Digital Product Engineering9.4 Information Architecture
VOL. IX · CH. 9.4 · ENTERPRISE ARCHITECTURE

Information Architecture

Not to be confused with a single site's navigation structure — this is data ownership across an entire enterprise.

DivisionEnterprise Architecture
DifficultyAdvanced
Prerequisites9.1 5.1
Related9.2
2 min read · 342 words

9.4.1Definition

At the enterprise level, information architecture defines what data entities exist across the organization, who owns each, what the authoritative source of truth is for each, and how data flows between systems (9.2) — distinct from a single product's content or navigation structure, which is a different, product-level use of a similar term.

9.4.2Why It Exists

Without a clear model of data ownership, the same entity — a customer, a product, an order — often ends up defined and stored independently in multiple systems, each with a slightly different, drifting version of the truth. Enterprise information architecture exists to establish a single authoritative source for each data entity, so every system knows where the real answer lives rather than maintaining its own, inevitably diverging copy.

9.4.3Core Concepts

  • System of record — the single authoritative source for a given data entity, which every other system references rather than duplicates.
  • Master data management — processes ensuring core entities (customers, products) stay consistent across every system that touches them.
  • Data lineage — a traceable record of where a piece of data originated and how it has moved or transformed across systems.

9.4.4Common Mistakes

  • No designated system of record for a core entity, leaving multiple systems each maintaining a slightly different version of "the truth" with no way to reconcile them.
  • Data silently duplicated and drifting across systems, discovered only when two reports disagree and nobody can say which is correct.
  • No data lineage tracking, making it impossible to trace a data quality issue back to its actual origin.

9.4.5Best Practices

  • Designate an explicit system of record for every core data entity, documented and enforced across integrations.
  • Implement master data management processes for entities used across multiple systems.
  • Track data lineage for critical entities, especially those feeding regulatory or financial reporting.
Real-World ExampleA CRM (2.12) is commonly designated the system of record for customer contact data specifically to prevent marketing, sales, and support tools from each maintaining their own drifting copy of the same customer's information.