Behaviours

Behaviours in Orbit give you recency, frequency, and value statistics (RFV) directly from your customer and transaction data. Rather than building complex Audiences before you can analyse your data, you configure a behaviour once and use it across your dashboards and audiences.

Each behaviour targets a different aspect of how your customers engage with your brand.

The key benefit of behaviours is that they give you factual evidence rather than assumptions. Whether you need to know if customer spend is declining, whether donation frequency matches your expectations, or how long the typical gap is since a customer last transacted, behaviours surface that evidence on the dashboard without requiring any audience building work.

Understanding customer cohorts¶

A common starting point with behaviours is to map your customer base into cohorts based on their engagement level. A typical segmentation looks like this:

• First-time customers: Contacts making their first-ever transaction, and what that transaction was worth
• Active customers: Contacts transacting within your expected frequency and value range
• Lapsing customers: Active contacts who have started to drop off and their recency, frequency, or value is lower than expected
• Dormant customers: Contacts who haven’t transacted for a significant period and need a re-engagement campaign

What counts as "lapsing" or "dormant" depends on your purchase or donation cycle.

Behaviours give you the data to calibrate your own thresholds against reality.

How Orbit resolves behaviours¶

When you apply a behaviour, Orbit performs an on-the-fly aggregation. It retrieves all child records (for example, bookings or purchases) linked to each parent record (for example, a person or household), applies any transaction-level filter you have specified, and computes the result at the parent level. That resolved value feeds into charts, filters, and audience counts.

Table hierarchy requirement¶

Behaviours require a parent-child relationship between tables in your data model. The transaction table must sit below the grouping table in the hierarchy.

Where you can use behaviours¶

You can add behaviours in several locations across Orbit. The options available to you depend on the tile or context you are working in.

Dimensions¶

Use behaviours as dimensions on charts, cubes, Venn sets, and data grid columns. As a dimension, a behaviour groups your data by the resolved RFV value.

Measures¶

Use behaviours as measures on charts, cubes, venn sets, and number tiles.

Calculated measures¶

Reference behaviours inside calculated measures to build proportional or comparative figures.

Venn diagrams¶

Use behaviours to define the sets in venn diagrams.

Dashboard filters and analysis filters¶

Add behaviours as user filters and dashboard filters to drive the data shown across an entire dashboard. This is useful for building an RFV summary view.

Analysis filters work the same way and are available when building profile tiles.

Audience filters¶

Add behaviours as filters when creating an audience.

When you add a behaviour to an audience filter, Orbit displays an RFV icon next to the clause, indicating whether it is a recency, frequency, or value behaviour.

Recency behaviours¶

Use recency to retrieve the most recent (or earliest) value of a variable for each contact.

Adding a recency behaviour¶

To add a recency behaviour:

1. Go to where you want to add a behaviour and click + Add Filter.
2. In the Select Filter dialogue, click Behaviours.
3. Select Recency.
4. Enter a Name for the behaviour.
5. Under Transaction, click + Add Variable and select the variable whose value you want to retrieve.

6. Under Per, select the relevant table.

   

7. Under Order by, click + Add Variable and select a date or numeric variable to order by.

8. You can now choose how many of the Earliest or Latest records.

9. Optionally, click + Add filter to restrict which records are included in the calculation.

   

   Tip

   Use a date range filter within the behaviour to scope recency to a specific window, for example, bookings from two years ago to one year ago. This lets you compare cohorts across time periods without building separate audiences.

10. Click Apply.

Frequency behaviours¶

Use frequency to count how many transactions each contact has, optionally scoped to a specific subset.

Frequency is not limited to purchases. You can count any repeatable transaction-level event: bookings, complaints raised to a call centre, abandoned baskets, and so on.

Adding a frequency behaviour¶

To add a frequency behaviour:

1. Go to where you want to add a behaviour and click + Add Filter.
2. In the Select Filter dialogue, click Behaviours.
3. Select Frequency.
4. Enter a Name for the behaviour.
5. Under Transactions, click + Add Table and select the child transaction table.

6. Under Per, select the relevant table.

   

7. Optionally, click + Add filter to restrict which records are included in the calculation.

   

   Tip

   A transaction filter is particularly valuable with frequency. Counting all transactions ever is rarely actionable. Filtering to a specific product type, date range, or destination gives you a count that maps to a real business question, for example, the number of accommodation-only bookings per person.

8. Click Apply.

Value behaviours¶

Use value to aggregate a numeric or currency variable across transactions.

Adding a value behaviour¶

To add a value behaviour:

1. Go to where you want to add a behaviour and click + Add Filter.
2. In the Select Filter dialogue, click Behaviours.
3. Select Value.
4. Enter a Name for the behaviour.
5. Under Transactions, , click + Add Variable and select a numeric or currency variable.

6. Under Function, select an aggregation function (see Aggregation functions on value behaviours below).

   

7. Under Per, select the relevant table.

8. Optionally, click + Add filter to restrict which records are included in the calculation.
9. Click Apply.

Aggregation functions on value behaviours¶

The following functions are available when configuring a value behaviour.

Transaction filters¶

Any behaviour accepts an optional transaction filter. This restricts which child records are included in the aggregation before the result is computed at the parent level.