Skip to main content

Create a Statistics configuration

A Statistics configuration defines which fields should be summarized and which measures should be calculated.

A good configuration starts with a clear business question and produces metrics that can be reviewed and explained.

Before you begin​

Before creating the configuration, write down:

  1. The purpose - What question should the summary answer?
  2. The numeric Fields - Which Fields should be measured?
  3. The grouping fields - Should results be broken out by category, period, owner, or another field?
  4. The measures - Which metrics are needed?
  5. The review plan - What results would require follow-up?

Build the configuration in this order:

  1. define the business question
  2. choose the numeric fields
  3. choose grouping fields, if needed
  4. select summary measures
  5. decide how to handle blanks, zeros, and outliers
  6. test on a representative sample
  7. save with a clear name

Step 1: Define the business question​

State the purpose in plain language.

Examples:

  • Summarize total invoice amount by vendor.
  • Count open cases by status and region.
  • Review average processing time by team.
  • Compare transaction totals by month.

The clearer the question, the easier it is to choose the right measures.

Step 2: Choose numeric fields​

Identify the fields that should be measured.

Common examples include:

  • amount
  • quantity
  • balance
  • duration
  • score
  • countable record IDs

Confirm that the selected fields contain values that can be summarized reliably.

Step 3: Choose grouping fields​

Grouping fields let you break summary results into meaningful categories.

Examples include:

  • region
  • department
  • status
  • month
  • vendor
  • account type

Only group by fields that make the summary easier to understand.

Step 4: Select summary measures​

Choose measures that answer the business question.

Common measures include:

  • count
  • sum
  • average
  • minimum
  • maximum
  • range

Avoid adding metrics just because they are available. Extra measures can make the output harder to review.

Step 5: Plan for blanks, zeros, and outliers​

Decide how unusual values should be interpreted.

Ask:

  • Are blank numeric values errors or expected?
  • Does zero mean none, missing, waived, or not applicable?
  • Should unusually high or low values be reviewed?
  • Could duplicates inflate counts or totals?

Step 6: Test the summary​

Use a sample that includes:

  • normal values
  • blanks or zeros
  • unusually high or low values
  • multiple groups, if grouping is used
  • known totals or counts you can verify

Naming guidance​

Name the configuration by business purpose.

Better examples:

  • Monthly Invoice Totals by Vendor
  • Open Case Counts by Region and Status
  • Average Processing Time by Team

Weaker examples:

  • Stats 1
  • Summary Test
  • Numbers

Save checklist​

Before saving, confirm:

  • the business question is clear
  • numeric fields are correct
  • grouping fields are meaningful
  • summary measures answer the question
  • the configuration has been tested on representative data

Next step​

After the Configuration is ready, continue to Run Statistics.