Start learning with our library of video tutorials taught by experts. Get started
Viewed by members. in countries. members currently watching.
In Google Analytics Essential Training, Corey Koberg shows how to use the Google web analytics platform to generate and evaluate information about the visitors to a web site, including data on site traffic, user behavior, and marketing effectiveness. This course covers the out-of-the-box functionality, from account creation to reporting fundamentals, and explains how to glean insights from the vast array of data available.
The real point of web analytics is not just to collect data; it's to get insights. And through a feature called Intelligence, Google Analytics is going to help make analysis of our site for easier and help us draw accurate conclusions faster. To get to the intelligence reports, we navigate by clicking over here on the Home tab, click INTELLIGENCE EVENTS, and we have the option of the Overview, Daily Events, Weekly Events, or Month Events. For now let's just look into Daily Events. There are two different ways that I use this tool. The first I'll call forensic and the second we'll call insights.
In forensic mode here what I am really looking for are things that are jumping out for me but I can't necessarily explain it. As I look at this graph right here of visits, certainly this day jumps out. There is a large jump in visits here and I want to figure out what that is. That's not necessarily a particularly easy thing to do. Often what we are trying to do here in analysis is not things that couldn't ever possibly be figured out another way, but ways that we can do things faster, easier, quicker. We only have a certain amount of time in our day for analysis; we need to be as efficient as possible, so I need to figure out quickly what's going on here.
As I put my mouse over this date, I can see the number of visits have gone up, and down below you will see this green bar. This green bar indicates the number of intelligence alerts that have been detected on that particular day. In our case we have three alerts. If I click on this green bar, the bottom of the screen is going to update to show what those alerts that Google Analytics found for us. So there are three things of interest here. First is that page views have gone up, visits have gone up, and visits particularly from the source of reddit have gone up. If I look over here I will see this idea of Importance.
Importance is also known as significance, and the idea here is this is going to tell us how different this is from what was expected. In our case, there was a large increase in the number of pages views and visits. Normally we would expect to get between 0 and 275 visits from Reddit. In this case we had 2500 visits on that day-- over 500% increase. If I click on the little icon right here, I will actually isolate the graph above to show just that. In this case it becomes very obvious, yes, this is a major event that happened on that date.
When we look at just the visits here from reddit.com, we can see that this was something way out of the ordinary. If I click back to our original screen, we see that that bump in 2500 visits is a large contributor to our overall visits here. It's a quick way for us to understand what's happening when we see something out of the ordinary. But I think the real value of this tool isn't so much in explaining what's already obvious to us, but in uncovering insights that we may never have seen before. If we are diligent analysts, we might log in every single day, we might go through our list of hundreds of different reports, and we might analyze every single type of medium, every single type of source, every keyword, every campaign that we are running, all the different traffic from different areas of the world, different cities, different states, different countries, and we might look for all these little anomalies and differences.
When we clicked on reddit.com, it became apparent that there was something strange that had happened there, but we wouldn't necessarily know to go click on visits from reddit.com that day or click every other source that brought us traffic that day. This isn't something that we as human analysts are perfectly good at is going through report after report after report. However, this is the perfect job for a computer, to churn through all of these reports every day, looking for something out of the ordinary and then alerting us when that happens. The problem is, if we are writing the computer program to do that, how are we going to tell the computer to sift through all this data and alert us? We could do it by quantity and we could say when a certain change in quantity happens to alert us.
The problem there is if we are thinking it's something like page views, right down here we see there are 20,000 page views on one, other pages on our site may only get a couple hundred or even a couple of dozen pages. So if we set a page-view limit of let's say an increase of a hundred pages, we probably are going to alert every single day for certain pages that get thousands of page views. Or we may never get an alert for a page that only gets a couple of dozen but it goes up to a hundred and that would be significant for that particular page. The other option is we could look at things via percentage.
The problem there is things that have a low quantity are going to have high percentages. For example, if we are used to getting one conversion and all of a sudden that goes to three conversions, percentage- wise that's going to be a very large jump. But in reality I don't necessarily want to get alerted when three things sell instead of one thing. That's not going to be significant to me in the overall case of my business where I'm selling thousands of items per day. So percentages can be problematic as well. What we are really trying to say, in English, is I want to be noted when something significant happens. So what we are really looking at is something that's different from the expected.
The way Google Analytics is going to look at this is through essentially standard deviations. It's going to look and see what was expected, and then it's going to look at what was actually happening, and it's going to look at how different those things were. That's what we have over here in the gray bar of Importance. When things are significantly different from what was expected, it's going to be more important. So this algorithm isn't based purely on quantity and it's not purely based on percentage either. It's closer to what you might think of a standard deviation. The key to this is that it's going to stop us from getting false positive predictions, because as analysts if day after day we get false positive predictions of things that are supposed to be important or supposed to be significant but actually aren't, we are going to start ignoring those.
In fact, we have the ability to control how much of those we get or don't get by the slider up here where we can say we want the alert importance to be low or high. If I select over here to low then I'm going to see more alerts here that don't necessarily meet a high threshold of importance. If I say, listen, I am very busy today, I only want to see the high level alerts, then I can move this over to high. I'm going to get far fewer alerts, but the alerts that I do get will be of very high importance. In this case, we can see things that were predicted to be in a certain range but their actual was very far away from that.
So in this case we expected to have 11% to 12%; instead I got a 24. In this case 3.7 to 4.7; instead I got a 26%. These things have a high level of importance and therefore probably something that we're going to want to be alerted to. So let's go ahead and put this to use. Let's scroll up here. Let's set my alerts somewhere here in the middle and back to our overall graph, and we can see our visits here and some medium alerts that have come across. One of the things that I think is critically important for this is when I talk about insights, we are really looking for that needle in a haystack, except I don't even necessarily know that the needle exists.
What I mean by that is if I were to see this large spike here, even without the Intelligence reports, I would probably figure out what that was. I would do some searching. I would go through my reports and I would see that something happened there of significance and I would go and figure that out myself. The real benefit of the Intelligence reports is for uncovering insights that I probably never would have found, because I had no reason to believe that anything out of the ordinary was happening. As I look across my visits graph, there are a couple of days that kind of jump out of me as large spikes. But as I look down here across my alerts I have a few days where there aren't necessarily anything in the visits graph that would make me go and look at that if I didn't otherwise have a reason to do so.
For example, let's take a look here at one of these. On November 15, up here in this graph the normal visits top line data over time graph doesn't give me any reason to believe this is a day out of the ordinary. In fact, it looks to be a little bit of a low-performing day. If I scroll down here what I see is something out of the ordinary I probably never would have found. The prediction algorithm expected between 0 and 140. In other words this is not a page that gets a lot of traffic. But in this particular day instead of 0 through 140 as it was expected, there are over 1100 visits to the Go Gopher Figurine page.
Similarly, if I come over here and take a look on November 28, I see that this was again not necessarily out of the ordinary day by any stretch, a little higher than the ones around it, but nothing on the course of the months that would be anything of interest. But as I scroll down here, I see that there were some high-importance events. In this case in terms of orders being completed, it was expected to be somewhere in the low $200 range, and was up over 950. Revenue was expected to be between $1000 and $2600; instead it was $22,000--over 500% revenue boost there.
I think this is really one of the major values of this. If I click on this particular report, I see that there is a significant event on that day, but I may have never known to find this needle in the haystack if I wasn't particularly looking for this, which frankly I wouldn't be. What we are doing here is utilizing the power of Google servers to search through and find these things that I may not otherwise find and to surface those up so that we as analysts can spend our time looking at things that are interesting and different rather than searching through numbers, which is what the computers can do. Intelligence can be used in many ways, such as to find these insights that were buried beyond our view or as we saw on the first example, a bit more forensic, to explain something that we saw but couldn't necessarily explain.
So the best way to find out what type of intelligence Google has found on your site is to simply open these reports and start digging.
Find answers to the most frequently asked questions about Google Analytics Essential Training.
Here are the FAQs that matched your search "":
Sorry, there are no matches for your search ""—to search again, type in another word or phrase and click search.
Access exercise files from a button right under the course name.
Search within course videos and transcripts, and jump right to the results.
Remove icons showing you already watched videos if you want to start over.
Make the video wide, narrow, full-screen, or pop the player out of the page into its own window.
Click on text in the transcript to jump to that spot in the video. As the video plays, the relevant spot in the transcript will be highlighted.