Audience Insights

Audience Insight Reports offer insights into your customers based on your customer and event data. You can see a breakdown of your customer base by any customer attribute. Additionally, these reports can help you understand how different customer cohorts (based on any customer attribute of your choice) are contributing to your top line and bottom line metrics like revenue, orders and other goal metrics.

Using Audience Insights, you can identify target audiences who are driving the most impact and allocate your marketing budget effectively. You can get the valuable insights needed to assess customer lifetime value, analyze performance, compare behavior, and identify cohort patterns to drive business decisions.

Consider the following use cases:

• You can visualize your customer distribution by joining month. For example, what percentage of your customers are new (joined less than a month ago), what percentage of customers joined less than a year ago and what percentage of your customers joined more than a year ago.
• For an e-commerce or retail brand, you can evaluate order distribution by state or the average order value by first time customers.
• For a media brand, you can analyze total hours watched by subscription tier or average hours watched by sign-up month.
• For a consumer finance brand, you can examine the average number of credit card applications by customer age or total revenue by customer type.

Anatomy of a Report

When you create a report, you can set the following configurations:

• Name of report
• Type of report
• Metric for the report
• The time duration
• The function applied to the metric
• Group by attribute
• Filter users by Segment
• Chart for visualizing data
  • Chart type
  • Hover over data points to view data values
  • Toggle any series on/off
• Table - Tabular representation of the results
• Download report in CSV format
• View report in full screen mode

Report Metrics

You can generate a report based on metrics that are preset for your account on the Blueshift platform and might include the following:

• Customer Count - the total number of customers. This metric is availabe for all accounts.
• Other metrics - In addition to customer counts, you can also choose to analyze goal metrics for your account. Contact your CSM to set these up for your account. Some examples of such metrics include:
  • Purchases (i.e. number of orders)
  • Revenue
  • Visits

Filter by Time

You can filter reports by the specified time period:

• Last 7 days
• Last 30 days

Note: For Customer Counts, you cannot filter by time or apply a function to the data.

Function

You can apply the following function to the selected metric data:

• Sum - Sum of the selected metric (for example, revenue) across all customers in the cohort.
• Average - Sum of the selected metric (for example, revenue) across all customers in the cohort / Total number of customers in the cohort.
• Max - Maximum value of the selected metric (for example, revenue) across all customers in the cohort.
• Min - Minimum value of the selected metric (for example, revenue) across all customers in the cohort.

Additional options for the Average metric

If you select the Average function you can Sort your results by the Average values (the computed value) or by the Customer Count (the Denominator).

If you sort by the Average value, you can additionally specify the Minimum sample size. This is the minimum number of customers who must be present in a group in order for that group to be considered for the analysis. This setting helps you to filter out outliers.

For example consider that you have 3 customer groups:

• High Value Customers with 10,000 users and $20,000 revenue - Average revenue = $2/customer
• Mid Value Customers with 20,000 users and $30,000 revenue - Average revenue = $1.5/customer
• Low Value Customers with 2 users and $10 revenue - Average revenue = $5/customer

The ‘Low Value Customers’ group has the highest average revenue but has only 2 customers in the group and hence the average may not be representative of the group. So it might make more sense to filter out groups where the sample size is that small.

Group By

Compare different customer groups based on standard or custom customer attributes like city, country, age group, gender, customer tier and so on.

Select the Include ‘unknown’ category option to see how the group of customers with no value for the selected group by attribute compares to other groups (i.e. those with a definite value for the group by attribute).

Boolean type attributes

For attributes of type Boolean, the grouping is by attribute value: true, false, unknown.

String type attributes

For attributes of type String, the grouping is by attribute value. You can set the Limit for the maximum number of groups to display in the report. For example, if you are grouping by cities and you set Limit to as 5, then the top 5 cities will be displayed in the report.

Note: It is only possible to Group By string type attributes which have a cardinality less than 10,000.

Numeric or Date type attributes

For numerical and date attributes, you can create bins based on the range of the attribute value. You can select the number of bins by adjusting the Limit setting. Additionally, you can adjust the minimum and maximum values for each bin.

Example 1: You want to see a breakdown of your customers based on their year of birth. You are considering the following groups of customers:

1. Baby Boomers: Those born before 1965
2. GenX: Those born between 1965 and 1981
3. Millennials: Those born between 1981 and 1997
4. GenZ: Those born after 1997

To get the results, create an 'Audience Insights' report with the following settings:

• Metric = Customer Count
• Group by = birth_year
• Limit to = 4 (since you have defined 4 bins or user groups)
• Bins: Bin 1: Less than 1965, Bin 2: 1965 to 1981, Bin 3: 1981 to 1997, Bin 4: Greater than 1997

Once you've created a report, you can also edit the bins at a later time, if the situation so demands. You can use the '+Add Bin' button to add a bin between 2 existing bins. You can also click on the trash/ delete icon to delete an existing bin.

Example 2: If you want to update the analysis in example 1 to add a new customer group for Generation Alpha (i.e. those born after 2013), you can easily do so by clicking on the '+ Add Bin' button between Bin 3 and Bin 4.

Absolute vs Relative Values

When you group by a customer attribute, you can choose to either plot the absolute values or the relative (i.e. percentage) values. Simply select the '# Absolute' or '% Relative' button next to the chart type selector on the top right corner of the chart.

Example: You want a breakdown of your customers by gender. When you use 'Group by' = 'gender', you might see results like: 

• Male: 291,350
• Female: 186,450

This is the default analysis showing the absolute values/ counts of users in each group. If you are only interested in percentage breakdown, you can click on the '% Relative' button. It will now present the same chart as percentages. For example:

• Male: 61%
• Female: 39%

Filter by segment

Instead of analyzing your entire customer base, you can narrow down your analysis to one or more segments of customers. Not selecting a segment is the equivalent of running your analysis on your entire customer base.

Note: Click Analyze Segment when you are building your segment to generate Audience Insights for that segment and to identify the right set of customers.

Types of Charts

Display the report as a Bar or Column chart. Use the checkboxes in the legend to toggle data series on or off.

Tabular report

The table displays the information for the selected metric that is grouped by the selected attribute. 

Limitations

• Audience insights offers grouping by any customer attribute on the customer profile with the following exceptions:
  • The attribute must have a cardinality under 300,000. (In other words, the attribute you want to slice your customer or event data by cannot have more than 300,000 unique values)
  • Predictive scores are not supported. (In other words, you cannot slice and dice your customer or event data by a predictive score)
• When you're analyzing event data using audience insights reports, you will have access to 6 months of historical data. So for example, if you want to slice and dice revenue, orders, visits or other goal events by some customer attribute, you'll be able to do so for the last 6 months of activity