There are several approaches when it comes to leveraging data. You have those who swim through an ocean of data in the hopes of finding treasure. Stumbling on something sounds like pure luck, but according to Jim Sterne, Chairman of the Digital Analytics Association whose considered the “godfather of web analytics,” today’s machine learning and artificial intelligence systems can spare you the trouble by surfacing what is most relevant in your data. One such solution might be Outlier.ai, and other startups as well as established big players actively working to address this challenge.
The idea that machine learning can minimize that workload by bringing relevant data to the surface is tempting, however this is typically only possible at higher stages of analytics maturity, once you have mastered the fundamentals and have a certain level of comfort that your data is “good enough” (re small margin for error) in the first place.
Machine Learning also tends to be more efficient when there are vasts amounts of data, so small anomalies and noise are ignored, and only the data of real interest is surfaced. Even if a tool surfaces exciting tidbits of data, you, as the analyst, must be able to turn this data into information and insight.
|Get broad exposure to key technologies and skills used in data analytics and data science, including statistics with the Post Graduate Program in Data Analytics.|
A second, more structured approach to leveraging data is to start with a problem to solve or a goal to achieve. The shortest path to data-informed decisions is to apply a methodology that has been proven over time, such as Lean Six Sigma. We’ll dive further into this concept momentarily.
Conceptually, businesses are a complex combination of processes with specific inputs, producing specific outputs at every stage of transformation, be it physical products, services, government, academia, or any organization. If you break those processes into manageable components, you can understand those processes, measure, analyze, improve, and control them.
In math terms, we would say y = f(x); the output is a function of the input. Business analysts are trained explicitly for “enabling change in an organizational context by defining needs and recommending solutions that deliver value to stakeholders” (IIBA). As you can imagine, digital analysts and business analysts have many things in common, particularly the ability to understand requirements and break down the various steps required to achieve a result.
One way to overcome complexity is the Lean Six Sigma methodology. It prescribes five simple steps: Define, Measure, Analyze, Improve, and Control. It involves defining the problem (or opportunity), measuring and getting relevant data, analyzing correlations and patterns, and being creative about possible improvements that can be measured and controlled over time.
Lean Six Sigma uses qualitative and quantitative techniques to drive process improvements. Practitioners are encouraged to learn about Lean Six Sigma tools, many of which are derived from statistics, business analysis, and quality management.
Growing Beyond Digital
Your web analytics platform, on its own, is a goldmine of data. Given the instrumentation is appropriately done, you can track an unprecedented amount of information detailing how your visitors interact with your websites or mobile applications. Those digital properties offer a unique conduit to observe who your visitors are (Audience), where they are coming from (Acquisition), what they are doing (Behavior), and whether they are successful (Conversions). This is conveniently reflected in the reporting structure of Google Analytics.
There is an infinite number of business questions you can answer by properly leveraging data. For example, by applying the Lean Six Sigma methodology, an analyst can quickly asses that the “single source of truth” for sales might not be Google Analytics or another web analytics platform, but the back-office system is storing transactional details. The question is then to find the proper balance between “good enough data to inform a decision” and the cost and time to get to a hypothetically perfect answer. Directionally, and despite issues affecting data quality such as ad blockers, cookie deletion, data spamming, or mere tracking errors, web analytics data is often enough to inform decisions. As you grow in digital analytics maturity, you will uncover more opportunities and justifications to merge online behavioral data with your back-office data and other relevant sources.
The challenge doesn’t stop at data integration needs; visualization complexity and analysis requirements tend to grow exponentially. But maybe even more significant, the organization needs time to absorb the cultural impact involved with becoming more data-savvy.
Becoming "data-informed" is a corporate culture shift. Businesses must realize that data alone won't help; it can create more confusion—which is why we need people who can turn this data into useful information and knowledge that will drive change.
The Path to Wisdom
In the Data-Information-Knowledge-Wisdom pyramid, we can identify “data” as being the raw observations gathered from websites and mobile applications. Information is provided through the numerous reports available in Google Analytics. That is the extent to which a tool—any tool—will help the analyst.
Even if and when there is a level of machine learning and artificial intelligence at work, there is always the need to have a human being involved in making sense of the data.
Digital analysts are faithful “knowledge workers.” Knowledge is the unique value digital analysts can offer by wrapping business context around the information. Finally, wisdom is the active managers would take given the newly gained knowledge but counterbalanced with all the other constraints of a business.
Digital Analytics Maturity
The Digital Analytics Maturity Model is a framework that helps organizations assess their current situation and provide a structured, actionable path towards improving competence at leveraging data and analytics for enterprise-wide business decision-making.
The Digital Analytics Maturity Model provides you with an unbiased and easy-to-understand representation of your organization’s expectations of, and commitment to, your analytics infrastructure and initiatives. From your objectives, scope, and resources, to your methodology, tools, and management, the model will illustrate whether your organization is engaged in the critical data-gathering and customer insight-sharing activities required for your business’ success.
Taking a step back to think about your business’s strengths and weaknesses, whether you use this model or another one, will allow you to:
- Expose areas for improvement in the alignment of people, processes, and technologies with the declared objectives and scope of the analytics initiatives.
- Create a roadmap of specific and achievable improvements that will move your organization further up the path toward digital analytics maturity.
- Provide upper and middle management with a persuasive communication and persuasion tool that can be used to drive a message of organizational change.
As an analyst, your first step in your analytics journey is to conduct a digital analytics maturity assessment to identify strengths and weaknesses and develop a comprehensive roadmap. Starting with the business in mind, you first identify audiences, lifecycle, and desired outcomes.
Next, the onus is on you to define objectives using the new Product Thinking concept (instead of the classic SMART approach), embrace Agile ideas of DMAIC to overcome complexity and stay focused on quick, iterative value delivery, and introduce your business to your web analytics platform (like Google Analytics).
Gradually expand to include other relevant data sources, primarily from third-party ad providers, and your sales and CRM systems.