Analytics is a major part of the OMCA exam as it covers how to measure the result and the return of the other disciplines. Of course, with this section comes a healthy dose of terms, but you’ll have heard many of these already. In this video, I’ll give you the formal definitions of activities you do daily when working with a website. You may still hear people use the term hits to describe how popular their website is. Unfortunately, it only refers to how many files have been downloaded from your web server. For example, if someone downloads a homepage it is comprised of dozens of files, so it’s a meaningless term. Let’s start with the people coming to your website. Obviously, a visitor or a user, these terms are used interchangeably. This is someone who comes to your website. If it’s the first time they’re visiting your site, it makes them a unique user. If they return within a measurable timeframe, then they are a returning visitor. The ways visitors find a website is important in analytics. If they come from another website, that is called a referral meaning they were referred to your website by clicking a link on Pinterest, someone’s blog, or the New York Times. If they typed in the URL of the website in their browser then they are a direct visitor. The page they enter on is called the landing page. If they only viewed the landing page and leave the site, then they’ve bounced. That means a visitor only viewed a single page of the website. If they view multiple pages and then leave, the last page they saw is the exit page. In marketing, what you want the visitor to accomplish on the website is a conversion or a goal. This can be a sale, subscription, registration, any action that benefits the business. Now, when analyzing this information, you’ll use an analytics program like Google Analytics. The first thing you see when opening your Analytics is the Dashboard. It’s the centralized report of activity on your website and insights of your goal performance. When analyzing visitor data, Analytics lets you break up the data for better analysis. This is called segmentation. Segmentation allows me to view a certain subset of visitors based on a particular characteristic. That specific characteristic is called a dimension. This could be the source of their visit, such as a particular website. It could also be what device they used, their regional location or city, or an action they performed. Basically, any method of separating visitors is called segmentation and the unique separating factor is the dimension. Then, when viewing the data from the segmentation, you’ll see the metrics. The metrics are the accompanying data for each dimension: users, sessions, bounce rate, pages per session and conversions. In this example, I am segmenting the primary dimension of visitor source. The channels are the methods that people found the website, then the accompanying metrics for each channel are shown in the subsequent columns. This is the Analytics view of website visits which provides comparison data of visitor acquisition, behaviors and resulting actions. Now you know the primary terms and concepts of analytics. Now that was a lot of information, fast. So don’t hesitate to watch this video again to make sure that you have everything clear.
So, where does all of this analytics data come from? How do we know that you visited a website, looked at four pages over the course of five minutes, and watched a video on one of those pages, and then subscribed to an email newsletter? Let’s take a look at how analytics data is collected. It all starts with tracking data. This is done with cookies. Cookies are a small text file that are delivered to your browser from the website you visit. It confirms you viewed a page, and communicates information about your computer or device. Analytics can track this cookie to know the last time you visited the site, and how you found it. Of course if you delete your cookies, then you’ll show up as a new visitor. Now tracking does have limitations. The analytics program can only track when the page is requested, not when or why anyone leaves. Also, cookies are limited to devices. So a visitor on a smartphone who then returns to the website on their desktop is recorded as two unique visitors. Of course, this creates an immense amount of data, and to report every single visit would require an enormous amount of computing time and resources. Because of this, many analytics programs use data sampling to present information. Especially for large websites with millions of users. With a sample of the overall information, analytics programs can present an accurate depiction of the trends without the resources required to represent all of the data. Now, to track specific campaigns, you can add additional code that separates and defines anyone who responded to that campaign. These are called tracking parameters. Here’s how this happens. Let’s say you are sending a promotional email with the a link to a product page. With the tracking code, you’ll know how many people clicked the link from the email, went to the website, and eventually purchased. The tracking code informs the analytics of the campaign source, which enables that activity to be segmented for reporting. You can also use this code in any link for a paid search campaign. Each of your ads can have a unique tracking code, distinguishing the campaign, ad group, ad, and keyword. This lets you see all of your campaigns in a single report, and compare the effectiveness of each. So getting analytics data doesn’t happen by itself. Code needs to be added, and specific campaigns require additional tracking. But the extra effort enables clearer comparisons and better analysis.
Segmentation, a major principle of analytics, is dividing your visitor data based on common characteristics. In this video I’ll show you how segmentation works. You see, without segmentation you are viewing the data of everyone who has visited your website in one big report. Of course, seeing everyone in one report is confusing regardless of how they found your site, what they wanted, what they saw, or what they did. No wonder people freeze when they look at analytics reports. When you separate your data, it creates context. For example, I can compare my visitors from different locations, countries, regions, states, or cities. I can also segment the source of my visitors. I can segment by source and look at all of my search traffic. I can also create a subsegment and view my search visitors by which search engine they used. I can also segment by device and see how visitors accessed my site, how many used their mobile compared to a desktop or a tablet. Then I can see how effective my site is for those mobile users. I can also segment by actions or campaigns so I can make a segment of those who clicked on a link from a recent paid search campaign. I can also segment anyone who purchased a product and compare it to those who didn’t. In this example I can see that Google search provides the largest number of visitors and subsequent revenue to my website. But I can compare that to visitors from other search engines or sources. You may find that some deliver fewer visitors, but they may have a higher conversion rate and they might spend more per visit. Only by separating these visitor sources am I able to see these behaviors and draw conclusions about my marketing approach. Segmenting provides context in viewing visitors rather than evaluating everyone altogether. By dividing common characteristics you can make better evaluations of the effectiveness of your marketing or of your website.
To many people, the Key Performance Indicator or KPI is simply a metric that has to be tracked. It’s merely additional data that clutters the report. I’ll show you how KPIs are used when analyzing campaign data. By definition, a KPI has to be useful and provide an indicator of something. Otherwise it’s ineffective. A KPI should align to business goals and provide context as to how well a campaign is meeting those goals. Let’s say my business goal is to sell a product. The KPIs should then measure the progress of visitors towards that goal of buying. Well a KPI is not a goal and it isn’t necessarily directly related. It provides insight for a particular step of the sales process. If it starts with a paid ad, then my first KPI is impressions. The number of people that see my ad. If my impressions are too low, then I’ve made my target too small. The number of impressions in this case isn’t a direct correlation to sales, but it’s a useful KPI. My second KPI would be the CTR or the Click-through rate to my website’s landing page. And if the Click-through rate is too low, then it indicates that my ad is not performing well. Page views of the product page from the campaign is the next KPI. Then, I measure the actual goal, transactions. KPIs let me see what’s called the funnel. And understand how to adjust campaign settings based on the indication measured. Typically in a funnel report it starts with the broadest KPIs to the subsequent KPIs down to the final action or measurement goal. The funnel report provides a visual approach to seeing the rates of visitor progress through the intended milestones. By understanding the relationship of KPIs to each subsequent milestone analysts can determine actions to take to improve the campaign and improve the return on investment.
How do we know what information is the most important in analytics? Unfortunately, analytics has suffered from information overload. Too many times have lengthy spreadsheets, charts, and graphs been used to communicate information resulting in drowsy eyes and confusion. I’ll show you how to avoid reports with meaningless data and how to get to the actionable issues. One of an analyst’s most important skills is finding relevant information. But even more important is communicating it. Typically, most people rely on dashboards for their initial information. With a few graphs and charts, people can get a visual view of the right now status of meeting campaign goals. The problem is that many analytics programs have default dashboards, which is a one size fits all approach. The dashboard may not contain the most important information for your business and your business goals. Most analytics products allow customers to develop dashboards. These are particularly helpful since they can be customized for the use of different people in the organization. IT can have their performance dashboard. Marketing, a detailed campaign dashboard. And the executive level, the high level summary dashboard. Customizing data presentation to particular groups is vital to reducing unnecessary information overload. When creating a dashboard, the idea is to present meaningful information, that is, information that can be acted upon, such as a trend that could impact a business or a campaign. For example, if I see that I have a lot of mobile visitors, but they are leaving the website quickly and without engagement, it is indicating a problem with my website on a smartphone. Action needs to be taken. Now, in order to get approval for that action, information must be presented quickly and clearly. One thing I’ve learned over the years is that decision makers don’t respond to spreadsheets. They respond to stories and reports that show how to improve sales. So if I take that same information about my mobile users and I repackage that into a report that shows how much revenue is being lost by day, by month, and by quarter, now I have their attention. Then I can propose a solution that shows the financial benefits of improving the mobile experience. One warning, however. To better understand reporting in your analytics program, you need to understand that attribution, that is, how the final conversion is attributed. Let’s say a visitor first finds my website from a Google search. They visit a few pages and then leave. But a few days later, they come back. But this time, they remembered the website, typed it in, and came directly. From that visit, they found the product they wanted and bought it. Now most analytics programs employ what is called last-click attribution, that is, in this scenario, the direct visit gets the credit as being the source of the sale. Of course, that’s not correct. They first found it through search. There are multiple attribution models that can be used for different types of businesses, first click, last click, linear, which gives credit to each source. The key is finding which model works best for your needs and provides an accurate representation of your visitors.