fact_treble_sessions

Each row is one execution of a conversational flow by one user — from the moment it starts (a campaign send or an inbound message) to the moment it ends. This is the anchor table of the warehouse: messages, journeys, and variables all join back to it through session_id.

What questions does it answer?

How many conversations ran, inbound vs outbound?
Which flows (poll_name) generate the most conversations?
Where do users stop? (last_node_id, resolvable against dim_poll_nodes)
Which keyword triggered each inbound conversation?
How long do conversations last, and how do they end?

Columns

Column	Type	Description
`session_id`	Int64	Unique identifier of the conversation flow
`company_id`	Int32	Your company (filtered automatically)
`created_at`	DateTime64	When the conversation started
`cellphone`	String	User’s phone number
`country_code`	String	User’s country code
`status`	String	Who handled the conversation: `AI` (stayed with the automated flow), `HumanHandover` (handed over to a human agent), `Rating` (went through an agent and reached the satisfaction-rating step)
`is_valid`	Bool	`true` once at least one message of the conversation was confirmed delivered to or read by the user — the conversation actually reached the phone
`poll_id`	Int32	Flow identifier
`poll_name`	String	Flow name
`channel_id`	Int32	WhatsApp line identifier
`channel_cellphone`	String	WhatsApp line number
`direction`	String	`INBOUND` (user started) or `OUTBOUND` (campaign started)
`contact_id`	Int64	Treble contact; `0` if none
`finished_at`	DateTime64	When it ended; `NULL` if still open
`trigger_keyword`	String	The keyword that triggered an inbound flow, when one matched; empty otherwise and for outbound
`last_node_id`	Int32	The last flow node this conversation reached; `0` if it never reached one
`synced_at`	DateTime64	When this row was last written/corrected

Example queries

Where do conversations end, per flow?

SELECT
    s.last_node_id,
    n.node_type,
    n.question_text,
    count() AS conversations_ended_here
FROM fact_treble_sessions AS s
LEFT JOIN dim_poll_nodes AS n
    ON n.company_id = s.company_id AND n.node_id = s.last_node_id
WHERE s.poll_id = {your_poll_id}
  AND s.created_at >= today() - 30
GROUP BY s.last_node_id, n.node_type, n.question_text
ORDER BY conversations_ended_here DESC

Inbound conversations by trigger keyword

SELECT trigger_keyword, count() AS conversations
FROM fact_treble_sessions
WHERE direction = 'INBOUND'
  AND created_at >= today() - 7
GROUP BY trigger_keyword
ORDER BY conversations DESC

Get Started

Meta

Build with Treble

Main Platform

Agent Platform

Integrations

Analytics Warehouse

Data Warehouse (Legacy)

fact_treble_sessions

fact_treble_sessions

What questions does it answer?

Columns

Example queries

Where do conversations end, per flow?

Inbound conversations by trigger keyword

​fact_treble_sessions

​What questions does it answer?

​Columns

​Example queries

​Where do conversations end, per flow?

​Inbound conversations by trigger keyword

fact_treble_sessions

What questions does it answer?

Columns

Example queries

Where do conversations end, per flow?

Inbound conversations by trigger keyword