# Attribution dashboard **Attribution: which channel gets credit?** follows each conversion back through the visitor's journey of the 90 days before it, and shows how first touch, linear and last touch share the credit between channels. Use it to see which channels open journeys that others close, and what each model makes a paid campaign's conversion cost. [Attribution modelling](/docs/analytics/attribution-modelling) explains journeys, touchpoints and the models. ## Good to know - Two switches at the top, **Organic** and **Direct/Internal**, keep those channels in every journey or leave them out, for every tile at once. They open with Organic included and Direct/Internal excluded. - There is no Channel filter on purpose: filtering a journey to one channel would hand that channel all the credit. - A journey is one device's, and Safari clears the device cookie after 7 days without a visit, so a journey with a longer gap splits in two. ## At a glance How each channel's credit for a conversion moves between first touch, linear and last touch, and what each model makes a paid campaign's conversion cost; which channels open journeys that others close; and the paths of channels journeys take to a conversion. - **Tabs**: 3, Credit, Assists and Paths - **Charts and numbers**: 27 - **Built on**: the [Attribution](/docs/analytics/attribution) explore, with some tiles on [Campaigns](/docs/analytics/campaigns), as each says ## Filters The filters across the top apply to every tile they can, unless a tile says it has its own range. - **Date range**: the last 30 days, today included. - **Conversion**: not set when the dashboard opens, so it keeps everything until you pick a value. It does not apply to 1 tile on the Credit tab. - **Device type**: not set when the dashboard opens, so it keeps everything until you pick a value. It does not apply to 1 tile on the Credit tab. ## Tabs ### Credit ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | Conversions | Number | Purchases, leads, demo requests, newsletter signups and trial starts in range, each kind once a visit: what every model on this dashboard hands out. | | Revenue | Number | The revenue of the purchases in range, each once, in the order's own currency. | | Multi-visit journeys | Number | The share of conversions whose journey's touchpoints were in more than one visit in the 90 days up to it. | | Models disagree on | Number | The share of conversions first touch and last touch credit to different channels. | | Visits per journey | Number | Visits a conversion's journey's touchpoints were in, averaged over conversions. | | Channels per journey | Number | Channels a conversion's journey came by, Direct and Internal included, averaged over conversions. | | Touchpoints per journey | Number | Touchpoints in a conversion's journey, its visits' entrances and the campaigns that reached them mid-visit, averaged over conversions. | | Credit to mid-visit touches | Number | The share of conversions last touch credits to a campaign that reached the visitor mid-visit, such as a retargeting ad opened in a new tab, which visit-based attribution gave to how the visit arrived. | | Credited by fallback | Number | The share of conversions whose journey the Organic and Direct/Internal switches left nothing of, credited to their own visit's entrance whatever its channel. | :: #### How does the credit move between models? ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | Share of conversions credited to each channel | Dot plot | Each channel's share of the conversions under first touch, linear and last touch, the channel with the most linear credit on top. | | Model comparison | Table | Each channel's share of the credit under first touch, linear and last touch, and how many points last touch moves it from first touch. | :: #### Which sources and campaigns get the credit? ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | Sources and campaigns under each model | Table | Each source / medium and campaign: the conversions it took part in, the conversions each model credits it with, and its linear revenue. | :: #### What does a conversion cost under each model? All kinds, all devices ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | Paid campaigns under each model | Table | Every campaign with uploaded spend: its spend, and the conversions, cost per conversion and ROAS it is credited with under first touch, linear and last touch, within 90 days. From the [Campaigns](/docs/analytics/campaigns) explore. | :: **How to read this dashboard** A **journey** is the **touchpoints** of one device in the **90 days** up to a conversion: how each of its visits arrived, and every campaign that reached it later in a visit, such as a retargeting ad opened in a new tab, up to the conversion. A visit keeps the channel it arrived by everywhere else. A **conversion** is a purchase, a lead, a demo request, a newsletter signup or a trial start, each kind once a visit, as the Conversion dashboard counts them, dated on the day of the visit it happened in. The **Organic** switch (Organic Search, Organic Social, Referral and AI Assistants) and the **Direct/Internal** switch keep those channels in every journey or leave them out of it, path and credit alike. They open on Organic Included and Direct/Internal Excluded, which credits as Kontala always has; a journey they leave nothing of is credited to its own visit, whatever its channel. **First touch** gives each conversion to the first touchpoint of its journey, **last touch** to the last, and **linear** shares it equally between all of them. Every model hands out each conversion exactly once, and last touch is how Acquisition's Campaigns tab credits campaigns. **Took part in** counts the conversions a channel or source had a touchpoint in, so it adds up past the total. The **paid campaigns** table counts every kind of conversion on every device, so the Conversion and Device type filters do not reach it, and matches spend to visits as Acquisition does. There is no Channel filter, because it would cut journeys apart. Bot traffic, visitors who declined consent and origins the property does not allow are left out. A device is a first-party cookie, which Safari clears after 7 days without a visit, so there a long journey splits in two and first touch reads late. ### Assists ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | Assisted conversions | Number | Conversions whose journey had a touchpoint before the last one. | | Share with an assist | Number | The share of conversions whose journey had more than one touchpoint. | | Assisted revenue | Number | The revenue of the purchases whose journey had more than one touchpoint, in the order's own currency. | | Handed to another channel | Number | Conversions whose journey one channel opened and another closed. | :: #### Who opens journeys, and who closes them? ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | The part each channel plays in converting journeys | Stacked bars | Each channel's touchpoints in converting journeys, as shares: opening a journey, mid-journey, closing it, or its only touchpoint. | | Who opened, and who closed | Flow | Conversions with more than one touchpoint, flowing from the channel of the first to the channel of the last. | :: #### Which channels assist, and which close? ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | Assisted conversions by channel | Table | Each channel's conversions taken part in, assisted, closed by last touch and opened by first touch, its assists per close and its linear revenue. | :: **About assists** A touchpoint **opens** a journey when it is the first of several, **closes** it when it is the last, and is **mid-journey** between them; a journey with one touchpoint has it as its **only touch**. An **assisted conversion** is one a channel had a touchpoint in before the last, as Universal Analytics counted assists, so a channel can assist a conversion and close it too. **Assists per close** divides a channel's assisted conversions by its last-touch ones: above 1 it mostly opens and carries journeys, below 1 it mostly closes them. A channel the switches leave out plays no part, unless a journey had nothing else, as its only touch. ### Paths ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | Distinct paths | Number | How many different channel paths the conversions in range took. | | Converted on the first visit | Number | The share of conversions on their journey's only visit in the 90 days up to them. | | Median days to convert | Number | How long a journey typically took, from its first visit to the visit it converted in. | | Journeys of 5+ visits | Number | The share of conversions whose journey had five visits or more. | :: #### How do journeys unfold? ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | Journeys of more than one step, step by step | Flow | Converting journeys with more than one step, flowing from each step's channel to the next: steps one to three, then every later step folded into the channel the journey ended on. | :: #### Which paths convert? ::field-table | Tile | Kind | What it shows | | --- | --- | --- | | Top conversion paths | Table | Each channel path the conversions took, with its conversions, their share and revenue, and how many visits and days its journeys took. | | Visits in a converting journey | Column chart | The share of conversions whose journey had 1 visit, 2, 3-4, 5-9 or 10 or more in the 90 days up to them. | :: **About paths** A **path** is a journey's channels in the order its touchpoints came, those the Organic and Direct/Internal switches keep. Consecutive touchpoints from one channel are one **step**, such as Email ×2, and a path of more than five steps shows its first two and its last two around how many it leaves out. The flow numbers a journey's first three steps and folds every later one into the channel the journey ended on, as **4+**. **Days to convert** run from the journey's first touchpoint to the visit it converted in, at most 90. A journey is one device's, and Safari clears the device cookie after 7 days without a visit, so long journeys read short there.