Cost-to-income math: where 2pp actually comes from.

A CEE tier-1 bank's cost-to-income ratio improved from 54% to 52% in 18 months without redundancies. Anonymous waterfall analysis: where every basis point came from. Why the C-to-I story is not about cutting headcount but about redeploying it.

The CFO question

European universal banks in 2026 are operating at an average cost-to-income ratio of roughly 54%. The challenger neobanks operating in the same markets are at 35–40%. The board's strategic plan for 2027 typically targets ≤ 50%.

That sentence — "target ≤ 50%" — gets written in board decks every year. What does not get written is where the four percentage points actually come from. The honest answer to that question is what separates an executable cost programme from a strategic-plan-as-aspiration. The dishonest answer is "digitisation savings" with no line-item breakdown. CFOs see this every year.

This essay is the line-item breakdown for one anonymous tier-1 CEE universal bank that moved from 54% to 52% cost-to-income in 18 months without redundancies. Two percentage points — 50% of the way to the strategic target — from a single Wave-1 and partial Wave-2 deployment.

The bank's profile, for context:

• 6.8M retail customers across 6 CEE markets
• €140B total assets
• €1.9B net interest income + fees / year
• 54% cost-to-income at start of programme (Q1 2024)
• 52% cost-to-income 18 months later (Q3 2025)

The −2pp translates to approximately €38M / year of recurring savings on the same revenue base. The programme cost: approximately €4.2M one-time + €1.8M/year recurring (Coreal SaaS + bank-side operational change). Payback period: 8 months on the recurring savings line.

This is the waterfall.

The 2pp waterfall

I'll start with the headline numbers, then break each line down.

The five lines are all on the operational side. None are interest expense, none are funding costs, none are capital structure. That matters because operational savings flow straight to the C-to-I numerator without rate sensitivity.

Below I take each line and walk through where it came from.

Line 1 — KYC labour reduction: −€14.4M / year

The bank's KYC operation pre-Wave-1 employed approximately 240 FTE across the 6 markets: document reviewers, manual liveness reviewers, case-management specialists, AML specialists handling step-up cases, and operations team handling exceptions.

Wave-1 (covered in the 90-day field note) replaced the legacy KYC pipeline with Coreal's orchestrator. Auto-approve rate moved from 62% to 80% on retail customer onboardings. The manual review queue dropped by approximately 47% (combination of higher auto-approve + lower false-positive escalation).

The labour math:

• Pre-Wave-1: 240 FTE, average cost (loaded) €110k/year per FTE = €26.4M/year
• Post-Wave-1: 130 FTE handling residual cases + supervisor function + escalations = €14.3M/year
• Programme-cost-attributable saving = €12.0M/year
• Customer-experience and ops-quality improvements that the bank booked separately = €2.4M/year (lower drop-off in onboarding funnel → higher product conversion → more activated customers per month)

Crucially, none of the 110 FTE were made redundant. The bank redeployed them. Specifically:

• 35 FTE moved to the bank's growing financial-crime investigation team (where senior judgement is increasingly needed)
• 28 FTE moved to onboarding-experience and customer-success roles (the bank had been understaffing this for years)
• 22 FTE moved to AML programme leadership and policy roles
• 18 FTE moved into wealth-banking and SME-banking onboarding (smaller volumes but more complex cases)
• 7 FTE retired or moved on by attrition over the 18 months

The cost saving in this line came from not backfilling attrition + not opening the new roles that would have been opened to do the redeployed work, not from terminations. This is the structural point: a bank that says "modernisation will require redundancies" is making a different strategic choice than "modernisation will let us redeploy without further hiring." We have always advised banks to make the latter choice. It is operationally smoother, politically easier, and structurally superior because senior compliance and customer-success talent is the bottleneck in growth, not in cost.

Line 2 — Sanctions-screening false-positive review: −€8.1M / year

Pre-Wave-1, the bank's sanctions screening process generated approximately 14% false-positive rate per screen. Each false positive required a manual review by a sanctions specialist. The bank had 32 FTE dedicated to sanctions disposition.

Coreal's screening pipeline uses three independent sanctions providers (Refinitiv / Dow Jones / LexisNexis) behind a unified gateway, with a rule library tuned for the bank's customer profile. The false-positive rate dropped from 14% to 2%. The same screening volume now generates 86% fewer manual reviews.

The labour math:

• Pre: 32 FTE × €110k = €3.5M/year
• Post: 4 FTE × €110k = €0.44M/year (plus the gateway licensing cost)
• Programme-cost-attributable saving = €3.1M/year of labour

The remaining €5M/year of this line came from indirect savings: the cases that were false positives often required follow-up customer outreach (creating support tickets), occasionally caused customer churn (when a customer was wrongly flagged twice), and required compliance-audit attention. Eliminating 86% of false positives compounds across these downstream costs.

The 28 FTE freed up here joined the redeployed pool described in Line 1.

Line 3 — AML-evidence-pack labour: −€6.2M / year

When the bank's national FIU or a competent authority requested an AML evidence pack on a specific customer, the bank's pre-Wave-1 process required pulling records from multiple systems (T24, the legacy KYC platform, the case-management system, the document archive), reconstructing the decision trail, packaging the evidence into the regulator's format.

Average labour per evidence pack: 18 person-hours. Volume: approximately 8,400 packs / year across all six markets. Total labour: 151,200 person-hours = approximately 75 FTE-years.

Post Wave-1, the decision journal in Coreal captures the trail for every KYC decision automatically. The evidence pack for a Coreal-handled customer is generated from a single source with one query. Labour per pack: 2.4 person-hours. New volume: 20,160 person-hours = approximately 10 FTE-years.

The 65 FTE-year saving translated to:

• 45 FTE-year of direct labour cost = €5.0M/year at the bank's loaded cost
• 20 FTE-year of avoided over-time and contractor cost = €1.2M/year

Coreal-handled customer share grew from 0% (pre-Wave-1) to approximately 90% over the 18 months (Wave-1 plus Wave-2 onboarding new customers directly on Coreal). The residual 10% — legacy customers whose data still lives in the old systems — continues to consume the higher per-pack labour, but on a shrinking footprint.

This line will continue improving toward year 2–3 as the legacy-customer share of evidence-pack volume drops further.

Line 4 — Branch onboarding labour: −€5.4M / year

Pre-Wave-1, retail customer onboarding at branch took approximately 45 minutes of branch staff time per customer (document collection, system entry, identity verification, account opening, product configuration).

Post Wave-1, the customer can complete most of the flow themselves on the bank's mobile app, with the branch staff only verifying the last steps. Average branch-staff time per onboarding: 8 minutes.

Annual branch onboardings: approximately 320,000 across all six markets. Time saving: (45 − 8) × 320,000 / 60 = 197,333 hours of branch staff time. At loaded cost of €50/hour for branch staff (lower than centralised compliance specialists), that's approximately €9.9M/year of gross saving.

Net saving on cost-to-income line: −€5.4M/year. The difference (€4.5M) went into:

• Increased branch-staff time available for advisory and sales activities (bank's strategic priority)
• Reduced branch-overtime expenditure
• Slower headcount growth in the branch network (some branches did not backfill departures)

No branch closures. No branch-staff terminations. The bank's branch footprint actually grew slightly during the period (two new branches opened in growth markets) — but with fewer FTE per branch on transactional work and more on advisory.

Line 5 — Internal audit + compliance prep: −€3.8M / year

The combined cost of preparing for DORA, BCBS 239 and AMLA audits — across compliance, risk and internal audit teams — was approximately €10M/year pre-Wave-1. This included contractor and consultancy fees during the heavier audit cycles.

Wave-1 brought the unified-evidence-pack infrastructure online (covered in the BCBS 239 + DORA + AMLA essay). The bank's compliance org consolidated around shared lineage and a single decision journal.

The labour math:

• Pre: approximately 80 FTE across the three regimes, plus €1.8M/year of external consultancy
• Post: approximately 50 FTE, with consultancy reduced to €0.6M/year
• Net saving: 30 FTE × €110k = €3.3M/year + €1.2M consultancy reduction = €4.5M/year gross
• Net of investment in shared infrastructure team (4 FTE specialised, +€0.7M/year): €3.8M/year

The 30 FTE freed up included some who left the bank (attrition during a period of regulatory uncertainty), some who moved to specialist roles within the unified team, and approximately 12 who moved into broader risk and compliance positions in the bank's investment-banking arm (which had been chronically understaffed).

What the 2pp does NOT include

The €38M / 2pp number is what we measured directly. There are at least three secondary effects we believe contribute but couldn't isolate cleanly:

1. Customer-acquisition cost reduction. The bank's customer-acquisition unit economics improved during the period: better onboarding conversion meant lower CAC per activated customer. We estimate this contributed €4–8M to the marketing line, but the bank's marketing measurement methodology changed during the period, so we left it out.

2. Reduced customer churn. The bank's measured net promoter score improved by 11 points during the 18 months. Some portion of that improvement is from better onboarding; some is from other initiatives. Churn data is too noisy at the 18-month horizon to isolate.

3. Capital efficiency. Faster onboarding means faster activation, which means lower regulatory capital tied up in pending-application status. The bank's RWA-side gains are real but small (a few €M) and were absorbed into general capital planning.

If we credited even half of these, the C-to-I improvement would be closer to 2.4–2.6pp rather than 2.0pp. We stuck with 2.0pp because that's the number that's defensible to the auditor.

What the 2pp DOES include that some banks don't expect

Most of the savings are in compliance and risk operations, not in customer-facing operations. The five lines, ranked by contribution: KYC (38% of total), sanctions FPR (21%), AML evidence (16%), branch labour (14%), audit prep (10%).

The conventional narrative is that bank modernisation is about closing branches and replacing tellers. The actual savings, where they happen, are in the back-office: compliance specialists, sanctions analysts, evidence-pack assemblers, internal audit prep teams. These roles are senior, well-paid, and structurally hard to scale. Reducing the volume of work they do — without reducing their headcount — frees up their highest-leverage hours for the work that actually requires senior judgement.

A bank's CFO who is told "modernisation saves on the front line" should ask whether the programme is actually addressing the high-cost roles. If the programme's saving is mostly in clerical or branch staff, the C-to-I improvement is real but the strategic effect on the bank's risk and compliance maturity is small. If the saving is in compliance and risk operations, the C-to-I improvement is real AND the bank's senior judgement capacity has been increased.

What does NOT happen

I want to be specific about what the 2pp programme did NOT include, because banks often expect (and sometimes ask for) these and they are not part of the operational-leverage story:

No redundancies. The 200+ FTE of work that was eliminated was redeployed into growth functions or absorbed by attrition. Banks that conflate "modernisation" with "headcount reduction" are choosing a different programme. Both can deliver C-to-I improvement, but the redeployment path is more durable.

No branch closures. The bank's branch network footprint was essentially stable during the period (slightly expanded). The branch-labour saving came from per-branch productivity, not from network reduction.

No legacy core retirement. The bank's T24 still runs. The Wave-1 and Wave-2 programmes are adjacent to the core, not replacing it (see Adjacent vs replacement). The C-to-I improvement does not depend on the legacy retiring.

No new products with major revenue impact. Wave-2 shipped card-to-card transfers and FX (Wave-2 field note). These added approximately €1.4M/year in net fee revenue — meaningful but not a major driver of the C-to-I numerator. The C-to-I improvement story is overwhelmingly about cost-side operational leverage, not revenue-side product growth.

The revenue story will become bigger in Wave-3 (months 12–36), as new products on the Coreal ledger compound. But the 2pp at 18 months is mostly a cost story.

The strategic line for CFOs

A bank's strategic plan that targets ≤ 50% cost-to-income by 2027 needs an executable programme that puts numbers against the gap. "Digital transformation" is not a programme. "Modernise customer onboarding using Wave-1, redeploy 110 FTE, save €14M/year on KYC labour" is a programme.

The 2pp we measured at 18 months for this bank was the first half of a 4–5pp programme that, by year 3, gets the bank to ≤ 50%. The next 2–3pp comes from:

• Wave-3 deeper compliance consolidation (additional 0.5–0.8pp)
• New revenue products on Coreal ledger raising the C-to-I denominator (0.5–1.0pp)
• Legacy-core retirement decision in years 5–7, which becomes operationally trivial once the new-product volume is dominant (1.0–1.5pp)

None of these are aspirational. Each one has its own waterfall. The full programme is a 36-month plan with quarterly milestones, signed off by the CFO, the COO and the CTO.

The thing the CFO should ask for is not "more digital transformation". It is "show me the waterfall, signed by the COO." If the waterfall is empty, the strategic plan is aspirational. If the waterfall has named cost lines with measurable owners, the programme will deliver.

We help write the waterfall, scoped to the bank's specific core platform and operational baseline. Wave-1 is 90 days. The first 0.7–1.0pp of C-to-I improvement is visible by month 12. The full 2pp is visible by month 18.

Engagement-side disclosure: I am the founder of Coreal, which delivers the Wave-1 and Wave-2 programmes described. The C-to-I numbers are from an anonymous tier-1 CEE universal bank engagement; they are indicative for similarly-profiled banks but vary with the bank's specific cost baseline and product mix. For a waterfall scoped to your specific institution, book a working session →. For the architectural foundation, see solutions for banks.