The real test of embedded distribution is origination economics. Reach is the easy part.
Most banks still win customers through channels they own: branches, digital marketing, payroll relationships, referrals and direct sales. Those channels still matter, but they are highly competitive, especially when the intended customer already holds one or more established banking relationships.
There is another point at which customer acquisition can begin. A person moving countries may engage an employer or relocation company before opening a local account. A student deals with a university before arranging domestic financial services. A self-employed worker may receive income through a platform long before approaching a bank for business products. None of these starts as a banking conversation, yet each one ends up needing a bank.
Embedded finance has demonstrated that financial products can be placed inside these non-financial journeys. The strategic question for a bank is this: when does access to such a journey produce a durable customer relationship with attractive economics, rather than merely another distribution expense?
Distribution and origination are not the same
It is relatively straightforward to place an account-opening link, a card offer or a payment button inside a partner’s application. A bank can report accounts originated through the channel and measure partner-generated transactions. Neither metric by itself establishes that the bank has acquired a valuable relationship.
True origination requires the institution to know whether a customer was acquired incrementally, whether the customer has become active and whether meaningful funding and payment activity continues after the introductory offer. It also requires an understanding of who controls communication with the customer, how financial data may be used and whether the bank can offer additional products when needs develop.
Consider international students. A university may be well placed to refer a local wallet because students already rely on it for admissions, accommodation and tuition information. The bank benefits from timely access to a defined cohort. But if the student uses the account solely for a tuition payment and then switches to another bank, the arrangement may produce little lasting value. In the end, the commercial result depends on account activation and retention.
For a gig platform, the starting service may be a payout account or card linked to earned income. Again, the bank needs to establish whether the account becomes useful outside the platform. If it doesn’t, the bank has built an outsourced payments service and called it customer acquisition.
Scale can be misleading
A telecom operator with millions of subscribers, a retailer with high app usage or a large employment platform seems to offer an obvious distribution advantage. Yet a big audience doesn’t mean people want a new bank.
The customer may already have an adequate bank account. Moving a salary, direct debits and savings can be inconvenient. A financial incentive to switch may attract customers whose balances leave when the incentive expires. A familiar brand gets the offer noticed. It does nothing for primacy or profitability.
One useful example is T-Mobile MONEY in the United States. The product offers banking services through Coastal Community Bank; T-Mobile itself states that it is not a bank. It is a clean example of a model in which the telecom customer proposition and a regulated bank’s capabilities are combined. What it doesn’t show, at least from the outside, is whether it has become a real alternative to a primary bank account, or whether it makes money.
Sources T-Mobile MONEY, account terms and banking partner
Reach only counts if it turns into a real financial need, and then into everyday use. Without both, distribution scale is a poor basis for valuation.
The opportunity lies in the timing of the need
Not every non-financial business will be an equally useful origination partner. The strongest candidates are those that meet a customer at a genuine financial transition point.
A relocation provider may be involved when a professional needs a domestic account and an international transfer. An employer can introduce a salary account, although employees must retain freedom of bank choice. A university encounters students at the point when overseas funding, tuition and local spending need to be organised. A software platform serving sole proprietors can connect collections and disbursements with a business account or working-capital services.
What makes these partners valuable is timing and relevance. Any app can offer another screen to advertise on. A bank that appears when the customer has a clear operational requirement may convert more economically than one trying to displace an established relationship without a reason.
Some partners will want to keep the customer interface and use the bank only as a balance-sheet or regulated-service provider. This can be commercially sensible for the bank if volumes, economics and risk allocation justify it. It should not automatically be described as customer acquisition. Wholesale services supplied to a platform are a different business from winning the end customer’s primary banking relationship.
The agreement should follow the intended business model
Some partnerships are referral arrangements: the bank acquires and serves the customer directly. Others are embedded propositions under the partner’s brand, with the bank supplying accounts or cards. A third category consists of infrastructure and settlement services provided to a financial or non-financial platform.
These models produce different rights, obligations and returns. Referral economics may be simpler but less integrated into the customer journey. White-label or embedded arrangements may generate greater activation but require additional technology integration, servicing and controls. Infrastructure arrangements may deliver predictable volume but little direct access to the end customer’s future products.
I would urge management to ask one question at the time of approval: is this deal meant to deliver new primary accounts, deposits, payment income, credit origination, or fee income from providing regulated infrastructure? Success measures and contractual provisions will differ for each.
Pricing needs to reflect this. A partner may seek a fee for every application, an activated account, ongoing transactions or a share of economics. The bank should avoid paying acquisition economics for customers who never fund or use an account. Where the partner also expects marketing incentives or revenue sharing, the aggregate cost can exceed direct acquisition even if the apparent cost per lead looks low.
Regulatory accountability remains with the regulated institution
Partnership-led banking adds risk even where the customer experience appears simple. The bank may depend on the partner for marketing, onboarding, technology, customer communication, transaction data, complaints or parts of the operating ledger.
US banking regulators’ joint third-party risk-management guidance addresses due diligence, contracts, monitoring and termination across the life of a relationship. A further joint statement on bank arrangements for distributing deposit products highlights operational, compliance, liquidity, concentration and consumer risks, including potential confusion about deposit insurance. None of this argues against embedded banking. It does mean the bank owns whatever is done in its name.
Sources Federal Reserve, interagency third-party risk-management guidance; Federal Reserve, joint statement on bank-third-party deposit arrangements
Canada’s OSFI Guideline B-10 makes the same broad point: federally regulated financial institutions remain accountable for outsourced activities. Europe has its own regulatory architecture, including conduct, outsourcing and digital operational-resilience requirements. The precise rules differ by market, but the commercial convenience of a partnership does not remove the regulated entity’s obligations.
Sources OSFI, Guideline B-10: Third-Party Risk Management
This has direct implications for customer experience. The bank and partner need a consistent explanation of who provides the financial service, who holds customer funds, who deals with complaints and what happens if the partnership terminates. Reconciliation, record ownership, subcontractors, access rights and exit provisions sound like back-office details. They are usually where partnerships break, and when they do, the economics and the trust go together.
Data sharing introduces a separate constraint. A partner may sit on useful behavioural or commercial data. That doesn’t give the bank a free hand to use it for marketing, underwriting or relationship management. Permissions, privacy law and fair-lending considerations must be built into the service design. The model should work on its commercial merits without assuming unrestricted data pooling.
Proving incremental economic value
A credible partnership business case should compare customers acquired through the ecosystem with a similar cohort acquired directly. It should look beyond applications and account openings to funded-account activation, salary and payout penetration, retained balances, payment volume, risk outcomes, cost to serve and attrition.
The comparison must also take into account selection effects. A university referral programme may deliver high activation because the students have an immediate need, but the same cohort might have opened accounts elsewhere without the programme. The incremental benefit is what matters. So does persistence: if switching keeps getting easier, a customer acquired through a partner can leave as easily as they arrived, and the comparison should track how many are still primary customers after two or three years. For a multi-product bank, the analysis should also follow customers into appropriate lending, investments or business accounts rather than ending at the initial channel’s P&L.
Give it time, by all means, but keep the assumptions tight. An embedded proposition can be commercially successful as a transaction or infrastructure business even where it does not lead to wider banking relationships. The measurement should reflect the actual objective, rather than retrofitting a lifetime-value explanation to weak early results.
The next distribution advantage will be selective
I do not expect banks to replace their owned channels with non-financial ecosystems. Nor do I expect every large digital platform to become a meaningful banking distributor. The stronger opportunity is selective: partner where a customer has an identifiable financial need, the bank can deliver a clearly better service, and the economics of acquisition or servicing are superior to credible alternatives.
A good partnership does not have to involve common ownership. It requires complementary assets and a sustainable sharing of costs, risks and value. The bank brings regulated financial capabilities and product depth; the partner brings customer access at a relevant moment. Whether the arrangement produces primacy, infrastructure revenue or simply referrals should be clear from the outset.
This is the distinction I would preserve between embedded finance as product placement and ecosystem banking as a customer strategy. The latter is harder to execute and is more valuable when it starts from real customer needs rather than distribution scale alone.
The final article in the series looks at one of the most visible historical examples of financial services built around non-bank distribution: mobile money. Its record in Africa is substantial, but developed-market experience suggests that the economics cannot be transplanted without examining the market conditions.
Explore Comviva’s digital wallet and payments platform, mobiquity® Pay, built to support digital wallets, mobile money and payment experiences across markets.
Continue reading: Why Mobile Money’s Next Chapter Will Look Different
