Performance Issues in Customer-Centric Power BI Model (Large Dataset)

Hi all,

I’m building a customer-centric PBI dashboard using a large dataset for the 12 months (43,617,492 rows) already aggregated at customer-month level (not transactional).

It includes multiple dimensions (country, region, product, months etc.).

The report has time trends, composition charts, and many slicers.
I need dynamic KPIs like active, new, lost customers, churn, retention, and YoY %, depending on user-selected periods (1+ months), plus averages per customer. Example of measure for Active Customers:
Active_Customers =
VAR Customers =
SUMMARIZE(
DB_CUSTOMER,
DB_CUSTOMER[Customer_id]
)
RETURN
COUNTROWS(
FILTER(
Customers,CALCULATE([Current Year Shipments]) > 0))
I’m facing: Slow performance, Query execution conflict errors,High Capacity Unit (CU) consumption spikes

Constraints:

• Must keep full customer base
• Metrics rely on customer-level calculations over time

What I’m Looking For

I’d appreciate guidance on:

• Best practices to optimize this kind of customer-based KPI model at scale
• Whether my issue is more likely due to:
• Data model design
• DAX complexity (especially YoY + status calculations)
• Capacity / Fabric configuration
• Suggested approaches such as:
• Further aggregation strategies
• Hybrid tables / incremental refresh
• Calculation optimization
• Alternative modeling patterns for customer lifecycle metrics

Thanks in advance for your help! Happy to provide more technical details if needed.