Choosing the right stack: core, digital and lending platform selection done properly

A bank's technology destiny is set in a handful of selection decisions, each made in a short window under vendor-managed conditions. Here are the four traps we see constantly — and the method that avoids them.

The most expensive decisions are made earliest

A bank's technology destiny is largely set in a handful of selection decisions: the core platform, the digital channel stack, the lending origination system, the payments infrastructure. Each is a long commitment made in a short window — typically under pressure, with imperfect information, while vendors put their very best people in the room for exactly as long as the evaluation lasts.

We spend a meaningful share of our advisory work helping banks make these choices well. The traps we see are remarkably consistent.

The four traps

1. Demo theater

Standard vendor demos are rehearsed performances on curated data. They prove the platform can do what the vendor chose to show — not what your bank needs daily. The fix is simple and rarely applied: write your own scenarios, drawn from your real products and your most awkward edge cases, and require every vendor to demonstrate against them. Watch where the presenter hesitates; that is where the customization budget lives.

2. The TCO illusion

License fees are the visible tip. Implementation, integration, data migration, environment costs, upgrade cycles, specialist staffing and exit costs are the iceberg. A credible comparison models five years minimum, with assumptions on the table — and includes the cost of the integration landscape around the platform, which routinely exceeds the platform itself.

3. Checklist scoring

Five hundred functional requirements scored yes/no produce a tie. Every mature platform "supports lending." The real differentiators are weighted, scenario-based and architectural: how it supports your lending, how it behaves at your volumes, how cleanly it exposes APIs, how upgrades land, and how strong the local partner ecosystem is when something breaks at quarter end.

4. Selecting the core in isolation

The core, the digital channels and the lending stack are one decision wearing three costumes. A core selected without an integration architecture becomes the bottleneck for every channel decision after it. We insist on selecting against a target architecture — API-first, event-capable, cloud-ready — so each component is chosen for how well it plays with the others, not how it demos alone.

"Vendors sell platforms. Banks live with ecosystems. Select the ecosystem."

A method that holds up

Our selection engagements follow a disciplined arc. Capability mapping first: the bank's products, processes, volumes and growth ambitions translated into weighted requirements — so the scorecard reflects this institution, not a template. Then structured evaluation: client-scripted demos, architecture deep-dives with the vendor's engineers rather than its sales team, and reference calls with banks of similar size and market. In parallel, a five-year TCO model and an integration impact assessment. Finally, negotiation support and a board-ready recommendation with the reasoning — not just the verdict — documented.

Run properly, the whole arc fits in eight to twelve weeks. That is fast enough to hold executive attention and thorough enough to withstand the question every board should ask: "how would we defend this choice in three years?"

Independence is the product

One thing matters above all in selection advisory: the advisor must have nothing to gain from the outcome. A selection report that conveniently lands on the advisor's own implementation practice is not advice; it is marketing with a methodology section. We hold ourselves to documented reasoning, declared interests and recommendations we are prepared to defend in front of any board — years later.