What is direct traffic in Google Analytics 4, and how to reduce it

What is direct traffic in Google Analytics 4

Google’s definition of the channel: “Direct is the channel by which users arrive at your site/app via a saved link or by entering your URL.” The rule behind it is mechanical. A session goes to Direct when its source exactly matches (direct) and its medium is (none) or (not set) (default channel group, checked October 2026).

GA4 writes (direct) when it has nothing else. It looks for campaign parameters in the URL, such as UTM tags or a Google Ads click ID, and for a referrer sent by the browser. When both are missing, the session is direct, whatever the visitor actually did.

One detail softens the number in some reports. For session-scoped dimensions Google states that “sessions initiated by direct entrance are attributed to the UTM values for that user”: a returning visitor who comes back directly can be credited to their earlier campaign. Direct is the answer of last resort.

Where to find direct traffic in GA4

Open Reports, then Acquisition, then Traffic acquisition. The default dimension is Session default channel group, or Session primary channel group if your property has set a custom channel group as primary. Direct is one of its rows. Add Session source / medium as a second dimension to see (direct) / (none).

The User acquisition report shows the same for the first visit of each user, under First user default channel group.

Which pages receive direct traffic

The landing page tells you whether a direct session can be real. People type a home page address. Nobody types the URL of a filtered category or a blog post with a long slug.

1. In the Traffic acquisition report, type direct into the search box above the table.
2. Click the plus sign next to the dimension and add Landing page + query string.
3. Sort by sessions.

Direct sessions that land on deep pages are mislabelled traffic. That list is your work list. Rows where the landing page is (not set) are a separate problem, covered in what (not set) means in GA4.

What causes direct traffic

Only the first row is traffic you cannot label. Every other row can be reduced.

How to reduce direct traffic

1. Tag every link you control. Email campaigns, newsletter footers, social posts, QR codes, links in PDFs, partner placements. A UTM-tagged URL carries its source in the address, so it does not depend on the referrer. Agree on one naming scheme first; the rules are in the guide to UTM tags.

2. Use short, tagged URLs for offline media. A QR code or a printed address should redirect to a URL with UTM parameters. Check that the redirect passes them through.

3. Test your redirects. Open a tagged URL that redirects and look at the address bar after the page loads. If the UTM parameters are gone, the redirect rule drops the query string. This is a server setting to fix once.

4. Serve every page over HTTPS. If any landing page still answers on HTTP, links from secure sites arrive without a referrer.

5. Put the GA4 tag on every page. Landing pages built in a separate tool, checkout subdomains and old sections of a site are where the tag is usually missing. The setup is described in GA4 tags in Google Tag Manager.

6. Exclude your own team. Employees open the site from bookmarks all day. Define internal traffic under Admin, Data streams, Configure tag settings, and activate the data filter.

Check the effect a few weeks later in the landing page view above. A fix shows up as fewer direct sessions on deep pages, not as a smaller Direct total overnight.

Where GA4 stops

Three limits remain after the fixes:

• GA4 reports the label, not the evidence. You see that a session is direct. You do not see whether it had a referrer GA4 ignored, whether it started without a page view, or what the same visitor did the week before.
• The channel rules are Google’s. You can build a custom channel group, but the logic that turns missing data into (direct) is not yours to change.
• Past sessions stay as they were recorded. Tagging from today does not relabel last quarter.

For a marketer this matters at budget time. If a third of sessions sit in Direct, every paid channel looks less effective than it is, and the difference cannot be reassigned inside the interface.

Check the raw sessions in BigQuery

GA4’s native BigQuery export contains every event with its page URL, its referrer and its campaign parameters. That is enough to count, per landing page, the sessions that arrived with no referrer and no campaign source. The query runs as written on Google’s public GA4 sample:

WITH sessions AS (
  SELECT
    user_pseudo_id,
    (SELECT value.int_value
     FROM UNNEST(event_params)
     WHERE key = 'ga_session_id') AS session_id,
    ARRAY_AGG(
      (SELECT value.string_value
       FROM UNNEST(event_params)
       WHERE key = 'page_location')
      IGNORE NULLS
      ORDER BY event_timestamp
      LIMIT 1)[SAFE_OFFSET(0)] AS landing_page,
    LOGICAL_OR((
      SELECT value.string_value
      FROM UNNEST(event_params)
      WHERE key = 'page_referrer') IS NOT NULL)
      AS has_referrer,
    LOGICAL_OR((
      SELECT value.string_value
      FROM UNNEST(event_params)
      WHERE key = 'source') IS NOT NULL)
      AS has_campaign_source
  FROM
    `bigquery-public-data.ga4_obfuscated_sample_ecommerce.events_*`
  WHERE _TABLE_SUFFIX BETWEEN '20210101' AND '20210131'
  GROUP BY user_pseudo_id, session_id
)
SELECT
  REGEXP_EXTRACT(
    landing_page, r'^https?://[^/]+(/[^?#]*)'
  ) AS landing_path,
  COUNT(*) AS sessions,
  COUNTIF(NOT has_referrer
          AND NOT has_campaign_source)
    AS no_origin_sessions
FROM sessions
GROUP BY landing_path
HAVING sessions >= 500
ORDER BY no_origin_sessions DESC

The first rows on the sample:

A quarter of home page sessions arriving with no origin is plausible. Six in ten sessions on an apparel category page arriving with no origin is not people typing URLs. Something links to that page without tags, and now you know which page to investigate.

This is a diagnostic, not a copy of GA4’s channel logic, so the totals will not equal the Direct row in the interface. Linking a property to BigQuery is covered in connecting GA4 to BigQuery.

Make the check part of the weekly report

The query answers the question once. The marketer’s real need is to see it every week next to spend: which landing pages lose their source, and whether last week’s tagging fix worked.

OWOX Data Marts is built for that step. An analyst saves the query as a Data Mart on the warehouse that holds the GA4 export, describes what “no origin” means in it, and publishes it. A report on the Data Mart is delivered to Google Sheets or Data Studio (the former Looker Studio) on a schedule. The marketer opens a sheet, not a SQL console.

OWOX Data Marts does not collect GA4 data and does not reassign direct sessions to channels. It publishes the analyst’s SQL on the tables the export already wrote. If the same warehouse also holds ad spend, the two can be joined there, as shown in blending GA4 conversions with ad spend.

GA47 analytics tactics to get more traffic and sales using GA4April 30, 2024→GA4E-commerce tracking in Google Analytics 4 with Google Tag ManagerOctober 7, 2026→Data TrackingFirst-party, second-party and third-party data explainedDecember 13, 2024→

See all articles →