Groups of sources in Yandex.Metrica
Each session has its own source. Yandex.Metrica organizes these sources into groups:
- Direct traffic results from users who type your website address directly into their browser or have your site bookmarked.
- Link traffic is derived from users who leave a webpage address as a referrer when visiting your site.
- Search engine traffic is generated when users are directed to your site from a search engine results page. Yandex.Metrica can detect the majority of search engines. The search query is also recorded with the name of the search engine in most cases.
- Traffic from social networks comes from links to the site on social networks. Most social networks identify the user or group that posted the link.
- Ad traffic can be traced for Yandex advertising services (Yandex.Direct, Yandex.Market, or Yandex.Display), media platforms, and for a range of “tags” including UTM, Openstat, and Google AdWords. Detailed information is available in the “Ad systems” report.
- Cached page traffic means that the user clicked through from a page saved locally on the computer (for example, this session will have the referrer file:///С:/Documents/User/page.html), or the entry page domain contains localhost or 127.
Internal traffic is registered if the user's activity on the site is interrupted for a longer period of time than the set session timeout, but the user does not close the browser window. When the user comes back, the new “session” will be recorded as internal traffic.Note.
For accurate detection of internal traffic, make sure that the correct website address is shown in the counter settings (in the Common section).
- Mailing traffic means that the user visited the site from a mailing list.
The Summary shows accumulated data for these groups.
In some cases, the source might be unknown. The string Yandex: Unknown means that some of the session characteristics were not detected, although the session itself was registered in the system.
For example, for sessions originating from Yandex.Direct, this means that the session that Yandex.Metrica registered couldn't be traced to a specific click-through from Yandex.Direct. So besides the advertising system itself, the rest of the data couldn't be detected either: the campaign, keyword, query, and so on.
- The user clicked an ad but left the source page open on a browser tab, and later refreshed it.
- The user clicked a link from an ad, but then forwarded it to someone else. A repeat visit using this link won't be associated with the click from Yandex.Direct, either.
- Redirects on the site may cause yclid tags to disappear from the link and lose the referrer.
There are many reasons why sessions might be “undetected” in Yandex.Metrica. For example, the reason for Yandex: Unknown appearing in a Yandex.Direct source might be a gap in time between generating the yclid tag and registering the session in Yandex.Metrica. This can happen in the following situations:
How Yandex.Metrica detects the traffic source
Yandex.Metrica uses a tag to detect the traffic source if there is a tag in the data for the first pageview of the user's session. If there isn't a tag in the first pageview, the traffic source is detected from the referrer.
For example, a user clicked through to the site from a browser bookmark (the first pageview). Then the user left the site, but came back within 30 minutes by clicking a link with a tag in an email (the second pageview during the session). This session is considered direct traffic to the site.
|Traffic source||How it is detected|
Cached page traffic
Search engine traffic
Traffic from social networks
|Referrer and UTM tags|
|Yandex.Direct traffic||Conditions for linking a counter to a Yandex.Direct campaign|
|Traffic from other ad systems||Referrer and tags. Full data is collected only if UTM and Openstat tags are used|
|Other advertising: identified by tags||
UTM tag. The URL must contain the utm_medium parameter with one of the following values: cpc, ppc, paidsearch, cpv, cpa, cpp, display, cpm, banner.
The utm_source parameter can be omitted or can have a value that is not in the list.
Questions and answers about traffic sources
- Why isn't data being collected for ad systems?
- Why is there a lot of direct traffic to the site?
- How can I find out where internal traffic is coming from?
- Why do my Yandex.Metrica reports show a lot of traffic with the source Unknown?
- Why isn't traffic from certain sites registered?
Why isn't data being collected for ad systems?
In order for Yandex.Metrica reports to show information about Yandex.Direct campaigns, you need to take these steps in Yandex.Direct on the Campaign parameters page:
- Enter the counter number in the Metrica counters box.
- Keep the Tag links for Yandex.Metrica option enabled (URLs from Yandex.Direct ads will have the yclid=... tag appended to them).
To collect data from other ad systems, you must use tags.
Why is there a lot of direct traffic to the site?
- This traffic could have been generated by robots. You can see information about robot sessions in the Robots report under the Monitoring group. If you have discovered surges of unnatural traffic in the Robots report and you have reason to believe that this situation will reoccur, we recommend enabling robot filtering using strict filters and behavior.
- It's possible that this traffic came from sites that use the HTTPS protocol, so the referrer wan't transmitted. In this case, we recommend using the utm_referrer tag.
How can I find out where internal traffic is coming from?
Formally, a session is labeled as Internal traffic when the session begins with a click-through from a page on the domain that is specified as the site's domain or additional domain in the counter settings. For example, the counter settings list the site address http://mysite.example/. The session starts from the page http://mysite.example/page1.html, and the click wasn't made from an external source, but from a page on the same site, such as http://mysite.example/contacts.html (in other words, the page contacts.html was identified as the referrer when going to page1.html).
There are several reasons why this situation may occur:
The user navigated to a page (contacts.html in this case), but didn't perform any actions on the site for the duration of the time set as the session timeout in the counter settings (by default, this is 30 minutes). The user's next click-through on the site (to page1.html) is a new session, but the source is the same site.
One possibility is to increase the session timeout. On the other hand, first you need to determine the usage patterns for your site. Let's look at an example with email (Yandex.Mail).
The user logged in, checked email, and left the page open in the browser. An hour later, the user came back and checked email again, then replied to messages. In this case, it wouldn't be correct to combine all the user's actions in a single session. There really were multiple sessions on the site. So the presence of internal traffic is an indicator of the site's usage patterns.
A user followed a link from an external source to a page that doesn't have a counter (for instance, contacts.html), then immediately (during the same session) navigated to a page with a counter (in this case, page1.html). This means that two conditions are met: the session starts from page1.html (because Yandex.Metrica doesn't have data about the contacts.html pageview due to the absence of the counter on this page), but the page the user navigated from (contacts.html) belongs to the same site that is specified in the counter settings (http://mysite.example/).
We recommend installing the Yandex.Metrica counter on the page. In this case, it may be difficult to tell which page is missing the counter. You can use segmentation to help you find it.
Why do my Yandex.Metrica reports show a lot of traffic with the source “Unknown”?
This situation usually occurs if the counter code is installed incorrectly (the part of the code that should be inside the
<noscript> tag is located outside of this tag).
Why isn't traffic from certain sites registered?
More and more sites are switching to secure HTTPS connections. At the same time, those sites that haven't switched to secure connections yet have a growing percentage of direct traffic. From a technical point of view, direct traffic represents sessions on your site when the user's browser didn't pass the HTTP Referer header. This header normally contains the URL of the previous page that the user navigated to your site from. If this field is missing, it means there wasn't a previous page. For example, the user entered the URL in the browser's address bar.
The W3C (World Wide Web Consortium) has published a Referrer Policy, which contains recommendations on transmitting this header: the referrer should be transmitted when navigating from HTTP to any type of site, and from HTTPS to HTTPS, but the referrer isn't sent when going from HTTPS to HTTP. This means that if you leave the secure loop of HTTPS sites, the browser doesn't openly transmit the URL of the last page you visited in unencrypted format.
In practice, this means the following for site owners: if your site uses HTTP, you won't see the traffic source if this source uses a secured HTTPS connection.
We recommend switching your site to a secured connection. You can get an encryption certificate for free through the Let's Encrypt project, which is a non-profit certification center supported by many prominent representatives of the internet industry.
If you own a secured internet resource but you want the Referer header to always be transmitted for traffic from your site even when going to unsecured sites, you can put special markup on the site's pages that link to unsecured resources. For more information about how to mark pages, refer to the “Referrer Policy” document (see the section Referrer Policy Delivery). According to the data of the Can I use project, this link markup is already supported by over half of the browsers in use.