Ecommerce Analytics — Analyze and Improve the Impact of Your Digital Strategy — Ch2 (2023)

The Ecommerce Analytics Value Chain

Achieving successful analytical outcomes requires thinking of analytics as a value chain, which is called “the Analytics Value Chain.”

In a simple way, it is about how to think about the nature of analytical work from a managerial perspective. The value chain suggests a way to abstract an activity-based categorization of what is generally required, in most cases but not all, to do professional analytical work. The value chain answers the question, “What activities does my analytics team need to do to ensure quality analytical output and outcomes?”

The Analytics Value Chain is summarized into the following phases:

• Understanding what to analyze: The first phase requires gathering business requirements and goals, understanding the current data environment and the data within it, and developing a plan to execute work.
• Collecting and verifying data: The second phase involves determining if data is available and accurate and, if not, making it so. This work generally involves data engineering and implementing new data collection, data models, or databases. Data governance and master data management are implemented.
• Dashboarding, reporting, and verifying: The third phase is when dashboards, reports, and other artifacts that show data are created, verified, and made available.
• Analyzing, communicating, and socializing: During the fourth phase, the team analyzes data beyond the creation of reports and dashboards. Business questions are answered. Insights are generated. Analysts talk to stakeholders, meet with them, and ensure they understand the analysis. Narratives and stories are constructed and presented for discussion.
• Optimizing and predicting: The fifth phase is when the “data science” occurs and advanced analytical methods are applied and used to predict what could happen and recommend what should be done next.
• Demonstrating economic value: The sixth phase is when the analytical outcomes are gathered, and the business impact is quantified financially. The value of the analytics team and work is demonstrated by showing how it increased revenue or reduced cost.

The value chain represents the specific phases and work performed by an analytics team. It suggests the types of activities that managers want to align with when building organizations. It applies to ecommerce analysis because it is generalized in nature and at a high level. For ecommerce analytics, the value chain can be made even more specific because of the targeted nature of the ecommerce work.

• Identifying and Prioritizing Demand: Determine and frame the business requirement by asking a question, understanding the fundamental elements of domain knowledge and getting stakeholders to provide detailed, explicit, and relevant business questions.
• Developing an Analytical Plan: Demonstrates a rigorous commitment to a systematic, scientific approach by starting with a question, articulating a hypothesis, and explaining the logical approach and timeline for the analysis to be performed; Lists the business questions to be answered and the data set to be examined in a formal, structured way, such as a bulleted list or descriptive table. The goal of the list of questions is to ensure the right frame for the future data analysis; Provides a high-level, yet transparent road map, for the analytical work. The plan should suggest the objectives, data, resources needed, technology required, capabilities achieved, and the timelines for doing so; Defines the analytical objectives and the decisions made to deliver them. The objective cited in the analytics plan should tie back to financial impact — and should list the reasons why stakeholders chose these objectives as important. Aims to be objective by not introducing bias or opinion into the plan. The goal is to be consultative and helpful with analytical work. You might know the answer already or think the work request is not helpful. Although those conclusions and feelings are helpful and should be explored with fellow analysts, the analytical plan sticks to the facts and is objective. It shouldn’t be politicized; Provides a centralized artifact for focusing work as the work progresses. Some analytics teams will continue to update an analytical plan as work commences, as sprints ensue, and as decisions are made that change the scope or objectives of the work. Important decisions that impact analytics documented in a formal project management plan may be cross-referenced or cited in the analytics plan. **
• Activating the Analytics Environment and Collecting and Governing Data and Metadata: Deploying, using, managing, and maintaining infrastructure, technology, and tools (in terms of data collection, data sources, data processing, storage, and virtualization, data administration, data query and data preparation, data analysis, visualization, presentation) and the governance of data within those systems and architectures.
• Preparing and Wrangling Data: Data preparation like cleaning data via visually profiling, examining the data to show distributions, finding outliers, identifying erroneous or suspicious data, suggesting suitable join key
• Analyzing, Predicting, Optimizing, and Automating with Data: 1) The data is sufficient for answering the business question and doing the analysis. 2) If it is not, data needs to be collected within the analytics environment you activated. 3) The type of analytical work may be exploratory, explanatory, confirmatory, or predictive: i) Exploratory analysis is inductive - During exploratory analysis, the data is observed and explored to understand and explain it, perhaps using descriptive methods, against known premises. ii) Explanatory analysis can be both deductive and inductive in attempting to explain why something happened. iii) Confirmatory analysis seeks to accept or reject a conclusion or pre-conceived idea, hypothesis, or belief. iv) Predictive analysis attempts to identify what could happen given a set of data. It is deductive in nature but may include inductive thinking. Prediction may also lead to “prescriptive” analytics, in which recommendations or suggestions about the best course of action to take based on prediction are determined.
• Socializing Analytics: The positive perception of stakeholders about the helpfulness and support of analytics is enabled through socialization and talking to people, which enables the analyst to communicate answers to business questions, including conclusions, recommendations, insights, actions, and next steps.
• Communicating the Economic Impact of Analytics: Log the actions taken and decision made as a result of analysis, ask business stakeholders to estimate the financial impact, track the revenue or cost data and resulting derivatives, such as margins, use a financial model for capital budgeting, like Net Present Value (NPV) or Internal Rate of Return (IRR).

** Before creating an analytical plan, you need to consider the following:

• What is the business question?
• What is the business impact and how will it be financially qualified?
• What is the purpose of the analysis for the business question?
• What is the timeline?
• What analyses will be provided?
• What resources are available in terms of people and technology?
• How large is the total data set or sets?
• What types of data are available?
• Who is the audience and who are the stakeholders?
• What is the best way to deliver analysis to the audience?
• How will the audience judge the effectiveness of the analysis?

An analytical plan may contain the following:

• Business question(s) and hypotheses about potential answer(s)
• The purpose of the analysis
• The sources of data and the specific data sets you want to use
• The way you will integrate data
• Business rules used to include and/or exclude data about the analysis
• Specific key data elements to be used as variables in the analysis
• An indication of the type of analytical method that may be used

The value in analytics is created by analyzing data and helping people make decisions and create digital experiences based on data. There isn’t much value in analytics technology without business-focused analysis, so concentrate on creating analysis, not only on technology.