An attribution table stores one row per touch: every conversion, the sessions that led to it, and the credit each one earns under five models, so ordinary tiles can answer attribution questions.

What an attribution table is

An attribution table is a composite table built from raw events. For every conversion it finds the visits that came before it, within a lookback window, and precomputes how much credit each visit earns under five models. Because the credit is stored as ordinary metrics, any tile (Table, Bar, Line, KPI) can show it.

Model	Credit
First Touch	All credit to the first visit in the path
Last Touch	All credit to the last visit
Linear	Shared equally across all visits
Position-Based	40% to the first, 40% to the last, 20% shared across the visits between. With two visits, 50% each
Time Decay	More credit to visits closer to the conversion, halving every Time-Decay Half-Life days

This is different from the Attribution tile, which counts users per channel on the fly with a fixed last-touch rule.

Before you start

You need one of:

a GA4 BigQuery export and a Google service account with BigQuery Data Viewer and BigQuery Job User on the project. Queries are billed to that project.
a Lens event table with one row per event, such as Echo events, with a user ID, a session ID and an event name.

Create the table

Go to Configure > Data Tables, click + New Table and choose Attribution Table under Composite Tables.
Under Details, enter a Table Name.
Under Data Source, pick where raw events come from: - GA4 BigQuery Export: pick the Service Account (or the default TagPipes service account), then enter the Project ID and GA4 Export Dataset. Touches use GA4's session source, medium, campaign and channel group. - Lens Event Table: pick the Source Table, then the User ID Field, Session ID Field, Event Name Field and Channel Field. Source Field, Medium Field and Campaign Field are optional. Each session's first event decides its channel, source, medium and campaign.
Under Conversion, enter the exact Conversion Event, for example purchase. Optionally: - Only When: click Add Condition to narrow the conversion, for example fulfillment_type is delivery. - Keep These Conversion Fields: up to 10, comma separated. Each becomes a conv_ column you can filter tiles by, so one table answers delivery, pickup and all orders. - Revenue Field: without it, credit is counted in conversions only. - De-duplicate On: a conversion repeated with the same value counts once, such as a transaction ID.
Click Check Setup (below).
Set the Attribution Rules (below), a Sync Schedule and, optionally, a Notifications group to email when a sync fails.
Click Create Table. The first sync is queued straight away.

Check Setup

Check Setup reads the last 7 days of data and saves nothing. It shows:

Line	Tells you
Conversions, last 7 days	How many conversion events matched. Zero means the table would be empty
Converting users	How many users converted
field filled	For each kept conversion field, how often it has a value
Sessions with no channel	Visits that will read as (not set)
Events GA4 dates on another day than zone	GA4 export only: how many events fall on a different day in your chosen Report Days In zone
BigQuery scanned per sync	GA4 export only: what each build reads, with an estimated cost per sync and per month at the chosen schedule

Warnings appear underneath, for example when a kept field is blank on most conversions. See BigQuery Cost for how the estimate is worked out.

Attribution rules

Setting	Default	What it does
Lookback Window	30 days	How far before a conversion a visit still counts as a touch. 7, 14, 30, 60 or 90 days
Conversions Kept	Last 90 days	The history the table holds: 30, 60, 90, 180 or 365 days. Rebuilt in full on every sync
Direct Visits	Count Them	Count Them gives direct visits credit like any other. Skip When Possible ignores direct when another channel touched
Channel Values That Mean Direct	`Direct`, `(direct)`	Comma separated, case-insensitive. A blank channel reads as (not set)
Time-Decay Half-Life (days)	7	A touch this many days before the conversion gets half the weight of one on the day
Most Touches Per Conversion	50	Keeps the latest ones when a path is longer
Report Days In	Inherit from Property	The day each conversion is counted on. For a GA4 export, use the GA4 property's time zone
Step Names	Source / Medium as recorded	GA4 only. Pick a custom grouping on source_medium, source, medium, campaign or channel to give steps friendly names
Channel Group	GA4 Default	GA4 only. Primary Group uses the GA4 property's own primary channel group; visits it does not cover use the default. Changing it needs a sync

What the table contains

Each row is one touch on one conversion. Visits that did not convert are also stored, so conversion rates can be worked out.

Columns	Examples
Touch	Channel, Source, Medium, Campaign, Source / Medium, Step, Touch #, Touches in Path
Path	Path (Channel), Path (Source / Medium), Path (Campaign), Path (Steps), First-Touch Channel, Last-Touch Channel. A path keeps its last 10 steps
Credit	Conversions and Revenue for each of the five models
Rates	Conv. Rate per model (credit divided by Sessions), Conv. Rate (Visitors)
Totals	Conversions, Revenue, Touches, Sessions, Visitors, Converting Visitors, Avg Days to Convert, Avg Touches per Conversion

Visitors, Converting Visitors and Conv. Rate (Visitors) count one path per visitor and are filled for GA4 export tables; use them with the path columns.

Credit adds up across rows, so a Bar tile of Conversions (Linear) by Channel is a linear attribution report.

Troubleshooting

Symptom	Fix
"Fill in ... first" when checking	A required field is empty. Fill the fields named in the message.
No conversions matched	The event name or a condition does not match exactly. Check the spelling and case.
Many sessions with no channel	The channel field is often blank in the source. Pick another field, or expect a large (not set) row.
Days do not match GA4	Set Report Days In to the GA4 property's time zone.
"Unknown time zone"	Pick a zone from the Report Days In list.

Create an Attribution Table