Advanced metrics#

Our metrics provide you with detailed information on how many unique users have viewed, hovered and clicked your images on and in all places where your images have been embedded. Additionally we measure the time users spend interacting with your images. You can view individual image statistics down to how many people hovered and clicked a specific image tag. You can also view all image metrics combined in one timeline, incorporating data from all the interactive images you have created.

Views, Hovers, Clicks#

We measure a "view" whenever the tags are displayed, on top of your image, to the user. A "hover" is measured whenever the user moves the mouse on top of an icon (and the associated hover area) so that the tag bubble shows up. A "click" is measured whenever the user clicks on the tag icon, the hover area, or the bubble. Note that some tag types - because they contain embedded content from other sites - cannot reliably measure clicks in all cases.

On mobile, the initial touch on the icon which reveals the tag bubble is counted as a hover. The second touch then becomes the click, since it activates the content.

We perform low-pass filtering on the hovers, so that multiple, repeated hovers within a short time period on the same tag are only counted as one hover.

Hover and click rates are calculated as the amount of hovers and clicks compared to views - either uniques or total.

NOTE that, if you are logged-in, the views, hovers and clicks on you own images are not counted.


We calculate uniques users using a combination of cookies and browser fingerprinting, using a 30 day rolling interval. We do honor the Do-Not-Track -setting of most browsers, in which case we store the cookies only during the session.

Most statistics are calculated either as total or per unique user, and you can choose which you view using the toggle on the dashboard.

Dwell rate#

Dwell rate is calculated as the portion of users who hovered for at least one second over the image (or, in the case of mobile, interacted with the image in any way).

Time on image#

The "time on image" measures the time the users spent interacting with the image on the average. Note that because it's not always possible to get the "hover out" time, this only represents a sample of the users, and with very low view counts it may not be entirely accurate.

NOTE that, time on image can be calculated as
- The average of each daily figure.
- The all time average
This is why you may get slight differences in the time on image stats.

Interaction time#

By tag "interaction time" we mean measuring the time during which the tag is open - starting from when the user enters the hover area (or taps on it on a mobile device), and ending when the tag is closed.
Note that the closing of the tag is not always available (e.g. when the user switches to a different site), so this represents a sampling from the users, and may be inaccurate with low hover counts.


We keep track of all the places where the image is embedded and provide you with the number of embeds, when they were first detected (starting from the date someone viewed the embedded image the first time), and view/hover/clicks for those places.

We do filter known proxies, Google Translate, development and internal sites out of this list, which is why the embed view counts don't necessarily match up with the total view counts. (If you'd like to add your internal development site on our filter list, let us know.)

Currently we're filtering at least the following embeds (but not limited to; this is about half of the list):

  • translate.*
  • webmail.*
  • preview*
  • staging*
  • testing*
  • dev.www*
  • *.dev
  • *.local
  • *.gnl

Question: Why aren't my embeds showing correctly?#

When you embed an image or slideshow inside an iframe (that is, you use our iframe embed code), that iframe comes from - so on that page, we need to rely on the "Referer" header sent by the browser itself to figure out where the iframe is embedded. If this information is missing, it may cause the view/hover/etc counts to be counted against instead of your own site.

Unfortunately there are quite a few situations where the Referer header will not be sent. For example, if you are constructing the iframe dynamically (as is quite typical in ad operations), or you're wrapping multiple iframes inside each other. Sometimes even browsers are configured not to send the Referer-header!

In such a case, you can pass a "referer" parameter to the iframe - for example

<iframe width="640" height="890" 
        type="text/html" frameborder="0" 
        webkitallowfullscreen mozallowfullscreen allowfullscreen scrolling="no"></iframe>

This also works with embedded channels in an equivalent fashion. The URL should be URL-encoded.

Note that the referer-parameter is only used when the browser does not pass the Referer-header.

Further information#

See also our premium features presentation.