Dashboards do not change companies. Cultures that question dashboards do.
We have audited dozens of enterprise BI deployments. The technical work is usually fine. The dashboards exist, the numbers are reconciled, the access is provisioned. The cultural problem is that nobody opens them. Real data culture is harder than the tooling.
The first signal: are decisions referenced in writing?
In a data culture, decision documents cite the dashboard or query that informed them. Not "we think users want X" but "Q3 retention dropped 4 points among users who did not complete onboarding step 3 — link." If decisions in your org are not documented this way, the dashboards are decoration.
The second signal: who can write SQL?
Pure consumer cultures have a small data team and a wide audience that consumes dashboards. Producer cultures have analysts in every function — marketing, finance, product, ops — who can write their own queries when the dashboard does not answer the question. The producer culture is the goal.
The third signal: are metrics defined?
In a data culture, "monthly active users" has one definition documented in a metrics layer. In a non-data culture, marketing's MAU and product's MAU and finance's MAU are different numbers. The work to reconcile them is the work that builds the culture.
The investments that matter
- SQL training for non-engineers across functions.
- Metrics layer (Cube, dbt Semantic Layer, Looker LookML) so definitions are central.
- Dashboard hygiene: retire dashboards nobody opens, version the ones people use.
- Decision documents: a template that requires citing data in every product / strategy / hiring decision.
The investments that do not
Ever-fancier visualization tools. Ever-larger data warehouses. AI-generated dashboards (so far). The technical layer is not the bottleneck.
The leadership move
Executives who reference data in their own writing teach the org that data matters. Executives who say "I trust my gut on this one" teach the opposite. There is no investment in tooling that overcomes the second message.