Direct traffic is one of the traffic sources that appear in website statistics. In Google Analytics we’ll see it as (direct) / (none). What does it really mean?
Previous article (part 4): The conversion window (lookback window)
Direct visits
Direct visits mean the user entered a site without the intermediation of other sources, simply by typing the site’s address into the browser or using the bookmarks bar (shortcuts to favorite addresses in the browser).
A direct visit will also occur when a user enters the site:
- from a link placed in a file (a Word, Excel, PDF document, etc.)
- from a link in an email read in a mail client (Outlook, Mail, etc.)
– provided these links weren’t tagged with UTM parameters that allow attribution to other traffic sources (information about tags is available in Google Analytics help).
For this reason it’s worth placing parameters in URLs wherever possible, to distinguish actual direct traffic from visits from mailings or documents promoted online.
Direct visits in Google Analytics can also appear when, for some reason, the referrer (information about the referring site’s domain passed on a link visit), a click parameter (gclid, dclid, etc.), or UTM parameters get removed.
It’s precisely based on this information that Analytics classifies a visit to specific sources (e.g. a visit with the referrer google.com will be classified as google / organic, unless the link contains a parameter indicating it’s an ad (e.g. the gclid click identifier) — then it will be google / cpc).
Such situations can happen, for example:
- when moving from an HTTP page to an HTTPS page,
- as a result of parameters being stripped by the browser or an add-on to it as part of protecting the user’s privacy (adblock/privacy solutions),
- in some cases of moving from a mobile app’s built-in browser to the system browser (“open in Safari”),
- as a result of an (incorrect) redirect performed by the site’s server on a visit.
Direct traffic will appear in every case where source data is missing. In other words, the source (direct) actually means the absence of a source — hence the medium name “(none)”.
Direct traffic means Google Analytics doesn’t know where a given visit came from.
Where does direct traffic come from?
Direct traffic in its pure form comes from a few categories of users:
- Returning users who visit the site after an earlier visit and remembered the site’s URL or saved it as a bookmark;
- People who guessed the address of a specific product’s or brand’s site (e.g. coca-cola.pl);
- People who learned the site’s URL from an offline campaign currently running (e.g. press, radio, TV, outdoor);
- Users who visited the site on someone’s recommendation;
- Returning users who previously visited the site from other sources, but hadn’t visited it for a period exceeding the lookback window.
To a large extent, direct traffic reflects the strength of the brand and the attachment of regular customers.
What does direct traffic tell us?
Direct traffic never arises on its own. It’s always the effect (direct or indirect) of earlier marketing activities.
Direct visits are a reflection of earlier investments in brand awareness
We can influence direct traffic only indirectly: by investing in brand awareness, stimulating loyalty and user recommendations (most simply through a good offer, high quality of products and services, and customer service).
Studying direct-traffic trends can be one of the indicators of changes in brand awareness and interest in it. However, due to the ambiguity of direct traffic described above, a better alternative indicator here may be the volume of brand-name searches in the search engine and visits from that keyword, based on Google Ads and Search Console data.
Direct visits in attribution modeling
Direct is the absence of data, and when we want to study traffic sources it’s better to focus on the data we have. That’s why attribution analyses most often try to avoid assigning value to direct traffic.
Last (non-direct) click
Google Analytics has long avoided assigning traffic sources and key events to (direct) / (none): in traffic acquisition reports, the source is determined using the last non-direct visit (last non-direct click) model. If a user visits the site from Google Search or via a link (medium = referral), or visits the site from a display ad, then the direct visits that follow will keep the existing traffic source, and the user will still be treated in traffic source reports as coming from the original source.
Non-direct attribution models
In attribution modeling too, starting with Google Analytics 4, all attribution models are described as non-direct (as of January 2024), which means that for the purposes of attribution modeling, direct visits are “treated as thin air” and no value is assigned to them.
Assigning key events (and traffic) to (direct) / (none) will only happen when there’s no other interaction on the user’s path besides direct traffic — i.e. there’s effectively no other option and it has to be assigned to (direct) / (none).
For this reason, attribution to direct traffic in Analytics doesn’t depend on the chosen attribution model, and across all models the revenue assigned to direct will be the same.
(Incorrect) sources treated as direct traffic
Sometimes we’d like to treat certain visits as direct visits even though they have a source assigned.
A typical example of such a visit is a user returning from a payment gateway (e.g. a bank’s or payment operator’s page) during a purchase in an online store. This really isn’t a traffic source at all. What’s more, such an entry usually appears on the payment confirmation page, which is often the transaction page — meaning we could assign such a “source” a share in the transaction, even though its role was purely technical.
To prevent such situations, Google Analytics makes it possible to define certain referrals as so-called unwanted referrals. If a referral from such a site initiates a session, the session source will be the source of the user’s earlier visit within the lookback window, and if there was no such visit, the source will be direct. When such a referral occurs during a session, it will be treated as an ordinary page view.
More examples of using this feature in the article Unwanted referral exclusions in Google Analytics you should enable.
Extended direct traffic (big direct)
For a long time now, a significant share of users, instead of typing a site’s address into the browser, search for its name in a search engine and click the link they find there. Most browsers, when you type a site name or an incomplete address into the address bar, will redirect such a user to the default search engine, where they should find our site in the organic results and/or sponsored links.
In practice, such a user doesn’t differ much from a user coming from “classic” direct visits — their intent is the same: to visit a specific site. For this reason it’s justified to treat them the same way. Extended direct traffic therefore consists of users coming from the following sources:
- direct traffic;
- paid search results for keywords containing the company’s name or its website address;
- organic (free) search results for keywords containing the company’s name or its website address;
The concept of extended direct traffic allows us to better attribute revenue to individual non-direct sources and estimate the actual profitability of campaigns.
Direct traffic is only the effect of earlier marketing activities and is generally not treated as a standalone traffic source. By also including in it search traffic related to searching for our own brand, we eliminate the inconsistency in analysis that comes from treating users differently only because they caused the display of the page they intended to visit in a different way.
More about big direct in the article on channel groupings.
Next article (part 6): Comparing attribution models
Worth reading: A guide to attribution in Google Analytics
