Default channels

The channels every dashboard starts with, from Direct and Organic to Unknown Source and Raw Traffic, and what lands in each

A channel is a top-level category of traffic on the dashboard: Organic, Direct Traffic, Google Ads. Each visit lands in one of them by where it came from, and the conversions and revenue that follow are credited to it. How channels are built from filters, and how to make your own, is covered in Channels, filters & groups.

Every dashboard starts with a set of default channels. You can see how any of them is configured, and change it, by clicking the settings icon next to it.

All Traffic

The first row of the dashboard. All Traffic is the sum of every attributed source: the visits in the date range, and the conversions and revenue the model credited to them, across every filter on your dashboard. It is the total after modeling, so it follows your attribution type, model and traffic options. For the exact conversion events and revenue captured in the selected date range, before any modeling, see Raw Traffic at the bottom of the dashboard.

Ad channels

Your paid channels appear as soon as you connect an ad platform. The most popular platforms, Google Ads, Meta Ads, Microsoft Advertising, LinkedIn and others, need no configuration: once connected, Attribution starts tracking their clicks, creates the channel with a filter per campaign as they exist on the platform, and pulls the spend every day so cost sits next to the conversions and revenue each campaign earned. See Ad integrations for what each platform tracks.

What every project starts with

A new project comes with channels for the common ways people reach a site without clicking an ad. A visit lands in one of them when it comes from that kind of site and carries no advertising parameters:

  • Organic: search engines such as Google, Bing, Yahoo, DuckDuckGo and Yandex, including Google Shopping listings.
  • Social: social networks such as Facebook, Instagram, LinkedIn, X, TikTok, YouTube, Reddit, Pinterest and Threads.
  • Email: clicks from webmail such as Gmail and Outlook. A link in an email you send should still carry UTM parameters, so it is credited to your own email campaign rather than landing here.
  • AI Assistants: visits referred by assistants such as ChatGPT, Claude, Gemini and Perplexity.
  • Messaging: links opened from WhatsApp, Telegram, Slack, Messenger and Microsoft Teams.

The lists behind these channels are maintained by Attribution and grow as new services appear. Click the settings icon on any of them to see exactly what it matches, or to add a site of your own.

The remaining default rows hold what matches nothing above, and the raw totals. They are described below.

Direct Traffic

A visit lands in Direct Traffic when it arrives with no referrer and no tracking parameters in the URL: nothing says where the visitor came from. The most common ways that happens, roughly in order:

  1. Type-in traffic and bookmarks. Someone types your URL or opens a bookmark. Neither carries a referrer, and this is the only truly "direct" way to reach you.
  2. Links from mobile apps. A link opened from an app, the Facebook or Instagram feed, a messaging app, an email client on a phone, usually carries no referrer.
  3. Email links without tracking parameters. Without parameters an email click cannot be traced back to the email. Always add tracking parameters to the links in your emails.
  4. Links in documents. A link in a PDF, a Word document or a presentation carries no referrer, so the visit is direct unless the link has tracking parameters.
  5. Secure to non-secure linking. If your site is on http and the visitor comes from an https site, the browser drops the referrer and the visit looks direct.
  6. A landing page without the snippet. If the page the visitor lands on does not carry Attribution's tracking and the first tracked page view happens on the second page, the referrer is your own site and the visit looks direct. A noticeable share of such self-referred direct visits usually means some landing pages are missing the snippet.

The Breakdown Direct Traffic button at the bottom of the dashboard splits the Direct row into these cases, which is the quickest way to see which of them you have. Adding tracking parameters to every link you control, and the snippet to every page, removes most direct traffic that should not be direct.

Whether direct visits receive credit at all is a view setting: the traffic options decide if direct traffic is included in the model, excluded, or included only until a cutoff event such as a signup. See Traffic options for the four options and who each suits.

Unsegmented Traffic

A visit that matches none of your filters lands in Unsegmented Traffic, and so do the conversions credited to it. Unsegmented traffic comes in two kinds:

Tagged Traffic is any visit carrying a parameter for which no filter exists yet. Clicking the Tagged Traffic link takes you to the Dashboard Builder, where you can create the filters that will then appear on your dashboard.

Referral Traffic is any visit with a referring URL that no filter covers yet. Click the Referral Traffic link to create filters for these visits.

Unknown Source

Unknown Source holds the conversions that cannot be credited to any visit. A conversion is credited to the visits before it; when an event arrives with no preceding visit for that user, there is no touchpoint to credit, and the event lands here. If Attribution is installed properly there is always a visit before a conversion, so a sizeable Unknown Source is a sign to investigate. Some Unknown Source is normal; a higher share, above 5% or so, usually points to an installation issue.

Common causes:

  • Users who predate your install. When you connect Attribution, your site or your integrations may already send events for existing users who have never visited with the snippet installed. This is common for users who rarely visit but whose events keep firing, usually from the server side (a monthly subscription, for instance).

  • A conversion integration without identified visitors. Leads, contacts or deals created in HubSpot, Pipedrive, Salesforce or Stripe can only be tied to a visit if the person was identified on your site with an email trait beforehand. Records imported or created by hand have no visit history.

  • Tracking the event before the page view. Sometimes your code looks like this:

    Attribution.track('Landing Page Event');
    Attribution.page();

    Technically correct, but the event fires a fraction of a second before the page view, so for a first-time visitor the event arrives with no visit to credit. Call page() before track():

    Attribution.page();
    Attribution.track('Landing Page Event');
  • Ad blockers. Most ad blockers block Attribution (or Segment) page views and identify calls but not server-side events, so those events arrive without a visit.

  • Other causes, mostly technical or an incorrectly set up snippet. If a significant share of events lands in Unknown Source, contact us to verify your installation.

Raw Traffic

Raw Traffic is not shown by default; the Show Raw data button at the bottom of the dashboard adds it as the last row. It is everything Attribution received in the selected date range before any modeling: every visit, every conversion event and all revenue dated inside the range, whatever source they came from and whether or not a visit could be credited. It ignores the attribution type, the model and the traffic options, and keeps only your view's conversion event and revenue group.

Raw Traffic is mostly used for debugging and validation. Modeled numbers cannot be compared with anything outside Attribution, because no other tool splits credit the same way; raw counts can. If the conversion events and revenue in Raw Traffic match your store, CRM or billing system for the same dates, the data is arriving correctly, and any difference in All Traffic comes from the model, not from missing data.

Raw Traffic and All Traffic differ by design. All Traffic is modeled: under Marketer Spend View it counts conversions that happened after the range because the visits that earned them are inside it, and traffic options can keep some visits from receiving credit. Raw Traffic counts by date only. When the two disagree, the gap is the effect of your view's settings, not missing data.