Transcript for:
Google Analytics Reporting TroubleShooting

  • Hi, welcome back to our video series all about reporting in Google Analytics. I'm Krista Seiden and I'm here today to help you troubleshoot some common issues and questions you might see when looking at your analytics report. Quick reminder that this video is actually part of a video series all about reporting in Google Analytics. Together we've been looking at how to find answers to your basic questions about how people are finding and interacting with your website or app. You can find all of the previous videos linked in the videos description on YouTube. Today is the final video of this week long series, but we'll be back with more reporting videos for you in the future. Comment, what questions came up for you this week, or which topics you'd most like us to cover going forward? - So today we're answering some reporting questions that you've already submitted and there are some clear common themes. We'll address your biggest question about values like not set or other and tables, and we'll also talk about how to approach trend oscillations and apparent data discrepancies. One caveat, the answers to these questions will vary and be unique to your situation. So while we ultimately can't guarantee a specific answer to your specific property, we'll still be able to cover some common causes and solutions to these issues. With that in mind, let's get started. First, let's end to the question, how do I validate my data? Many of you want to know how to make sure your data is flowing into Google Analytics properly and is working correctly. So there are a lot of ways to go about this, but let's start with implementation. If you are the one implementing and using Google Tag Manager, you should use the Google Tag Manager built-in preview mode to test your tags and ensure that data is firing correctly. If you're not the one implementing or if you just wanna see that data is actually following in on the Google Analytics side, a great resource directly in analytics is the real-time report. Let's briefly look over at the real-time report so I can show you what I mean. So the first thing to look for is that you actually have traffic coming into your site. So here we see we do have active traffic on our site, so that's a great sign, data is flowing. And specifically, if we wanna know if certain events are being collected, we can use the event count by event name card. Maybe we've just implemented a new promotion on our website and we wanna make sure that that's actually showing. We can come to this View promotion event, click in and then find the promotion name and look it up. So if we just launched the calling all YouTube fans promotion, we can see that yes, this is showing people are interacting with it in real time, which gives me a good sense that it's working as I intended. Another thing you can use the real time report for is to see that traffic is coming in from different sources. So in this case, we are looking at users by first user source. We can change that if we want, maybe source medium and see, okay, great, I've got some direct non-traffic, Google Organic, Google CPC, and if you've just launched a new marketing campaign with specific tags for source and medium, you would see those show up here if you're testing out those links. The next set of questions was all around undefined values and reports. So things like, not set and other. Let's address these one by one. So what does a not set event or an empty name actually mean? Not set often means that there's a setup issue preventing Google Analytics from receiving and reporting information for a dimension. If you see a not set value, this would be a great place to troubleshoot and dig in to ensure that all of your events are being properly tracked, all of your pages properly tracked, and that your implementation code is firing everywhere that you want it to be. The other common question we get here is what does the other row mean? The other row appears when the number of rows in a table exceeds the table's row limit. Analytics will show the most common dimension values in the table and condense less common values under the other row. The table's row limit will vary based on property type, individual report and complexity of your query and data. So there isn't a hard and fast rule for avoiding the other row. That said, there are a few best practices that might help you to reduce the likelihood of seeing the other row. A best practice here is to not use high cardinality data, meaning that the data has hundreds or even thousands of assigned values unless absolutely necessary. So things like a user ID or a client ID could have a lot of individual values. If you have a lot of users coming to your site, in which case these could cause high cardinality data. And along those same lines, don't create a custom dimension to distinguish individual users. Instead, use the built-in user ID feature. And finally, use standard reports when possible to help avoid the other row. We also get a lot of questions about data discrepancies and trend oscillations that you might see in your data. And while there isn't one hard and fast rule for what might be causing this, there are a few things that you can look at to help troubleshoot. So let's go ahead and look at that. If we move on over to the report snapshot, here we'll see a graph of our users over time. And specifically, we see this one spike in our data. Now if you see a spike in your data, this could be an anomaly. Google Analytics has built an intelligence to detect anomalies, and in this case here in the report snapshot, this tells us that while there was a spike in users on this particular day, it expected to see about 10,000 users, but actually it saw closer to 20,000 users, which was a 95% increase in the expected value. Digging in further, since we know this is a spike in users, if we go over to acquisition and user acquisition, we can try to dig into where the spike of users actually came from. So here we see that same spike and it looks like direct traffic was a big part of it, the biggest part of that spike, and also some from organic search. One cool thing that you can do for most reports in Google Analytics is actually open up the same report in explore. If we come over to this little button here that says edit comparisons down at the bottom, you can click explore. This will open the same report in explorer. Here we have our table report and we also have a line chart. And on this line chart we do see those same anomalies and this time broken out by new users per channel. So direct we see there is 7,800 visits from direct and expected was 2100, so a 267% increase. This looks like the majority of our traffic came from direct on this day. We can dig in farther by trying to understand if this is legitimate traffic or potentially spam traffic. If I go back over to the table report, what I wanna do is add a few dimensions here that can help me understand the legitimacy of this traffic. Now there are a lot of different dimensions that you might use to help understand traffic such as country, browser type, browser version, device type, operating system and more. And we'll just add a couple of those here to see if we can narrow it down. So I'll go ahead and add some of those dimensions such as country device category, browser version, and operating system with version. And since I know a lot of this traffic was from direct, I'm gonna go ahead and include only this selection by right clicking this row. And now I'm gonna add a couple of dimensions. So first I'm gonna start with operating system with version. I can double click that to add it to this table. And when I add this operating system with version, it does look like I'm getting a lot of traffic from Macintosh Intel 13.5 specifically. So perhaps some of this traffic might not be legitimate. You can continue to dig in using this method to try to narrow down the traffic that you would expect and the traffic that you wouldn't expect. And then you can create a segment of that traffic so that when you analyze the rest of your data, you can make sure to exclude that segment of traffic. Going back over to reporting, we do often see lots of trend oscillations in data and it's also important to know if those are expected or unexpected. So for example, maybe your business does most of its sales during the week. You would probably expect to see dips in your data over the weekend and that were to reflect on the graph. So if you see a sudden dip over the weekend, but it goes back up Monday through Friday, this might be expected behavior and not something to worry about. However, if you do see these little bubbles on these graphs that call out an anomaly, that's an area where you might wanna dig in deeper. Okay, so today we covered validating your data, interpreting report values, addressing unexpected data spikes or dips, and comparing data. To go deeper on any of these topics, check out the resources linked below this video on YouTube, including our help center and our analytics academy courses on Skill shop. If you have other questions about understanding analytics reports or where to find certain answers, comment below. We may not get to answer every question, but your comments can help inform future videos. In the meantime, bring your troubleshooting questions to our Discord server, join the community and the conversation happening on Discord, and stay tuned to our YouTube channel for more reporting videos. Habby measuring.