Visits (Sessions)

Tracking how people interact with your site

Intro

Visits (or sessions) tables record how people interact with your website. Visit history is required to derive an attribution model. This attribution model allows us to tie your growth levers to downstream elements of your growth funnel (e.g. signups and revenue). In other words, visit tables empower the "cause" of cause and effect.

Note that we use the term visits and sessions interchangeably; when relating these terms to what happens in real life:

• A visitor can visit your site multiple times.

• Each visit (or session) consists of multiple page views.

• During each page view, a visitor can interact or "do something" zero or more times, which we call events.

We support three methods to consume visit data:

• Internally modeled "gold" table

• Segment Pages

• GA4 BigQuery Export

Internally modeled "gold" table

Many customers already have a source-of-truth table for tracking visits to their site. When available, we prefer to derive any downstream modeling from such a table. However, to ensure compatibility with our systems, the visitstable you provide must adhere to the following schema:

Column Name	Type	Description
visit_id	varchar	Unique identifier of the session
visitor_id	varchar	Identifier of the anonymous visitor associated with this session (some systems call this anonymous_id or user_pseudo_id)
visited_at	timestamp	UTC timestamp of when the session started
user_id	varchar (nullable)	Identifier of the user associated with this session
referring_page_url	varchar (nullable)	Full URL of the referrer
landing_page_url	varchar	Full URL of the landing page on your site, including parameters.

Additional Constraints

The following additional constraints apply to the above table:

• visit_id is the primary key

We do not explicitly require UTM parameter columns, because we extract these directly from the landing_page_urlcolumn.

While we *highly recommend* that warehouse-native integrations with Roadway have an internal source of truth for visits, this is not always the case. Alternatively, Roadway can build your attribution model by directly consuming Segment's Page table or GA4's BigQuery Export data.

Segment Pages

If you are using Segment's analytics tag, issuing Page calls, and extracting the resulting data to your warehouse, we can consume a view of the resulting table (which is typically called pages,by default). This view should be a subset of columns that obscures any PII; it should expose the following schema:

Column Name	Type	Description
anonymous_id	varchar	Unique identifier of the session
timestamp	timestamp	UTC timestamp of when Page was viewed
user_id	varchar (nullable)	Identifier of the user associated with this session
context_page_referrer	varchar (nullable)	Full URL of the referrer
context_page_url	varchar	Full URL of the landing page on your site, including parameters.
context_page_title	varchar (nullable)	Title of the landing page on your site
context_page_path	varchar	Path of the URL of the landing page on your site
context_campaign_source	varchar (nullable)	UTM source parameter from page URL
context_campaign_medium	varchar (nullable)	UTM medium parameter from page URL
context_campaign_name	varchar (nullable)	UTM name parameter from page URL
context_campaign_content	varchar (nullable)	UTM content parameter from page URL
context_campaign_term	varchar (nullable)	UTM term parameter from page URL

Example SQL

select
    anonymous_id
    , timestamp
    , user_id
    , context_campaign_source
    , context_campaign_medium
    , context_campaign_name
    , context_campaign_content
    , context_campaign_term
    , context_page_url
    , context_page_title
    , context_page_path
    , context_page_referrer
from
    <your_segment_destination_schema>.pages

GA4 BigQuery Export

This method is only for customers who use BigQuery for their warehousing needs, AND have previously established a GA4 BigQuery Export.

If you are using Google Analytics and BigQuery, and you have set up the GA4 BigQuery Export for your own use, we can consume a view of the resulting tables. This view should be a subset relevant to visits (sessions) only, such that it obscures any PII and returns rows only for the session_startevent:

Column Name	Type	Description
event_name	varchar	The GA4 event name; should only be session_start (GA4-supplied)
event_timestamp	integer	GA4 timestamp, as micros since the epoch (GA4-supplied)
user_pseudo_id	varchar	Identifier of the anonymous visitor associated with the visit (GA4-supplied)
user_id	varchar (nullable)	Identifier of the user associated with the visit (GA4-supplied)
batch_page_id	integer	A sequential number assigned to a page that increases for each subsequent page within an engagement (GA4-supplied)
batch_ordering_id	integer	A monotonically increasing number that is incremented each time a network request is sent from a given page (GA4-supplied)
batch_event_index	integer	A number indicating the sequential order of each event within a batch based on their order of occurrence on the device (GA4-supplied)
event_params	record	BigQuery RECORD type that stores nested key-value event parameters, including page_location, `page_referrer`, etc.

Example SQL

select
    event_name
    , event_timestamp
    , user_pseudo_id
    , user_id
    , event_params
    , batch_page_id
    , batch_ordering_id
    , batch_event_index
from
    `analytics_<your_GA4_export_property_id>`.`events_*`
where
    event_name = 'session_start'

For more information on the `<your_GA4_export_property_id>`, see the Google Analytics DIrect Connector Guide.

A Brief Note on Visit Modeling

When provided with Segment or GA4 visit data, Roadway will conduct any sessionization internally, subject to the following rules:

1. A new visit starts whenever a visitor returns with a new page referrer.

2. A visitor returns after more than 30 minutes of inactivity.