Banking and finance digital marketing for accounts where CAC and LTV decide the next campaign.

The operative reference is still MAS FEAT Principles (Fairness, Ethics, Accountability, Transparency in the use of AI and Data Analytics in the Financial Sector, published November 2018). FEAT is principles-based, not rules-based, and the industry reads it to cover AI tools that drive decisions affecting customers, including AI-personalised marketing, AI-generated creative, and algorithmic targeting on paid platforms.

To operationalise FEAT, MAS published the Veritas Toolkit V2.0 in June 2022 with assessment methodologies. CMOs at MAS-regulated institutions are expected to evidence FEAT-aligned governance over the AI surfaces inside their paid stack, including Google Performance Max and Meta Advantage+. MAS Notice 626 updates effective 30 June 2026 tighten source-of-funds verification on digital channels, which materially affects credit-card and wealth-account acquisition flows.

The compliance pack a CMO should put in front of the CRO before launching an AI-personalised campaign: FEAT mapping per AI tool, Veritas-aligned testing methodology, model-validation log, third-party-vendor disclosure (Meta, Google, programmatic DSPs), and audit trail of any AI-generated creative.

Three different rules, three different mitigations.

Meta Special Ad Category. Credit ads opt in automatically; lookalike audiences are disabled, age and gender targeting are restricted, US ZIP-code targeting is removed, and detailed-interest targeting facets that enable proxy discrimination are blocked. Advantage+ Shopping and Advantage+ Lead campaigns auto-detect SAC eligibility; advertisers cannot opt out where Meta determines the ad is credit-related. The work moves to creative variety, first-party retargeting, and contextual placement.

ASIC RG 234 in Australia. Headline claims must not mislead by omission; risk disclosure must be given equal prominence to benefits. ASIC enforcement in 2024-25 has tested short-form social video and ruled that the format itself cannot strip prominent risk disclosure. ASIC DDO (in force October 2021) plus Target Market Determination review must run on every credit campaign; programmatic and lookalike expansion that pulls a credit product to a non-target consumer creates direct DDO exposure.

NAIC Model Bulletin. 20+ US states have adopted, including New York DFS Circular Letter No. 7 (July 2024). Insurers in adopting states must maintain a written AI Systems Program covering AI in rating, underwriting, claims, fraud detection, and marketing. Documentation must include third-party AI tools (Google Performance Max, Meta Advantage+, AI-driven CDPs) and the program must be examinable.

The EU AI Act (OJ 12 July 2024) phases in across 2025-2028. Prohibited practices and AI literacy obligations from 2 February 2025. Governance rules and General-Purpose AI obligations from 2 August 2025. Most provisions from 2 August 2026. High-risk-systems rules to 2027-2028.

Annex III explicitly classifies credit scoring of natural persons (point 5(b)) and risk-pricing in life and health insurance (5(c)) as high-risk. Article 2 extends application extraterritorially to providers and deployers whose AI output is used in the EU, the test that catches non-EU banks serving EU customers via cross-border digital channels.

The practical implication for a non-EU bank: if your AI tools score EU credit applicants, you fall inside the high-risk obligations from 2 August 2026 (risk management, data governance, human oversight, accuracy, technical resilience, post-market monitoring). A 90-day pre-applicability sprint to map AI tools, document data flows, and stand up the governance structure is the right shape of work.

The McKinsey Global Banking Annual Review, November 2024 publishes indicative payback windows for net-new retail customers: checking 18 to 30 months, high-yield savings 9 to 15 months when paired with cross-sell, credit cards 12 to 24 months on prime, wealth and investment accounts 24 to 48 months. Cornerstone Advisors, February 2025, places cost-per-funded-checking-account at US$200 to US$300 for US community banks and US$280 to US$425 for US digital-first banks.

Channel mix that compresses payback. Branded search and intent capture on Google for the active-buyer segment. Life-event prospecting on Meta with Conversions API live, Event Match Quality (EMQ) at 7.0 or higher, bound-account events fed back via offline conversion import. Meta Special Ad Category compliance baked in for credit cover. Programmatic CTV for the brand-trust signal a regulated category needs; premium PMP, not open-exchange. LinkedIn Account-Based Marketing (ABM) for commercial banking, paired with Google Search for active commercial intent.

The payback window contracts most sharply when the post-click application-completion funnel is fixed; channel reallocation cannot save a programme losing 85 percent of applications between submit and fund.

The two channels do different work. LinkedIn B2B Institute's "The Long and the Short of It" 2024 update reaffirmed the 95-5 rule: 95 percent of in-market firms are not actively buying at any moment; brand investment in the out-of-market 95 dominates long-run share for high-consideration banking products including treasury services, capital markets, and wealth management.

McKinsey B2B Pulse, March 2025 reports the average decision cycle for an SME or mid-market commercial-banking primary-relationship switch at 6.4 months, with 73 percent of decisions involving a buying committee of 3 or more. LinkedIn and peer referrals together account for 64 percent of first-touch attribution in commercial-banking deals over US$10 million in annual revenue.

IAB April 2024 reports CTV ad spend reached US$28.7 billion in 2024, up 13.7 percent year-on-year, with financial-services share up 23 percent year-on-year. CTV earns category brand-trust above the line. A bank that swaps LinkedIn for CTV on cost-per-lead optimisation alone tends to lose senior-decision-maker share within 18 months, because LinkedIn was the channel earning the consideration set, not just the click.

Both. The audit decides the allocation against your specific cohort.

Bain & Co., Customer Behavior and Loyalty in Retail Banking, October 2024 shows digital-only banks now win 47 percent of net-new primary-checking relationships among consumers under 35 in the US, but the growth rate of that win-share decelerated in 2024 versus 2023 as incumbent banks shipped credible digital onboarding.

The differentiator that still wins is application-completion friction, not channel novelty. CMOs at incumbent banks who fund yet another digital-first-style paid-acquisition push without first fixing the post-click funnel see flat marginal funded-account-per-dollar. The advice is inverted from the 2022-23 narrative.

The audit reads post-click drop-off in the application funnel and reports the marginal cost of every percentage-point improvement in completion rate. Programmes that hit 60 percent application-to-funded conversion are scaling profitably; programmes at 30 percent are subsidising a leak.

Three layers, all running simultaneously.

1. In-platform attribution per channel as the always-on baseline; understood to overstate by 30 to 50 percent versus incremental lift.
2. Server-side measurement. Conversions API on Meta, Enhanced Conversions for Leads on Google, Events API on TikTok, Conversions API on LinkedIn, UET plus Offline Conversions Import on Microsoft. Bound-account event fed back via offline conversion, not application-submit.
3. Marketing-mix-modelling refreshed quarterly on 24 months of weekly spend and outcome data; the only honest cross-channel read. Pair with quarterly Conversion Lift Studies on the channels with volume.

For banking specifically the highest-leverage adjustment is the funded-account feedback loop. Application volume trains the algorithm on the wrong unit; bound and funded train it on the right one.

Payback is gross-margin divided by fully-loaded customer-acquisition cost, expressed in months. Fully-loaded includes media plus platform fees plus tooling plus sales-team allocated cost on RM-led accounts plus creative production.

Retail. Checking 18 to 30 months (McKinsey Nov 2024), high-yield savings 9 to 15 months when paired with cross-sell, credit cards 12 to 24 months on prime, mortgages 6 to 14 months on origination fees alone but materially negative when discounting for primary-relationship acquisition, wealth and investment accounts 24 to 48 months.

B2B / commercial banking. 18 to 36 months on primary-banking relationship value, with the longer end on treasury and capital-markets accounts. LinkedIn-sourced commercial accounts typically run higher CAC and higher persistency than search-sourced direct policies.

Quarterly cohort analysis (acquisition cost by month, revenue realised by month, persistence to year 2) is the format your CFO will recognise.

For retail banking on a primary-checking acquisition focus, a 60:40 performance-to-brand split is the conventional starting point. For wealth and SME commercial banking the split inverts toward brand 60, performance 40, because of longer consideration cycles and the 95-5 rule. The right answer depends on the specific cohort.

Brand earns the next dollar when category awareness is structurally low (Brand Lift Studies show the gap), when trust signals are absent or buried, or when financial-strength credentials are not surfaced in branded search. Performance earns the next dollar when intent capture has plateaued (search auction saturated, retargeting frequency peaked) and the application-completion funnel is clean.

Edelman Trust Barometer 2025 reports financial services trust at 56 percent globally, down two points year-on-year, with banking specifically at 60 percent. A trust deficit is a brand problem, not a performance problem.

The JD Power 2024 US Retail Banking Satisfaction Study (May 2024) reports a 138-point gap (on a 1000-point scale) between primary-bank customers who receive proactive personalised guidance and those who do not. Loyalty correlates more strongly with proactive financial guidance than with rates or fees.

The implication for marketing: the highest-LTV cohort is built post-acquisition through guidance content, structured cross-sell, and life-event-triggered relationship deepening. The campaigns that drive loyalty in the JD Power data are not the ones that win first-account acquisition. Treat post-acquisition marketing as its own programme with its own budget line, its own creative, and its own incrementality testing.

Programmes that spend everything on acquisition and treat onboarding as customer-service overhead leave the 138-point gap unclosed.

Three controls running in parallel.

Asset-level governance. AI-generated creative passes through the same regulator-aware review as human-built creative. ASIC RG 234 equal-prominence rule applies. FTC Endorsement Guides revised June 2023 require clear and conspicuous disclosure of material connections including AI synthesis. MAS, FCA, and BNM apply existing fair-dealing and misleading-conduct rules to AI-generated creative without separate AI-specific guidance as of January 2026.

Tool-level governance. Every AI tool in the stack documented: model name, vendor, data inputs, prompt library, output review steps. NAIC bulletin and OSFI E-23 both require third-party-AI vendor documentation for material business uses.

Output-level review. No AI-generated creative ships without named human approval logged in the workflow. FINRA Notice 24-09, 27 June 2024 made this explicit for member firms; the same supervisory bar applies in practice across all five anchor markets.

Three triggers.

1. The platform stack has shifted under you. AI Max migration ahead of September 2026, Advantage+ defaults, Smart+, Microsoft PMax transparency, plus Special Ad Category constraints tightening on credit cover. A team running the 2024 playbook widens the gap every quarter.
2. The compliance frame has changed. MAS FEAT operationalisation through Veritas. APRA CPS 230 in force from 1 July 2025. OSFI E-23 in force from 1 July 2027 with phased build now. NAIC Model Bulletin adoption progressing. EU AI Act phasing in. Teams that have not refreshed their pipeline are shipping non-compliant assets into the new environment.
3. The board is asking for a causal read. If nobody has run Conversion Lift Studies, geo experiments, or marketing-mix-modelling refreshes, there is no evidence the spend is incremental. Reported ROAS is overstated by 30 to 50 percent against an incremental-lift study.

The audit reads which of the three is actually breaking the programme. Findings document yours regardless of next steps.

Senior practitioners run paid media across the five anchor markets (Singapore, Malaysia, Australia, US, Canada) with a banking-specific compliance pipeline. Diagnostic audit reads the programme end to end; build sprint restructures the surfaces flagged in the audit; managed subscription runs the day-to-day with weekly senior review.

Channels we run for banks: Google Search + Performance Max + Demand Gen, Meta Advantage+ Shopping and Advantage+ Lead, LinkedIn ABM and Thought Leader / Document Ads, programmatic CTV via DV360 and Trade Desk, Microsoft Performance Max with Copilot diagnostics, YouTube Brand Lift Studies for awareness, TikTok Smart+ where the buyer skew fits. Measurement stack: Conversions API + offline conversion import on every platform, server-side GTM, Enhanced Conversions for Leads, Consent Mode v2.

A senior practitioner is named on every engagement. For most accounts that means one of the four leadership team members below sits inside the account from the first conversation:

• Siddharth Surana, Founder and CEO, 18+ years. Ex-Regional CDO Havas, ex-COO Media360, Programmatic Pioneer APAC 2011.
• Sundeep Surana, Managing Director, 16+ years.
• Ratnakar Nemani, Operations Director, 11+ years, Google Ads Certified.
• Nitesh Sanghvi, Search and Social Director, 11+ years, Google Ads & Google Analytics certified.

50+ combined years across the leadership team. No account-manager handoff after the pitch.

Yes, on a structured-handover pattern. Three phases.

1. Discovery and audit (weeks 1 to 3): we read the existing account end to end, document the structure, identify lift opportunities and compliance risks (especially MAS FEAT mapping, ASIC DDO exposure, NAIC bulletin readiness where applicable).
2. Handover (weeks 3 to 5): account access, asset library, tracking setup, underwriter / compliance-officer pipeline, reporting handoff. We work with the outgoing agency on a professional handover where possible.
3. First sprint (weeks 5 to 12): highest-priority fixes shipped, measurement baseline locked, first Conversion Lift Study scheduled.

We do not poach accounts mid-contract or pitch in competition with an active incumbent.

Eight things 90 days before launch.

1. Conversion tracking dual-tagged across browser-side and server-side on every platform with Event Match Quality (EMQ) 7.0 or higher; bound-account event firing via offline conversion import.
2. Regulator compliance review of every ad asset against the applicable frame (MAS FEAT, FAA-N16, FAA-N03; ASIC RG 234, DDO, TMD; CFPB Circular 2022-03; FINRA Notice 24-09; NAIC bulletin where state-applicable).
3. CRM stage-progression mapped to platform conversion events; bound-account event fed back via offline conversions.
4. Application-funnel UX audit with form-friction analysis at every step; identity verification and KYC paths under five minutes mobile-first.
5. Catalog feed quality where commerce applies (credit cards, term deposits, motor loans, mortgages).
6. Lead-quality feedback loop from CRM back to platforms on fund and on activation, not on application.
7. Cross-platform attribution model with cost-per-funded-account as the chosen north-star.
8. FEAT / NAIC / EU AI Act readiness pack for the compliance officer; AI tools documented, prompts logged, vendor disclosures in place.

Most launches skip 4 of these 8 and pay for it for two quarters.

Banded by engagement type rather than percentage of media spend.

• Diagnostic audit: 2 to 3 week fixed-scope read, findings document yours regardless.
• Build sprint: 6 to 8 week fixed-scope restructure on a specific surface (signal quality rebuild, compliance pipeline build, channel-mix rebalance, post-click funnel fix).
• Managed subscription: ongoing day-to-day management with a senior practitioner named on the account.
• Embedded retainer: strategic and technical direction with weekly cadence.

We do not mark up tool subscriptions or platform media spend. International engagements billed in equivalent currency on every invoice.

Three signals.

• Regulated-sector operating experience plus MAS FEAT, FAA-N16, FAA-N03 fluency. Most agencies have run financial accounts; few have run them through the post-FEAT and Veritas Toolkit governance expectations.
• Technical depth on the 2026 measurement stack (Conversions API, server-side GTM, Enhanced Conversions for Leads, Consent Mode v2, EMQ tuning, AI Max migration planning, offline conversion import keyed to bound-account events).
• Senior practitioner involvement on the account, not an account-manager funnel.

leapbuzz leadership: Siddharth Surana (Founder/CEO, 18+ yrs, ex-Regional CDO Havas), Sundeep Surana (MD, 16+ yrs), Ratnakar Nemani (Ops Director, 11+ yrs, Google Ads Certified), Nitesh Sanghvi (Search and Social Director, 12+ yrs, Google Ads & Google Analytics certified). 50+ combined years across leadership.

Travel Guard Singapore is our named live client; engagement covers paid media across search and social channels.