Organisations often pursue a single source of truth as a data ideal: one database, one dashboard, one authoritative number. This pursuit is understandable. Fragmentation is frustrating. Inconsistency feels like failure.
Yet in dynamic systems, a single source of truth is not only unrealistic — it is often harmful.
Truth Is Contextual
Different roles require different representations of reality. An operator, an analyst, and an executive may all need valid — yet incompatible — views of the same system.
Forcing convergence into a single metric or model:
- Strips away contextual nuance
- Suppresses legitimate disagreement
- Privileges one perspective over others
Uniformity masquerades as clarity.
Synchronisation vs Understanding
Single-source architectures often prioritise consistency over comprehension. When discrepancies arise, they are treated as errors to be eliminated rather than signals to be examined.
In practice, divergence between views often reveals:
- Measurement uncertainty
- Temporal misalignment
- Differing decision horizons
These are not problems to be resolved — they are realities to be navigated.
The Political Nature of “Truth”
Declaring a single source of truth is also a power move. It determines who defines reality, whose metrics matter, and which interpretations are legitimised.
When data authority is centralised, contestation diminishes — sometimes at the expense of accuracy.
Designing for Plurality
A more robust approach accepts multiple, purpose-specific truths:
- Operational truth for action
- Analytical truth for understanding
- Strategic truth for direction
These truths coexist, overlap, and sometimes conflict. Design must make these relationships explicit rather than forcing artificial alignment.
From Authority to Coherence
The goal of data design should not be singular truth, but coherent pluralism — multiple perspectives that can be compared, reconciled, and debated.
The question shifts from: “Which number is correct?”
To: “Why do these numbers differ — and what does that tell us?”
In complex systems, disagreement is not a failure of data.
It is often the most valuable signal available.
Leave a Reply