Web Analytics (Chapter 9)

**[Chapter 9: Basic Web Analytics and Web Intelligence]** **Web Analytics:** Web analytics, the practice of collecting and analyzing data from websites, continues to evolve with new technologies and techniques. Web analytics can be used as a tool for business and market research and assess and imp...

**[Chapter 9: Basic Web Analytics and Web Intelligence]** **Web Analytics:** Web analytics, the practice of collecting and analyzing data from websites, continues to evolve with new technologies and techniques. Web analytics can be used as a tool for business and market research and assess and improve website effectiveness. **Features of Web Analytics:** - **User Journey Analysis:** Web analytics now focuses on understanding the complete user journey across multiple touchpoints. It involves tracking and analyzing user interactions, behavior, and conversion paths across different devices and channels, such as websites, mobile apps, social media platforms, and offline interactions. - **Advanced Analytics Techniques:** Web analytics incorporates advanced machine learning and predictive modeling techniques. These techniques enable organizations to gain deeper insights, predict user behavior, personalize experiences, and optimize conversion rates. - **Customer Segmentation and Personalization:** Web analytics can segment website visitors into distinct groups based on their behavior, preferences, and demographics. This segmentation allows organizations to deliver personalized experiences, targeted marketing campaigns, and customized product recommendations. - **Real-time Analytics:** Real-time analytics in web analytics is gaining importance. Organizations now have access to tools that can provide instant insights into website traffic, user behavior, and conversion rates. - **Privacy and Consent Compliance:** With increasing concerns over data privacy, web analytics is adapting to new regulations and user expectations. Organizations are implementing privacy-friendly practices, obtaining user consent, and ensuring compliance with data protection laws, including anonymization of data, transparent data usage policies, and user-friendly consent management mechanisms. - **Cross-Channel Attribution:** Attribution modeling has become more sophisticated, considering the impact of multiple marketing channels and touchpoints on conversions. - **Voice and Mobile Analytics:** As voice assistants and mobile usage continue to rise, web analytics is expanding to include voice and mobile analytics. It involves tracking and analyzing user interactions, searches, and conversions on mobile devices and voice-enabled platforms, providing insights into user behavior in these emerging channels. - **The Most Common Web Analytics Use:** - Optimize websites website performance against specific marketing goals and initiatives - Maximize the marketing placed on websites. - Learn how site navigation, content, and aesthetics affect the bottom line, which should align with business goals. - Learn from past marketing efforts on a website. - Optimize future campaigns to increase conversion on a website. - Recommend website or marketing changes based on an analysis of website behavior. - Implement site changes or recommend changes to those in authority to do so. **Why Study Web Analytics?** - Google Analytics becoming the dominant web analytics platform, and it is accessible to most users in a free version for use and for learning purposes. As the current version of Google Analytics (Google Analytics 4.0) has a functioning e-commerce sandbox that connects to an actual website (The Google Merchandise Store), we will focus more heavily on Google Analytics when we discuss web analytics. - Marketers use online data to address fundamental business problems by leveraging data relevant to their business. Below is information about web analytics that marketers use to help them understand its foundation and capabilities: - Using web analytics, business users can use data to measure success, instead of blind faith because the amount of capturable data has grown significantly. - The web provides access to an infinite amount of data at a low cost. - Web and data analytics provide qualitative and quantitative data about monitored websites and customers, including customer intent for the organization and, at times, competition and desired outcomes for online and offline business goals. - Web and Digital Analytics is a broad field that covers more than websites. Analysts are expanding their skill set to include R, Python, statistical programs such as IBM SPSS, and visualization platforms such as Tableau. - Web and digital analytics are not part of big data. Web analytics databases are structured by the platform vendors. Web analysts operate on structured data in reporting suites and produce detailed reports of website and marketing performance using segmentation against pre-set business and website goals. Web analysts do not need to be programmers whereas, with Big Data, it is a requirement. - In comparison, data scientists work with large sets of semi-structured and unstructured data (big data) to create interactive and predictive analytics intelligence against marketing goals, Data scientists must program in Python, R, or use IBM SPSS, SAS or another high-end statistical platform. - In addition to the programming skills, data scientists need massive amounts of data to work with and run algorithms on. - Web analysts work with the first-party data that is collected and structured for them and requires a different skill set, including business knowledge and business communications skills. - However, web analysts and data scientists share a passion for data; it's just a different set of data. - Web Analytics and Big Data are closely related. Both are solutions that deal with the complexity of data that is being generated and collected and provide stakeholders with information that leads to actionable results. - Ideally, an analyst should be able to operate across the spectrum of analytics. Conversely, big data must be cleaned, organized, and structured before it becomes useful for most applications. **Commonly Used Software Platforms Supporting Web Analytics:** - **R:** Statisticians and data miners use the R programming language for developing statistical software and data analysis. R's popularity has increased substantially in recent years. It has become one of the most commonly used languages for data analytics and Big Data projects. - **Python:** Python is a high-level dynamic programming language that allows programmers to create software with a minimum of programming code. - **IBM SPSS:** SPSS (Statistical Package for Social Science) is a statistical package that can perform highly complex data manipulation and analysis with simple instructions. - **Tableau:** Tableau is a standard data visualization platform for enterprise data. - **Perl:** Perl is a high-level, general-purpose programming language that is used especially for developing Web applications. **Web Analytics: Path of Value** - While Web Analytics provides great business value once it is set up well, there some things it is above to do very well and other things it cannot do, at all. - It is important for that stakeholders' expectations are managed realistically, when a web analytics platform is first implemented there is a list of functions it can do, and others it cannot. **What Digital Web Analytics is and what it can do?** - Analyze data from a website or mobile app or something else in order to improve business outcomes. - Analytics can tell us what, when, where something interesting happened on a website or application, but not why it happened. For example: Analysis of some, but not all questions a business has, such as what happened that resulted in a change of amplitude or frequency of data that is captured from a website or mobile application and where the source of the change came from. - Some third-party tools such as Microsoft Clarity can be combined with Web Analytics to better understand the "why's" and "where" of the root cause of changes happening on the website or application. **What Digital Web Analytics is not, and what it cannot [yet do:]** Digital Web Analytics usually cannot tell us WHY an event or interesting thing that we are monitoring in the platform is happening (it cannot know the ultimate source and cause of the things being analyzed, because it is beyond the scope of what analytics was designed to collect -- this usually needs a human to connect the dots, when possible, back to the ultimate cause of an event or interesting change on the website or app). **Determining goals for web analytics:** - Find out what the organizations business goals and look for those goals where web analytics can assist. - Once the business goals are defined and determined to be informated by the web analytics, its then time to ask the stakeholders how those goals will be met, going from the general to the more specific examples. - Once the business goals are defined and gathered by the analyst or analytics team they should be written down in a document that is accessible to all stakeholders and be used to guide the initial implementation. - As shown in Figure 9.2, a successful Web Analytics implementation moves down a path beginning with an overall strategy to lower level tactical business targets that can be measured. - Later, the results are contemplated and communicated with midlevel managers, generating digital strategies and key performance indicators that impact stated business objectives. - Finally, once the measurements are defined and campaigns implemented, analysts define behavior segments and business targets, as part of their day-to-day operations, to achieve the larger, stated goals. **The Cycle of Improvement:** - The "Cycle of Improvement" (refer to Figure 9.3) is the continuous improvement cycle that businesses operate in, as market conditions, competition, and technology change. - There are cycle charts for Search Engine Optimization and Search Engine Marketing that work in a similar way. - The digital measurement process involved in the Cycle of Improvement includes mobile, social, campaign, survey, competitive, and offline data, such as closed sales from online leads (see Figure 9.4). **Key Business Requirements (KBR) vs. Key Performance Indicators (KPIs):** - KPIs are a way to determine if an organizations business goals are being achieved or not (during a defined period of time), that's all they are, no need to mystify it. - Once a organization defines their business goal(s), they need a method to determine if the their business(es) are meeting their business goals (performance based). - Each KPI chosen must be measurable (something that can be captured in a web analytics report, ideally) and directly tied to one or more specific business goals (for example as shown in Table 9.1). Metrics in and of themselves can be or could not be useful, it all depends if the metric can meaningfully be part of KPI that is tied to a business goal or not. Web analytics metrics, for example, would not be of much use for digital analytics if they are not tied to a specific business goal. - Key performance indicators (KPIs) are the digital measurements marketers use to track progress, whereas key business requirements (KBRs) reflect the business requirement or goal that the processes are meant to realize. - For instance, referring to the schema shown in Figure 9.5, if KBR 1 is a business requirement to encourage types of member content to be shared more often, then the implementation of the share button and the number of times it is clicked on becomes one of the KPIs that informs KBR 1. **KPI EXAMPLE:** - Are page views per session that can be supplied by any web analytics platform a good KPI or not? (It depends.) - If the business goal is to acquire more business leads that lead to more product sales, then page views per session would not be a particularly useful KPI. In this case, there is not a clear correlation between pages viewed per session of a business lead and sales that resulted from viewing those pages. - But if the business is a type like a news website, where a major source of revenue is page views where there are ads displayed on those pages, then page views per session might be a great KPI (metric) in this case. - Different businesses tend to have KPIs that are more typically use for that kind of business and vice versa. **Types of Web Metrics:** - There are two kinds of web metrics, activity-based traffic versus conversion event-based metrics. Depending on which Web Analytics platform used, the reports may have different names, but the same functionality exists on almost every Web Analytics platform **Activity (out of the box) metrics vs. milestone metrics (configured):** **Correlation Filters:** - Adobe Analytics and Google Analytics offer report filtering, but they are evoked in various ways. Adobe Analytics has secondary dimensions but calls them "correlations." Analysts select several correlations at one time if desired. - For example, a correlation filter can be configured that shows how many confirmed registrations took place from 25--29-year old visitors to a fashion retail website during a given period. - If the data is collected in the Web Analytics platform, there is probably a breakdown or correlation filter report that will be able to inform a specific business question. **Choosing Relevant Web Metrics:** - Metrics are the currency of Web Analytics, which is an excellent "bean counter" containing much built-in intelligence about Web traffic and conversions taking place on a monitored website. Some of the more common metrics are visits, unique visitors, pageviews, revenue, and time spent. - Each business is unique, and KPIs should be tailored to the company's needs. - Knowing how an industry or organization operates makes it easier to determine what to look for and benchmark. **Intermediate Web and Social Media Metrics:** - Intermediate Metrics are marketing data that falls within the range from impression to purchase, such as bookmarks, views, the number of followers, shares, clicks, retweets, likes, pins or downloads. - Web Analytics, and social media analytics have created their own currencies but they are not interchangeable, and they should not be combined in a single metric with integrity (because they measure different behaviors from different audiences). - Consequently, while Intermediate Metrics are useful as a measure of audience engagement, they should not be considered the endpoint of a "Return on Investment" calculation. **Designing KPIs That Work:** There is an art to developing a useful KPI, and here is one approach: - Define the business problem (what are the main issues the organization faces?). - Chose a single issue (per KPI), and at least one designed KPI indicator for each issue. - Define three or four processes associated with the KPI (but don't get too specific, yet). - Define a statistical measure that will be linked to each of the: processes related to tine KPl. By \"instrumenting\" an Indicator and setting thresholds with a high/low action for each process, it becomes possible to place the KPl in a dashboard and act on It. In a nutshell, connect each business process with a data source that will inform It. - Define the measurement algorithm for the KPI as a formula when possible. - Define the procedure to conduct/measure the KPI.