Why Monzo's 2024 CAC doubled while its CPM fell 30%.

§01 · The headline contradictionTwo numbers in the same year that should not coexist.

Monzo Group plc's 2024 filings show blended customer-acquisition cost roughly doubled against its 2022 base.^{[1]Monzo Group plc · Annual Report 2024–25, p.74. Acquisition cost commentary; primary current account acquisition.} Over the same window, Meta and Google CPMs for UK financial services fell by approximately 30%.^{[2]SOI Digital UK Q4 2025 benchmark report; WARC Adspend Outlook 2026 (financial services vertical).} Both numbers are public. Both numbers are robust. The naive reconciliation — "iOS17 attribution killed everything" — is doing too much work.

§02 · The naive explanation, and why it's wrongAttribution loss explains some of the gap. Not all of it.

iOS17's ATT changes, finalised in late 2023, did meaningfully degrade multi-touch attribution for paid social.^{[3]Apple iOS17 ATT changes, finalised Sept 2023. See: Branch State of Mobile Attribution 2025, p.22.} If that were the full explanation, however, we would expect every UK neobank to show a similar CAC curve. They don't. Starling's curve is flatter; Revolut's is non-comparable for product-mix reasons. The variance demands a better model.

§03 · The actual mechanismFunnel saturation, not attribution loss.

The UK current-account customer who was easy to acquire was acquired between 2019 and 2022. What remains is a residual pool: lower intent, higher switching cost, less brand-sensitive. Every fintech below £1B valuation is now running 2019's playbook against 2025's pool. The CAC inflation is not a measurement artefact. It is a market-structure artefact.

Top-of-funnel inefficiency compounds at the mid-funnel: drop-off rates from prospect to verified account are up across the cohort. The naive response — increase top-of-funnel spend — accelerates the saturation. This is the trap.

§04 · What the filings show about Monzo's responseLending push, B2B expansion, US market.

Each move is valid. Each move also re-resets the CAC curve from scratch in a new product category. The 2024 filing^{[4]Monzo Group plc · Annual Report 2024–25, p.82. Lending product unit economics; US market entry commentary.} shows lending acquisition costs well below current-account CAC — but the lending pool is smaller, more credit-sensitive, and arrives with different LTV mechanics. Each successive product line is a smaller hill to climb, climbed in the right direction.

§05 · What this means below Monzo's scaleThe cohort effect.

Every UK neobank below £1B valuation has access to the same playbook Monzo used in 2019. None of them have access to the brand equity Monzo built between 2017 and 2021. The 2026 version of that playbook does not produce the 2019 result. We are not yet seeing the second-order effect: the writedown of CAC assumptions inside the boards of those neobanks. We expect it within four quarters.

§06 · Three things to do differentlyDefendable from public data. Each one points back to method.

One. Stop reporting blended CAC as a headline number to your board. Report fully-loaded CAC by acquisition cohort, segmented by intent signal. The blended number is now an obfuscation.

Two. Re-price your top-of-funnel against a 2025 pool, not a 2021 pool. The implicit assumption in most fintech growth models is that the acquirable pool is stable. It is not. Bake decay into the model.

Three. Build the brand engine that 2019 Monzo had, in the channel where your operators read. For most UK fintechs below Series C, that is now LinkedIn long-form and the operator-newsletter network, not Meta.

§07 · What we got wrongOne number we could not confirm. One source we wished we had.

We could not confirm the proportion of Monzo's 2024 acquisition spend allocated to the lending product versus the primary current account; the 2024 filing aggregates. We would have used Revolut's UK-segmented CAC, but the group reports at consolidated level only. Both gaps weaken the piece by perhaps 10%. We name them because the piece is not stronger for hiding them.