Aggregates are calculations made on data. They take into consideration all events for each customer separately. They help you calculate an information for every customer (e.g. minimum, maximum, sum, count, average, first, last, exist). Aggregates can be found in data management.
The process of calculating an aggregate is as follows:
- the algorithm runs through the data
- the algorithm gives you the value you asked for (it is important to understand because it differentiates aggregates from running aggregates)
Creating an aggregate
To calculate an aggregate, the process is as follows:
- Select a calculation operator (sum, first, last …)
- Select an event on which you want to calculate (e.g. count purchase start would count number of purchases for every customer) –> 1st aggregate
- If desired, add an attribute where you can filter for customer you need (e.g. count purchase where total price > 20€ would count only those purchases where the price was above 20€) –> 2nd aggregate
If a customer had 5 purchases for 15€, 25€, 35€, 45€, 55€, the first aggregate would calculate a result of 5 and the second – 4. The time span of these purchases can be specified as anywhere in the analysis using the time filter.
Aggregates in analyses
When using an analysis in reports, you can create an aggregate directly, without creating it beforehand in the data management.
When adding an event, you can switch to customer attributes and select one of those. Furthermore, there is a plus sign in the aggregates title which creates an aggregate within an analysis.
Aggregates can be used for filtering customers, segmenting customers, expressions and consequently getting insight or using it for the next campaign.
For instance, if I am interested to see customers, I could select the first utm_source of the customers who came to the website. Or, alternatively, I could use a filter to only see the customers who had at least 3 purchases in last 12 months. If I wanted to show a banner to customers who are not “first-time” visitors, I could filter people who have more than 1 session start. If you wanted to create a campaign based on the last product that a customer purchases, you can create such segmentation and base it on the “last purchase item name” aggregate.