In general, audiences should be created in Lytics using the audiences builder, then pushed to actioning platforms via built-in integrations. These audiences update in real-time and can be automated to keep downstream tools in sync.
However, in some cases, there are more narrow and targeted audiences of users who you need to target for one-off campaigns or pushes to downstream tools. In these cases, we recommend creating the audience in your data warehouse then pushing that group of users to Lytics.
Examples of audiences to build in Lytics:
- Users who have visited the site in the last week
- Users who purchased a product within the last three days
- Users who opened an email and have not logged in yet
Examples of one-off audiences to push from an external querying tool:
- Users who attended event A and purchased a product over $100 two quarters ago
- Users who visited a booth during conference A and, on average, ordered at least three products per month over the last two years
- Users who should receive a promotional offer based on a list provided by an outside agency
In this example, we’ll use Google BigQuery as our data warehouse and querying tool.
Once notifying your account manager, your Lytics instance will be configured to pull data from a table called lytics_custom_audience_push with the following schema:
- Create a table called
lytics_custom_audience_pushwith the above schema.
- Write your query. The query output must have three columns: the audience name, the audience value, and the customer id.
- Under query settings, set
lytics_custom_audience_pushas the destination table for your query results.
- Lytics pulls data from
lytics_custom_audience_pushon a rolling basis, so you should see the custom_audiences field in Lytics populate within roughly 6 hours (max).
Let’s say you need to push an audience of VIP users selected for targeting based on their purchase history and attendance at an event.
"vip_users" AS custom_audience_name,
"true" AS custom_audience_value
FROM user_db u
JOIN event_db e
ON e.id = u.event_id
WHERE u.purchase_avg > 50
AND u.purchase_num > 5
AND e.id IN ( 748, 573, 934 )
lytics_custom_audience_push table should look something like this:
After this query is pushed to the
lytics_custom_audience_push table, you can use it in the Lytics audience builder under the
lytics_custom_audience_pushtable is configured to accept strings as values. If you need to push a different data type for the value, let us know. We can also configure a table and mappings for integers, floats, and timestamps.
- One-off audience pushes are typically used for one downstream campaign. If you find yourself repeatedly targeting a one-off audience, talk to your Lytics account manager to have that audience reconfigured as a user attribute.
- If you are using a different data warehouse, you can send the table as a CSV to an SFTP server or S3 bucket for Lytics to pull.
Updated 8 months ago