What is Data Aggregation? How Summarising Data Enables Business Analytics

Every business metric you care about is an aggregation. Total revenue, average margin, count of customers, sum of inventory value — all are aggregations of individual records.

Understanding how data aggregation works helps you ask better questions and interpret analytics results more accurately.

Common Aggregation Functions in Business Analytics

SUM: Total of all values

• Total sales this month: SUM of all transaction amounts
• Total inventory value: SUM of (quantity × unit cost) for all items

COUNT: Number of records

• Number of customers who ordered this month
• Number of invoices raised

AVERAGE: Mean of all values

• Average order value
• Average days to collect payment

MIN / MAX: Smallest or largest value

• Lowest-performing product by sales
• Maximum single transaction in a day

MEDIAN: Middle value when sorted

• Median invoice value (less affected by outliers than average)

DISTINCT COUNT: Number of unique values

• Number of unique customers (not total transactions by customers)
• Number of unique products sold

Aggregation Dimensions: The "Group By" in Analytics

Aggregation is most powerful when combined with grouping — "total sales by product category", "count of orders by region", "average margin by salesperson."

In SQL, this is the GROUP BY clause. In BI tools, this is the "breakdown" or "dimension" selection.

Common aggregation dimensions for Indian businesses:

• By time period (day, week, month, quarter, year)
• By geography (city, state, zone, region)
• By product (SKU, category, brand)
• By customer (customer, segment, channel)
• By salesperson or team

The Granularity Challenge

Data granularity refers to the level of detail in raw data. The more granular the data, the more aggregation options you have.

Example: If you have transaction-level sales data (each individual sale with customer, product, quantity, price, date), you can aggregate it in any combination.

But if data is already pre-aggregated (stored as monthly sales by region totals), you cannot break it down further — you can't go back to the individual transaction level.

This is why data warehouses store raw, granular transaction data — so analytics tools can aggregate it in any way needed without pre-defining the analysis.

Aggregation Pitfalls to Avoid

Double-counting: Aggregating data that's already been partially aggregated can create inflated totals.

Average of averages: Averaging monthly averages to get a yearly average produces incorrect results unless all months have equal record counts.

Mixing metrics with different granularity: Adding transaction counts from a daily table with transaction counts from a monthly table requires careful handling to avoid incorrect totals.

See what is OLAP for how online analytical processing enables fast aggregation across large datasets, and drill-down analysis for navigating from aggregated summaries to underlying detail.