Kaltura provides various metrics on each media entry's use. This tutorial tells you how to view an individual video's analytics, and what all of the metrics mean. If you're interested in seeing the higher-level analytics, check out our knowledge base article on media gallery / channel analytics.
Any Kaltura user can view analytics for any of their entries, and you can look at analytics for any entry in either Canvas or MediaSpace. Arguably, however, these analytics will be most useful to Kaltura instructors, who share their videos in an authenticated environment where viewers' identities are logged.
If your video is able to be viewed by unauthenticated users, you won't see who they are. This may go without saying, but you'll only see the identity of specific users in the analytics if you're using it within Canvas or in a MediaSpace channel that only allows authenticated users to view it. If, say, you embedded the video on a public site, users' analytics will be tracked using the moniker "Unknown."
You can see the analytics for a Kaltura media entry in either Canvas or MediaSpace. Follow the first two steps depending on where you want to perform the action.
In Canvas | In MediaSpace |
---|---|
|
|
Now you should be looking at the analytics for your video. Read below to learn about what each section means. If it helps, use the list below to jump to a particular section.
This graph can show how many users were viewing the video at particular intervals throughout the video. At a glance, it can give you an indication of what parts of your video are generating the most interest.
In the "views on playback" graph above, views peaked about 81 minutes in when a guest speaker was introduced. At that point there were 106 views (81 unique authenticated views), but that number slowly drops off as the video approaches its conclusion. This graph generally tells us that viewership remained fairly steady throughout the video, increased somewhat with the guest speaker, but slowly declined after his initial introduction.
The "views on playback" graph is one of the first things that you might see on the analytics page for a Kaltura entry. It sits just below a copy of the video itself (which can be played if desired).
Here are the definitions of the terms used in the "Views on Playback" section of Kaltura video analytics.
Metric | Definition | Notes | Question it Addresses |
---|---|---|---|
Player impressions | How many times the video was loaded on a page for that user | The number of player impressions is irrespective of whether or not they clicked "play." This just measures how many times they landed on a page that had the video on it. | How many times did the viewer arrive on a webpage that had this video on it? |
Plays | The number of times the play button was hit | A "play" doesn't relate to how much of the video they watched, though they have to have watched more than just a few seconds to count as one. | How many times did the viewer click play? |
Total completion rate | The total percentage of the video watched, across ALL views | "Total completion rate" doesn't report on the same section of a video twice. So if a user watched the same half of a video twice, the 'Total completion rate' would be 50%. If the user watched the first half of the video one day and the second half of the video another day, the 'total Completion rate' will be 100%. | What percentage of this video did the student actually watch? |
Avg. completion rate | The average percentage of video completion, across all plays | Example: A student watched the video twice. The first time she watched 20% of it, and the second time she watched 80% of it. Her average completion rate would be (20+80) ÷ plays = 50%. | On average, what percentage of the video did the student watch per viewing session? |
"Engagement per user" shows you your individual viewers' viewing behaviors with the video. To view these data, click the "View Engagement Per User" link just below the first graph shown on a video's analytics page.
Clicking the "view engagement per user" link will expand a list of all the users who have viewed the video. Note the screenshot of the graph above demonstrates that despite the high number of unique viewers (16,929), nearly all of them watched this 39-second video from the start to the end.
There are 4 column headers in the "engagement per user" section.
Metric | Definition | Notes | Question it Addresses |
---|---|---|---|
Player impressions |
How many times the video was loaded on a page for that user | The number of player impressions is irrespective of whether or not they clicked "play." This just measures how many times they landed on a page that had the video on it. | How many times did the viewer arrive on a webpage that had this video on it? |
Plays | The number of times the play button was hit | A "play" doesn't relate to how much of the video they watched, though they have to have watched more than just a few seconds to count as one. | How many times did the viewer click play? |
Total completion rate | The total percentage of the video watched, across ALL views | "Total completion rate" doesn't report on the same section of a video twice. So if a user watched the same half of a video twice, the 'Total completion rate' would be 50%. If the user watched the first half of the video one day and the second half of the video another day, the 'total Completion rate' will be 100%. | What percentage of this video did the student actually watch? |
Avg. completion rate | The average percentage of video completion, across all plays | Example: A student watched the video twice. The first time she watched 20% of it, and the second time she watched 80% of it. Her average completion rate would be (20+80) ÷ 2 plays = 50%. | On average, what percentage of the video did the student watch per viewing session? |
The so-called "user engagement heatmap" is visible on a user-by-user basis. If you've clicked "view engagement per user," you should notice a little carat (triangle) next to each user's name in the list. If you click that carat, the engagement heatmap for that user will display below. You can view the heatmaps for multiple users on a page by clicking multiple carats.
The horizontal bar that spans the width of the window is meant to represent your video's timeline, from start to finish. The heatmap shows you what specific parts of the video the student watched, and whether certain parts were watched once, twice, or more than twice. The darker the blue, the more times that piece of the video was watched.
The graph titled "video performance over time" shows you a variety of metrics over the time period you identified at the top of the analytics page.
The "video performance over time" graph in the screenshot above indicates an initial surge of plays in early October just after the video's creation, and then an even higher number of plays in early December. These surges likely correspond to fall and winter quarter viewings of the video.
The "video performance over time" allows you to choose which metric you want to see mapped over the time period you identify at the top of the analytics page.
Metric | Definition | Notes | Question it Addresses |
---|---|---|---|
Player impressions | The number of times the video was loaded on a page, calculated on a daily basis | Impressions don't account for whether or not the user played the video - it's just the number of times a user landed on a page that had the video on it. | On a daily basis during the selected date range, how many times did someone come to a page with this video on it? |
Plays | The number of times the play button was hit, calculated on a daily basis | It doesn't relate to how much of the video they watched. They need to have watched it for 10-15 seconds for it to be counted as a "play," however. | On a daily basis during the selected date range, how many times did users click play on this video? |
Unique viewers | The number of authenticated viewers who watched the video, calculated on a daily basis | If you're only presenting your video in Canvas, chances are that most or all of your viewers will be unique/authenticated. All unauthenticated views will be tracked under the user moniker "Unknown." | During the selected date range, how many viewers watched this video? |
Minutes viewed | The number of minutes of the video were consumed, cumulatively from all viewers, calculated on a daily basis | Across all viewers within the selected date range, how many minutes of the video were viewed? | |
Avg. completion rate | What percentage of the video was completed, averaged across all users; calculated on a daily basis | Example: A student watched the video twice. The first time she watched 20% of it, and the second time she watched 80% of it. Her average completion rate would be (20+80) ÷ 2 plays = 50%. | On average, how much of the video was watched on a particular day within the selected date range? |
Avg. drop off rate | The average quartile of the video when users stopped watching the video, averaged across all users and calculated on a daily basis |
The "quartile" refers to which of the four parts of the video the user was watching when they stopped watching the video (the first, second, third, or fourth quarter of the video). The first quarter is treated as 0%, the second as 25%, the third as 75%, and the fourth as 100% So: Average dropoff rate = (quartile dropoff for first play + quartile dropoff for play 2, etc.) ÷ number of plays Here's an example. Let's say a student watches the first 20% of the video and then later that day watches the first 80% of the video. His average dropoff rate would calculated as follows: (0 + 75) ÷ 2 plays = 37.5% avg. drop off rate For the first view, his watch of 20% of the video means that he stopped in the first quartile, which means 0%, since he didn't reach the 25% "milestone." His second view (80%) is represented in the calculation as 75%, since he passed the 3rd quartile milestone. All users' daily average drop off rates are then themselves averaged to get the final figure. | On average, how much of the video did users watch on a daily basis? |
You can choose to view the graph calculated with monthly data points or daily data points. You can also elect to compare two metrics by using the pull-down menu near the top of the graph.
If you click "view details" beneath the "video performance over time" graph, you can see specific metrics based on three selectable criteria, all within the date range you identified at the top of the page:
The so-called "engagement funnel" attempts to illustrate the relationship between the number of times someone landed on a page with your video on it ("player impressions") and the percentage of video completion. Users can select completion quartiles in a pull-down menu towards the bottom of the illustration.
In the example above, we start at the top of the funnel and move down. Users landed on a page with this video on it 464 times (within the date range selected at the top of the page). Once arriving on the page, users clicked "play" 374 times - 81% of all player impressions. (So 81% of the people who landed on the page actually chose to watch the video.) Lastly, at the bottom of the funnel, there were only 96 times when someone landed on the page and watched at least 75% of the video (which is 21% of all player impressions).
Admittedly, the engagement funnel is a visual representation that may be more useful for marketing. In a way, it tells you a "conversion rate": how many people who saw the video player actually watched it, and how much of it. It might give you insight about whether your video gathered users' attention when they came to its page. In a marketing context, it might inform creative decisions about changing, say, the video length or content to improve the play-through rate.
There are only a few metrics used in the engagement funnel.
Metric | Definition | Notes | Question it Addresses |
---|---|---|---|
Player impressions | The number of times the video was loaded on a page, calculated on a daily basis | "Player impressions" doesn't account for whether or not the user played the video. | How many times did someone come to a page with this video on it? |
Plays | The number of times the play button was hit, calculated on a daily basis | It doesn't relate to how much of the video they watched. Users could have pressed play and then immediately paused the video or closed the window. | How many times did users click play on this video? |
Percent play-through (vod) | How much of the video a user watched, represented in quartiles. |
25% refers to users who watched at least 25% of the video. 50% refers to users who watched at least 50% of the video. 75% refers to users who watched at least 75% of the video. 100% refers to anyone who watched the entire video. | Roughly how much of the video did each user play? |
The "top countries" section of Kaltura video analytics tells you from what countries your video has been watched, based on player impressions (the number of times the video's page was loaded in a browser).
Particularly if you're teaching an online course, this might paint an interesting picture of how your students are distributed throughout the world. The darker the blue, the more the video has been watched in that country.
This section of the analytics tells you what devices students are using to consume your multimedia. This is based on plays, not impressions.
"Top domains" refers to where on the Internet your students are consuming your video, such as Canvas, MediaSpace, or any external website.
If your video is presented exclusively within Canvas, chances are there won't be much of interest to see here. ("Canvaskaf.ucsd.edu" refers to a video being presented within Canvas.) But if your video is presented in multiple places - for example, published to a Canvas course, visible in a MediaSpace channel, embedded on a webpage, etc. - then it may be more interesting to see where most of the traffic is coming from.
Note that depending on how you've distributed your video, some of the information may be difficult to decipher, and entirely unique to your publication situation.
You're able to click any domain in your list of top domains and drill into some data specific to that domain. When looking at specific domain URLs, you're shown two main sets of data: a graph, and a list of specific page URLs where the video can be found.
The graph shown is comparable to the "video performance over time" graph elsewhere on the analytics page You can elect to view plays, player impressions, player impression ratio, minutes viewed, or average completion rate over the time period you selected at the top of the page, all specific to the domain that you clicked on.
In addition to the graph, you'll also be shown various URLs within that domain where the video has been loaded, which may provide some additional clarity on and granularity to the graph above. Icons on each URL's row allow you to visit the webpage, which can be handy in determining where, specifically, people are watching your video.
Note that sometimes the URLs might be a little wonky or not work. Sometimes they refer to a video player that appears in a pop-up window when you're, say, configuring an embedded video. Or perhaps a co-publisher you added to your video published it in his own course, to which you don't have access.
In this example pictured above, we've clicked on the "canvaskaf.ucsd.edu" entry in the "top domains" section. In addition to the graph showing plays over time, three different page URLs are listed. Though difficult for to decipher, they can be clicked on and visited, and they refer to three different video locations: where it's embedded within a Canvas module, the "edit player" (the player on the page after you click the pencil icon within "My Media"), and the normal player page (when an owner clicks on the media to view it within "My Media").
The metrics in the "top domains" section may be familiar to you if you've looked at other analytics sections.
Metric | Definition | Notes | Question it Addresses |
---|---|---|---|
Player impressions | The number of times the video was loaded on a page in the selected domain | "Player impressions" doesn't account for whether or not the user played the video. It's just if they landed on a page with the video on it. | Within this domain, how many times did someone come to a page with this video on it? |
Player Impr. Ratio | The number of times the play button was clicked divided by the number of impressions | Short for "player impressions ratio." This number gives you an impression of how many people were interested enough in your video to click play when they landed on its page. | Within this domain, what proportion of users who arrived at the page actually clicked play? |
Plays | The number of times the play button was hit in the selected domain | It doesn't relate to how much of the video they watched, beyond the 10-15 seconds of watching required to count as a "play." | Within this domain, how many times did users click play on this video? |
Plays distribution | The number of plays with the selected domain divided by the total number of plays across all domains | This can give you an idea of what domain is getting the most views. | What proportion of viewers is coming from this domain? |
Min. viewed (vod) | The total number of minutes of video viewed on that domain | "VOD" stands for "video on demand." This distinction isn't meaningful to us since we don't use any of Kaltura's synchronous features (e.g. Kaltura Live). | How many minutes of video were viewed on this domain? |
Avg. completion rate | How much of the video was watched | If you've got your video posted on multiple domains, this could provide insight as to where viewers are more likely to watch more of your video. | On average, what percentage of the video did viewers watch on this domain? |
Kaltura allows you to filter your analytics search and export the results as a .csv (which is readable by Microsoft Excel). These options are available at the top of the page.
The filtering options are fairly limited and straightforward. In addition to being able to select the date range for your analytics data, you're able to filter by the following criteria:
Be sure to click Apply in order for your filter settings to take effect.
If you click the "export" pull-down menu at the top of the analytics page, you'll have the option to export one or more reports from this video's analytics, including:
Once you click "Export Reports," the system will send you an email that will contain links to each report you've selected. The files will be comma-delimited text files, with the extension .csv. Though these .csv reports can be viewed with any text editor, they're much more legible within Microsoft Excel. If you want to edit or manipulate the exported reports, it's highly recommended that you save the .csv file as an Excel file first (.xlsx extension).