Why are you so slow?

Speed and performance of a data platform are often core business needs.

• Waiting for a business intelligence report to load.

• Paying for long-running data pipelines or queries.

• Enduring daily data refreshes instead of hourly.

If you are facing slow performance and delayed data delivery to business teams, here are some places to start.

Reasons why data platform performance suffers.

• Wrong technology.

  • NoSQL instead of SQL or Spark. OLTP instead of OLAP. Direct Query instead of Import. Or vice versa. Limited caching.

• Poor modeling.

  • Dimensional vs. relational. Aggregate vs. atomic grains. Mismanagement of business keys, surrogate keys, and source system keys.

• Poor workload isolation.

  • Mismanagement of Individual users, group reporting, or ETL workload permissions. One user’s query crowds out other key workloads slowing down teams of people or other high priority tasks.

These can be highly technical challenges and design decisions, but they are rooted in removing friction from business users' data experience.

The design that could have worked last month or last year, may no longer serve the business well.

Always be architecting.

Always be designing.

Always be modeling.

I’m here,

Sawyer