I'm going to be talking about metrics. This lecture is somewhat misnamed because I'm not going to tell you absolutely everything about every metric that you can potentially collect about your site. But really focus on the metrics that are going to matter the most to you as a user experience, researcher, or designer. So, the first is, there are metrics that you should be keeping track of that have nothing to do with what people are doing on your site. If really is to get a better sense for what's happening in the industry in general. So, at the highest level, you're going to want to keep an eye on what's devices, traffic and browsers people are using, because that will inform some of your thought processes around where it is that you need to be thinking about focusing your design and development efforts given the breakpoints and platforms that people are using who tend to come to your experience or your site. At the highest level you're really interested in tracking three different types of metrics and those are really around clicks, where people are coming from so geolocation and then over the life cycle of a customer. So, for a particular individual, whether or not people are coming back and having different types of behavior over time. You may or may not be able to get a sense for what that behavior is like for a particular individual depending on whether or not you have your site instrumented to capture that information. When I say that, that essentially means whether you have cookies set to capture that information about individuals. When thinking about events you really want to get a sense for what are people doing on a particular page? So events are actions that happen within the context of people interacting with a page. So, whether that is loading the page, whether that is clicking a control on the page or whether that is capturing the keystrokes as they are interacting with a particular page. But all of those things can help inform whether or not your design works, whether or not people get it and how compelling the content is on that particular site with regard to driving people to behave in the way that either they want to or you would like for them to. Essentially, you collect metrics using cookies. Which you do need to be careful because a minority of users have either blocked or deleted them, but which do give you fairly robust data so that you get an understanding of where it is that people are going and what it is that people are doing on your site. I am going to talk a little bit about some of the key things that you really should be thinking about from a user experience perspective on a more tactical basis. So, hits. That's any interaction that generates data. So, we talked about events but you can also have page interactions, social interactions, or e-commerce. You can think about time on site or session duration. So the calculation is here but essentially it is taking the time of the first hit on the last page minus the time of the first hit on the first page. Again this is sometimes very approximate depending on where it is that people go when they are leaving a particular page. So, when they leave your site, for example, this might actually negatively impact the quality of this data for you. Engagement is something that is very unique to each site, which doesn't necessarily equate with being good or bad depending on how long someone is actually on the site. So, fundamentally you're looking at the relationship between the amount of time that people spend on site and the number of pages viewed. I've seen everything from a hospital website where the number of pages that people viewed was one and people spent a relatively short period of time on the sites are typically between 1.5 and two minutes and that was actually a great interaction for their users. I've also seen the opposite where for an e-commerce site typically people were spending 90 minutes on a site and looking at over 100 pages. You will need to spend some time actually getting to the best way to understand what good engagement means for your site because that's not something that is easily generated overnight. So, it will probably take you at least several months or even a year to get a sense for what engagement really is good for you. Obviously, you want to understand what people are doing during visits or sessions. So, essentially bottom line you're looking at a series of page requests, you're looking at an IP address level. One thing to think about is that often people are looking at multiple sites at once because they're using multiple tabs. So, whereas you might be looking at a particular session people might have multiple sessions running at the same time across multiple tabs while they're looking at your site. So, that's one thing to be looking at. So, is it very common for people to be using your site across multiple tabs at the same time? Then also looking at how many pages did they visit during a particular time frame? Again those averages are very much unique to particular sites and you will need to figure out what makes sense for you. Visitors is actually a somewhat slippery concept too because a lot of the time people get interrupted when they're viewing your site. So, think about this. You've got a parent who arrives at home and they are looking at your website at the same time they are trying to cook dinner, feed the dog, clean the house, take out the trash. So, while they're having this visit they actually may be going away from their computer and you will only see this as potentially a break in activity on the site. But you can't be sure what it is that people are doing during that time. So, you just have to think about how do you want to understand how it is that people are using their site. This is one of those situations where a combination of analytics and as well some ethnographic research will help to give you a better understanding of how it is that people are using your site. So, that mixed method approach. Other thing that you need to be thinking about is how it is that you are looking at the distribution of the types of visitors to your site? So, your analytics using those cookies are going to tell you whether or not you've got first-time visitors and then return-repeat visitors as well. You do want to track them differently because the behavior of those visitors is actually quite different. Much of the time for certain types of sites that are informational most of the people who are going to them are going to be first-time users. Whereas, for say a financial institution like a bank, the vast majority of users typically over 80 percent tend to be return visitors because they are customers and they're coming to look at their bank account balance for example. So, the next type of information that you will probably be interested in is something around impressions. So, that is typically the number of times an advertisement has appeared on your site. Now just because an advertisement appeared on your site does not actually mean that somebody saw it. Because it could be that the advertisement loaded below the page fault for example or that people had banner blindness. So, again this may not be one of those key numbers for you but you do need to be thinking about that combination of both analytics and then some usability testing to get a sense for did people actually see it so is that impression data really reflecting the reality of people's exposure. Bounce rate is an important metric to keep track of because it can give you some sense of how successful people were based on the number of pages which in this particular case is only one that they've viewed on your site. It could be a good or a bad thing. So, if people came to your site got their answer in one page, find that's great but most often you want people to be engaged. So, having a very high bounce rate on a particular page is typically not such a great thing. Page depth can really be measured in two ways, both at the individual level as well as as an average across people who are visiting your site. So, for specific individual, it's the size of a visit. So, how many pages did they actually go to? If you're looking at the average across an experience you're looking at the total number of page views divided by the total number of visits. So, the click path is information about the chronological sequence that people follow as they are moving through a site. This gives you a better sense for where it is that people are having happy paths when you use this information in concert with say a survey. So, attitudinal analytics or whether or not you're seeing conversion information that's associated with these click paths. So, might be that a particular pathway you have a conversion rate that is higher than another one and you want to see if you can push people to that path. The bottom line is it's going to take you some time to establish what's normal for your site. So, the thresholds for what's high, what's low, what's good or bad and on a seasonal level what those levels of quality are, are actually going to take you a year or maybe, even more, to understand better. You also want to be looking for trends, both whether or not things over the long-term or of the short-term are changing. So, again establishing those thresholds and really understanding what is normal for your site will help you to understand whether or not those trends are significant over time.
