Architecture & assurance

Make the Finance Landscape Work as One Architecture.

Most finance landscapes were assembled a decision at a time, each one defensible on its own. This engagement decides where data, calculations, planning logic, consolidation logic and reporting actually belong — and what that means for what you already have.

When this is worth doing

Nobody sets out to build an incoherent landscape.

It happens by accumulation. A reporting requirement is met where it is easiest to meet. A platform arrives for one purpose and is reused for three more. A calculation is added in the layer whoever needed it had access to. Every individual decision was reasonable and the result is a landscape nobody would have designed.

  • Two systems produce the same number and finance trusts one of them for reasons nobody can articulate
  • A new report means a new interface
  • Nobody can say where a figure came from without asking a specific person
  • Plan and actual will not reconcile because they were never designed to the same grain
  • A platform was bought for one requirement and has quietly become load-bearing for several
  • The landscape has grown by addition and never by decision

What the engagement answers

Eight questions, answered specifically for your landscape.

Which system is the source of truth for each finance data domain?

Not in principle — in practice, per domain, with a named owner. Most landscapes have two answers for at least one domain and nobody has noticed.

Where should transformations and finance calculations live?

The same calculation implemented in three layers will eventually disagree with itself. The question is which layer owns it, and what the others are allowed to assume.

How should actuals, plan and consolidated data be aligned?

Three datasets, three grains, three calendars and one variance report. The alignment has to be designed, not reconciled afterwards.

When is Datasphere or BW actually required?

Sometimes the honest answer is that it is not, and the landscape is carrying a layer it does not need. Sometimes it is the only sensible place for the logic.

Which reports belong in S/4HANA, Group Reporting, SAC or a disclosure tool?

Reporting placement drives half the integration effort in a finance landscape and is usually decided by whoever built the report.

How should hierarchies and master data be governed?

Who can add a cost centre, who can restructure a hierarchy mid-year, and what happens to comparatives when they do.

How should latency, performance and auditability shape the design?

Live, replicated and modelled approaches each buy something and cost something. The trade-off should be deliberate.

Where are point-to-point interfaces creating avoidable complexity?

Interfaces accumulate quietly. The cost shows up as reconciliation effort long before anyone counts them.

What you get

A target architecture with the transition attached.

  • Current-state landscape and data-flow map
  • Reporting inventory and ownership map
  • Target architecture and system responsibilities
  • Data-flow and integration principles
  • Calculation and semantic-layer placement decisions
  • Reporting-channel strategy
  • Risks, dependencies and a transition roadmap
Typical duration
2–6 weeks
Depending on the number of platforms, reporting products and source systems in scope, and how much of the current landscape is already documented.

Our bias, stated

Integrate where the process requires it. Stay separate where it does not.

Architecture work has a natural pull towards more integration, because integration looks like progress. It also creates dependency, and dependency is what makes a landscape fragile. Where two things genuinely need to connect, we connect them properly. Where they do not, leaving them alone is the better answer and we will argue for it.

Related

Where this sits.

Next step

Describe the landscape and the argument you keep having.

Which platforms are in play, which reports matter most, and the disagreement that never quite resolves. That last one usually points straight at the decision that needs taking.