Most organisations invest significantly in two things: the systems that capture data and the reports that present it. ERPs, CRMs, and financial platforms at one end. Dashboards, board packs, and analytics tools at the other. Both receive budget, attention, and governance.
Between them sits a layer that almost nobody designs deliberately. It’s the layer where raw operational data is transformed into meaning — where transactions become KPIs, where activity becomes trends, where thresholds are set and early warnings are defined. We call this the signal layer, and in most organisations, it has grown organically rather than been architected.
The accidental architecture
Here is how it typically happens. Finance needs a monthly revenue view, so they build an extraction from the ERP with their own logic for recognition timing, currency conversion, and entity consolidation. Operations needs a utilisation metric, so they build a separate extraction from the same systems with different aggregation rules. Risk needs a compliance indicator, so they build a third view with yet another methodology.
None of these are wrong. Each is internally consistent. But they were never designed to be coherent with each other. When leadership asks a question that spans two of these views — “why does the revenue trend not match the operational output?” — nobody can answer it without a manual reconciliation exercise, because the signals were extracted through different lenses from the start.
This is not a data quality problem. The data is fine. It’s a signal architecture problem. The same operational truth is being converted into meaning through multiple, uncoordinated processes.
What a designed signal layer looks like
A deliberate signal architecture defines three things. First, how signals are extracted from source systems — with shared rules for timing, granularity, and categorisation that apply regardless of which team consumes them. Second, how signals are normalised — so that a revenue figure means the same thing in the CFO’s dashboard as it does in the COO’s operational review. Third, how signals are distributed — so that every consumer receives the same signal, not their own reconstruction of it.
This does not mean every team sees the same dashboard. Different functions need different views. But those views should be built on a shared signal foundation, not on independent extractions that happen to start from the same data.
Why it gets skipped
The signal layer is invisible in most technology roadmaps. System implementations focus on the deterministic layer — getting the platforms in, the integrations working, the data flowing. Reporting initiatives focus on the interpretation layer — building better dashboards, automating board packs, improving visualisation. The signal layer sits between them, owned by nobody and governed by convention rather than design.
It also resists the usual fix. You cannot solve a signal architecture problem by buying a tool. A BI platform, a data warehouse, an integration middleware — these are infrastructure, not architecture. They provide capability without coherence. The architecture is the set of decisions about how signals are defined, extracted, and governed. Those decisions are organisational, not technical.
The return on designing it
When the signal layer is deliberately architected, the effects are immediate and measurable. Reconciliation drops — because teams stop producing conflicting versions of the same metric. Reporting accelerates — because the signals arriving at the interpretation layer are already consistent. And decision confidence increases — because leadership can act on the numbers instead of first debating whether they’re correct.
The signal layer is where most reporting problems actually originate. Not in the systems, and not in the dashboards, but in the undesigned space between them. Designing it is not a technology project. It’s an architectural one — and it may be the highest-leverage intervention most organisations have never considered.
If your teams are producing conflicting signals from the same data, we can help you see where — and what to do about it.
Begin a Conversation