Definition
Frequency is the ability of a data model to update, recalculate, and communicate information at the appropriate rhythm of the phenomenon it represents, without delays that distort interpretation or overloads that introduce noise.
A model with correct frequency resonates with the reality it measures.
In Power BI, frequency ensures that:
- data is refreshed when it should be,
- metrics reflect the expected state,
- and users interpret the “now” with temporal accuracy.
Nature
Temporal, operational, and perceptual.
Frequency is not just technical (refresh);
it is alignment between business time and analytical time.
It emerges when:
- the model understands the rhythm of the real process,
- updates are neither late nor excessive,
- and analysis avoids both obsolescence and noise.
Function
To ensure analysis arrives on time.
In practice, frequency allows:
- decisions to be timely,
- indicators to reflect the correct moment,
- users to trust data freshness.
A correct model delivered late
is analytically useless.
Consequence
- Interpretations aligned with the present.
- Reduced friction between analysis and operations.
- Trust in operational dashboards.
- Avoidance of decisions based on stale data.
Without proper frequency, the model loses synchronization with reality.
Signals of Correct Frequency
- Refresh cadence matches business rhythm.
- KPIs reflect the expected process state.
- Users know when data was last updated.
- No decisions are made “blindly”.
- Latency is known and accepted.
Signals of Lack of Frequency
- Data arrives too late to decide.
- Refresh is so frequent it introduces noise.
- Users doubt dashboard freshness.
- Operational metrics are treated as historical.
- Analysis becomes reactive instead of proactive.
This is not a data failure: it is temporal misalignment.
Conceptual Example in Power BI
Without frequency (❌)
- “Real-time” sales with daily refresh.
- Operational indicators calculated weekly.
- Users asking: “Is this updated?”
Result:
👉 loss of usefulness.
With frequency (✅)
- Operational KPIs with hourly refresh.
- Strategic KPIs with daily or monthly refresh.
- Explicit and understood latency.
Result:
👉 trust and action.
Frequency and Model Design
Frequency is strengthened when:
- Operational and strategic metrics are differentiated.
- Refresh is designed by layer, not globally.
- Expected latency is documented.
- The model avoids unnecessary recalculation.
- Users understand the “rhythm” of the data.
Frequency is not fixed with hardware;
it is designed with analytical judgment.
Interactions
- Persistence → continuity over time.
- Consistency → stable meaning.
- Coherence → internal logic.
- Congruence → intent–outcome alignment.
- Informational Scarcity → noise avoidance.
- Frequency → temporal synchronization.
Without frequency, the other properties arrive too late.
🔧 Samples — Frequency applied in Power BI
🔹 Sample 1 — Operational vs strategic KPI
Without frequency (❌)
Using the same refresh for everything.
With frequency (✅)
- Daily sales → hourly refresh.
- YTD sales → daily refresh.
- Annual targets → monthly refresh.
📌 Result:
Each KPI vibrates at its correct time.
🔹 Sample 2 — Explicit latency
Without frequency (❌)
Dashboard without temporal indication.
With frequency (✅)
Last Refresh :=
MAX(RefreshLog[RefreshDateTime])
📌 Result:
Users know exactly how current the data is.
🔹 Sample 3 — Frequency and noise
Without frequency (❌)
Refreshing every 5 minutes for stable metrics.
With frequency (✅)
Refresh aligned to the real change rate of the process.
📌 Result:
Less irrelevant fluctuation, more signal.
🔹 Sample 4 — Frequency by layer
Without frequency (❌)
Entire model refreshes at the same pace.
With frequency (✅)
- Operational layer → high frequency.
- Historical layer → low frequency.
📌 Result:
Efficiency and temporal clarity.
🔹 Sample 5 — Frequency and user perception
Without frequency (❌)
Users assume the data is “current”.
With frequency (✅)
The model explicitly communicates its rhythm.
📌 Result:
Informed decisions, not assumptions.
🔹 Sample 6 — Antipattern vs Pattern (Frequency)
❌ Antipattern — Desynchronized model
- Arbitrary refresh.
- Out-of-time KPIs.
- Distrustful users.
✅ Pattern — Synchronized model
- Defined rhythm.
- Known latency.
- Timely analysis.
📌 Rule of thumb:
If data does not arrive in time to decide, frequency is wrong.
Synthesis
Frequency does not accelerate the model.
It synchronizes it with reality.