions & data sources — Pana

Signups — user.createdAt, Pana Global (usertype = GLOBAL), signup-month cohort.
Wizard opened — first Amplitude event of type Bridge Onboarding Started or Viewed Onboarding Bridge Wizard, per user, from panaapp-16ce3.amplitude.EVENTS_396830. This is t1 in Pana's own canonical KYC event table — the true first step, ahead of anything Persona's backend tracks. Query diag-onboarding-wizard-started-monthly.
Persona application created — a personainquiry record exists for the user (BigQuery, planetscale_pana.personainquiry). This is t7 in the same canonical table — created by Pana's backend once the user has gone far enough into the wizard, not when they open it.
Approved — bankapplication.status = 'Approved'. Metabase queries funnel-3step and persona-funnel-flow (cross-checked, agree to within 0.3%).
Scope stops at approval. Funding/activation metrics are tracked separately and intentionally excluded here to keep this a pure onboarding/KYC read.
All figures pulled via BigQuery through the Metabase API, panaapp-16ce3.planetscale_pana and panaapp-16ce3.amplitude, Sep 8, 2026.
Correction, Sep 8 2026: the first version of this deck measured "Started KYC" from the Persona-application-created event (t7), because that was the earliest step the existing dashboard query (persona-funnel-flow) tracked. Claudia flagged that KYC should start counting from when Persona actually takes the user, which led to checking Pana's own canonical event table and finding the wizard-open event (t1) is the true start — 25pp earlier than what the deck first reported.

On the external benchmark figures

Compiled from secondary industry sources (fintech CX/CRO round-ups citing Signicat and neobank activation research), not one primary index — treat ranges as directional, not exact.
Definitions vary by source (e.g. "activation" sometimes means first transaction, sometimes 5+ transactions); ranges here were matched to Pana's stage definitions as closely as the source allowed.
No external source publishes a clean "signup → wizard opened" rate, so that stage is shown for internal diagnostic value only, with no scored verdict. Confirmed by a multi-pass check (own research plus independent cross-checks) against Sumsub, Alloy, Persona, Veriff, and Plaid — identity vendors define a technical "start" event (e.g. Veriff's started/7001, Plaid's IDENTITY_VERIFICATION_START_STEP), but their own funnels begin after a verification session is already created, not at account signup. That earlier handoff sits outside what any vendor or industry report measures, so it isn't a gap in this research — it's an industry-wide blind spot. It's exactly why "wizard opened → approved" is the right stage to benchmark against their "KYC completion" figures — it's Pana's closest match to what vendors mean by "started."

Sources

Perspective AI — "Fintech Customer Experience in 2026: Solving Onboarding, Trust, and Drop-Off"