Case study · Business cards & payments
Owning the opening integration roadmap after a top-tier US card issuer acquired a fintech spend-management platform, and deciding what to integrate, in what order, and what to deliberately leave alone.
If you read nothing else
A top-tier US card issuer with a strong small-business card franchise acquired a fintech spend-management platform, buying in one move the control-first product and shipping velocity that banks have struggled to build. I owned the opening integration roadmap: the first version of which capabilities converge, in what order, and which parts of the acquired platform get deliberately protected from integration. The core judgment was recognizing that the default corporate playbook, absorb everything onto the bank's stack, would methodically destroy the asset it was applied to. The roadmap's job was to capture the deal without breaking the thing the deal was for.
There is a divide running through business payments that everyone in the industry can recite. Banks lead with security, scale, and balance sheet: they are where the deposits, the credit, and the trust live. The fintech spend-management platforms lead with control: the ability for a business to shape a card's rules before the spend happens, wrapped in software that ships weekly instead of yearly. Businesses increasingly want both, and neither side has convincingly built the other's half. Buying it is the rational move, and my employer made it. Then comes the morning after, which is where I came in: someone has to write the first roadmap for how two organizations that won by opposite philosophies become one product family without one of them flattening the other.
The history of banks acquiring fintechs is mostly a history of expensive funerals. The pattern is well known: the acquirer pays for velocity and product, then integrates the acquisition into its own stack, its own release process, its own risk cadence, its own brand, and eighteen months later the thing it bought ships at the speed of the thing that bought it. Nobody decides to do this. It happens one reasonable-sounding convergence decision at a time, because every individual integration step looks like synergy and the cost, the slow suffocation of the acquired product, never shows up on any one decision's business case.
The opposite failure is quieter and just as expensive: leave the platform fully standalone to protect it, and the deal never earns its price. The customers stay in two disconnected worlds, the bank's balance sheet never reaches the platform's users, and five years later the acquisition is a line item nobody can explain. An opening integration roadmap is really a stance on this dilemma, and whoever writes it is deciding, capability by capability, which risk to take where.
Every convergence decision looks like synergy on its own business case. The cost, the slow suffocation of what you bought, never shows up on any single one.
The opening roadmap as a sorting exercise: converge the plumbing where the customer wins, and put an explicit fence around the things the deal was actually for.
The bet was to make the roadmap a sorting exercise rather than a schedule. Converge the plumbing where the customer clearly wins: one identity for a business instead of two, the bank's balance sheet and credit reaching the platform's users, shared money-movement rails, one risk and compliance floor underneath everything. And put an explicit fence around what the deal was actually for: the platform's release velocity, its control-first product philosophy, and the team that produces both. Not "integrate later," which is how absorption starts, but named, on the roadmap, as deliberately protected, so that unwinding the fence would require a decision as visible as the one that built it.
What I chose not to do was run the default playbook in either direction. Not the full absorption sequence, replatform, rebrand, align the release process, which is the well-trodden path to buying a fast company and owning a slow one. And not the trophy-on-the-shelf model of full separation, which protects the product by ensuring the acquisition never compounds. The senior part of the call was smaller and harder than either: accepting that the two failure modes are asymmetric in visibility. Nobody gets fired for integrating; the costs arrive slowly and unattributed. So the fence had to be built into the roadmap explicitly, because everything about large-organization gravity would otherwise pull one column into the other a quarter at a time.
Absorption never announces itself. It arrives one reasonable convergence decision at a time, which is why the things you will not integrate have to be written down.
An opening roadmap for an integration is less a document than a negotiation with everyone who will have to live inside it. The bank side read the fence as leaving synergy on the table, and some of them were measured on that synergy. The platform side read every convergence item as the first domino of absorption, and they had watched this movie at other companies. Both suspicions were reasonable, which is what made the room hard: this was not a misunderstanding to clear up, it was two accurate readings of two real risks.
What moved it was refusing to argue the philosophy and arguing the sequence instead. The first conversions on the roadmap were plumbing wins that both sides wanted and neither felt threatened by, the kind where the customer visibly gets something, the synergy line visibly moves, and nothing about how the platform ships changes. Each one bought credibility for the harder conversation behind it, and the fence, stated in writing from day one, gave the platform side a reason to engage with convergence instead of trench-fighting every item. My contribution was less any single decision than the shape of the document: principles first, sorting logic second, dates last, so the reasoning could hold the line after the roadmap changed hands, which, in an integration measured in years, it inevitably would.
I owned this roadmap for its opening stretch before handing it on, and the integration it describes is still being written, so the honest result is narrower than a finished number: the first convergence moves shipped without slowing the platform's cadence, and the fence survived its first quarters and its first handoff. Whether the whole bet pays off will be visible years from now in a simple test: does the acquired platform still ship like itself inside the bank that bought it. Attribution, honestly: the strategy that led to the acquisition preceded me, the teams on both sides did the building, and my part was the opening sort and the argument for the fence.
What I would do differently
I drafted the first version of the roadmap from the acquirer's side of the table, from strategy documents and capability maps, and then socialized it with the platform team. I would reverse that order. The acquired team's own roadmap, the things they were already planning to build before anyone bought them, turned out to be the best single input into what should be protected, and I read it later than I should have. You learn what you bought by studying what it was about to become, not just what it is.
The deal buys the company. The roadmap decides whether, two years later, you still own the thing you paid for or just its logo.
This problem is about to become the defining one in business payments. The divide that drove this acquisition, banks with the balance sheet, fintechs with the control-first software, is collapsing from both directions: every serious issuer is now buying or building spend management, and every serious spend platform is reaching for credit and deposits. Which means the industry's next few years will be decided less by who does the deals than by who can integrate one without breaking it. That is a product discipline, not a corporate-development one, and it is a muscle I got to build early.