Every region runs a proper PMO, and that's the problem. It's a counterintuitive one, because each regional PMO is doing good work. The difficulty only exists at group level, which is why it can take so long for anyone to own it. The goal was never one process, it's one vocabulary with room underneath it.

The pressure pulling regions apart isn't easing. Gartner forecasts worldwide sovereign cloud infrastructure spending to reach about $80 billion in 2026, up roughly 36% from 2025, with regulated industries among the biggest buyers after governments, a sign of how firmly data and governance are being tied to local jurisdictions. Aggregating across those boundaries is hard even under supervisory pressure. Deloitte, summarizing the Basel Committee's latest progress report on risk data aggregation, notes that only two of 31 assessed global systemically important banks were fully compliant, with fragmented IT landscapes, legacy systems, and manual processes slowing progress. Program data isn't risk data, but the aggregation problem looks familiar.

Good Local Governance, No Group View

Multi-region institutions grow governance locally, and local governance usually reflects real local requirements: a different regulator, a different legal entity structure, an acquired business with its own inherited practice, or a market where the program mix is genuinely different.

So four regions end up with four gate models, four reporting formats, four status vocabularies, and four working definitions of what amber means.

At group level, reporting becomes translation. Somebody maps each region's stages onto a common set, normalizes the RAG ratings, and assembles a portfolio view, and that translation is where the useful detail goes. By the time twenty programs across four regions are expressed in a common format, what survives is often a color and a date.

The group committee then makes decisions from the least informative version of the information that exists anywhere in the organization.

Why the Usual Fix Tends to Fail

The standard response is to mandate a single global model: pick one, roll it out, and enforce it.

It tends to fail for a reason that's easy to miss from the center. Some regional variation isn't cultural preference or legacy inertia, it's regulatory. A gate model that satisfies one supervisor may not satisfy another, and the region asked to abandon its model is often the one with the least room to do so.

So the mandate gets partially adopted. Regions comply at the reporting layer and keep their real process underneath, which produces the worst outcome, a group view that looks standardized and isn't.

Shared Structure, Local Configuration

The workable version separates two things that usually get conflated.

1. At group level, standardize the vocabulary: One set of stage definitions, one RAG rubric with actual criteria rather than judgment calls, and one definition of what counts as a gate decision. That's a small surface area, and it's where comparability comes from.

2. Underneath, let regions configure: Their own gate criteria, evidence requirements, and approval chains, plus additional stages wherever a regulator requires them.

The group view then rolls up from a shared structure rather than being translated into one, so comparability comes from the vocabulary rather than from forcing identical process for all.

This only works if regions can configure without a central change request. If every local adjustment queues behind a global backlog, regions will build shadow processes, which brings the original problem back with extra steps. Admin configurability is the enabling condition, not a convenience.

What Changes at Group Level

The portfolio view becomes a query rather than a compilation, available on any day rather than the week after month end. The detail survives, so a group committee can drill from the portfolio into an actual program in an actual region and see the real gate status rather than a normalized summary of it. And the translation effort largely disappears, which in many institutions is a meaningful share of what the group PMO spends its time on.

Why It Belongs Beside the Business Data

Regional programs touch regional customers, products, and entities, and most of those are already segmented in the institution's Salesforce org. In Cloud Coach, the group vocabulary sits at the top and each region configures its own gates and approval chains underneath it, inside that same org, so the regional boundaries already defined through sharing rules and entity structure are the ones that apply, rather than being recreated in a separate system. Regional admins can make those changes themselves without waiting on a consultant or a central backlog, which is what keeps the shared structure honest. For institutions with data residency obligations, that's far from a minor consideration.

How We See It

The goal was never one process, and mandating one is how many of these programs stall. The goal is one vocabulary, with room underneath it, and a roll-up that doesn't destroy the detail on the way up.

Frequently Asked Questions Related to Fintech

• How should a multi-region bank standardize PMO governance?

Standardize the vocabulary at group level, meaning stage definitions, a RAG rubric with real criteria, and what counts as a gate decision, while letting each region configure its own gate criteria, evidence, and approval chains underneath.

• Why do global gate models fail in regional PMOs?

Because some regional variation is regulatory rather than cultural. When a region can't abandon its model, it tends to comply at the reporting layer and keep its real process underneath, so the group view looks standardized without actually being so.

• How can group PMOs roll up regional program status without losing detail?

By running every region on a shared structure in one system, so the portfolio view rolls up from live program records and committees can drill into any region's actual gate status instead of a translated summary.

Sources Cited

Gartner, "Gartner Says Worldwide Sovereign Cloud IaaS Spending Will Total $80 Billion in 2026," press release, February 9, 2026. https://www.gartner.com/en/newsroom/press-releases/2026-02-09-gartner-says-worldwide-sovereign-cloud-iaas-spending-will-total-us-dollars-80-billion-in-2026

Deloitte, "BIS Assessment of Banks' Compliance to BCBS 239 Principles," February 23, 2024 (summarizing the Basel Committee on Banking Supervision's 2023 progress report). https://www.deloitte.com/us/en/services/consulting/blogs/bis-assessment-bcbs-239-principle-compliance.html