We’ve all experienced or observed it — that day when we were so overwhelmed that making a decision based on data seemed next to impossible. Analytics platforms like Google Analytics have been evolving exponentially; in respect to features, the interface and how data is collected. This creates an ever increasing challenge for marketers and analysts to take action.
Instead of focusing on business questions, people tend to enter a state of over-thinking and over-analyzing a situation so that a decision or action is never taken. This is known as analysis paralysis.
In this article we present an easy to follow 6 step framework to overcome this increasingly common problem with a structured analytical approach and the incredible data visualization tool – Tableau. This will empower you to generate a solid ROAI (Return On Analytics Investment) by improving your ability to take action on Google Analytics data. Gold star and a raise for you!
Framework to Drive Action from Data
The need to isolate “noise” information is vital to turn data into action. A newspaper organization reached out to us at Blast with an interesting challenge. How could they leverage Google Analytics data to understand if the articles they write engage their audience?
Step 1. Identify Primary Goal
It’s fundamental to identify the primary goal. Usually you can find it by asking, “why do you exist?” In this case, this newspaper organization exists to inform their audience with relevant, timely news and other valuable resources. So their goal is to continuously increase and retain readership by ensuring that their articles are being read. Not an easy task…
Step 2. Ask Questions
Don’t think about the data yet! Don’t look at the Google Analytics interface! Forget the vanity metrics! Instead, focus on the goal and ask questions that really matter. So let’s start with the goal in order to develop the primary question to ask.
Defined Goal: Increase article readership/engagement
Primary Question:
- Do visitors read our articles?
When you ask yourself how to increase reader engagement you need to embark on a journey of exploration and critical thinking that is focused on the primary question. This will help you to drill-in and develop secondary, follow up questions.
Secondary Questions:
- Does the percentage of an article read vary by category or journalist?
- Can we improve visitor engagement by understanding which journalists write more engaging articles for each category?
Step 3. Ask Questions
These questions then lead us on a quest for data. By default no analytics package measures how much of a page was read. Thanks to a custom Google Analytics implementation solution from Justin Cutroni; we can understand how far a visitor scrolls down the page and how long it takes to do so. Each article pageview is then classified as either a read or a scan.
Read Ratio = Total Read Pageviews / Total Article Pageviews
To rank each page we calculate a read ratio, which is defined by the total number of read pageviews divided by the total number of pageviews for that article. Along with this information, we send the article author and category to Google Analytics through the use of two custom dimensions.
Step 4. Explore Google Analytics Data with Tableau
Having identified the data needed to answer our questions and loaded it into Tableau; it’s time to use our creativity and Edward Tufte’s analytical design principles to create data visualizations. There is no magic recipe, Tableau is a forgiving tool that allows you to create multiple views of your data with very few clicks. Let your inner adventurer, explore!
The use of Tableau visualizations and dashboards can further enable a marketer to explore the primary question and drill-down to uncover answers to the secondary questions.
Primary Question:
-
Are our articles read? Answer: Around 58% of the article pageviews are considered reads. Secondary Questions:
-
Does article reading vary by category and journalist? Answer: Yes, journalists and categories have different read ratios. We can even understand which journalists write content with a higher read ratio for each category and vice-versa.
Step 5. Drive Action
This secondary question is actually a hypothesis in form of a question;
“Can we improve visitor engagement by understanding which journalists write more engaging articles for each category?”
This hypothesis can and should be tested, as this is what drives action and ultimately affects the primary goal (i.e. reads). What would happen if we promoted the most read authors in each category to write more content in the categories where they perform best?
Start small, benchmark your read ratio for a given category, promote the best performing authors to write more than the others. Identify each journalist’s favorite topics and how this relates to their article read ratio. Analyze read ratio differences between authors to identify what tips could be shared with lower performing authors to improve the performance of their articles.
Observe read ratio over time for a specific test category. If the test is successful share the results with the team and implement across all categories.
Going back to our previous Tableau “Authors & Category Read Ratio Performance” Visualization/Dashboard, we were able to come to the following data-driven insights that we can take action on:
-
Cameron and Myrtie are the best writers for the Sports category (as their read ratio is about 33% higher than other authors). Action: Cameron and Myrtie should write more Sports articles and share tips with other Sports category authors to improve their read ratio.
-
Haley’s read ratio is sub-par in the Entertainment category (as her read ratio is only 9% compared to an average of 33% for the other authors). Action: Haley should try writing for other categories or use tips from top performing authors to improve her read ratio.
-
Silas’ articles generate a higher read ratio for the Jobs category (as his read ratio of 28% is double that of Amina). Action: Silas should continue writing Jobs category articles and identify what he does different than Amina to help her improve her Jobs category read ratio.
Step 6. Iterate
Repeat the process by iterating through the steps.
- Identify other goals
- Ask questions
- Find data to answer the questions
- Explore data with Tableau
- Drive action
Why the Framework? Because Measuring Only Takes you Half-Way
Using any analytics package to measure traffic only takes you half-way. The ability to measure everything can easily turn you and your analyst into overwhelmed “reporting monkeys.”
Let’s step outside the digital world for a moment.
Think of a medical monitor. It continuously measures vital signs like heart rate, blood pressure and it may even repeatedly perform medical tests, such as blood glucose monitoring. Like other measuring devices, it requires a qualified and experienced individual to be able to interpret these signals into what could be a life-saving action.
Hospitals are increasingly data-rich and data-driven, but notice how sparse the medical monitor interface is. It contains only vital information for the technician, mainly in two different forms: a trend over time and summarized data. This simple interface allows the technician to immediately answer questions about his patient:
Primary Question:
- How is my patient’s health? Secondary Questions:
- Is the heart rate stable?
- Is the respiratory function working properly?
In the right hands, the data from this medical monitor may ultimately end up saving the patient’s life.
How Can We Emulate this Medical Analogy?
One of my favorite methods is to embody Edward Tufte’s 6 principles of analytical design.
Principle 1: Show Comparison
Show comparisons…always ask, “Compared to what?”
Example: The technician asking the questions has the know-how and experience to be able to mentally compare these values with previous patients and benchmark the current patient relative to others.
Principle 2: Show Causality
Use causal logic to reason a probable cause for what you are observing.
Example: The patient has a lack of oxygen in their blood, which could be due to CO2 poisoning. Looks like we better provide O2.
Principle 3: Show Multivariate Data
Explore data across the multiple variables and dimensions. This helps you get a better picture and uncover relationships between variables.
Example: Measure heart rate, CO2 levels and blood pressure over time.
Principle 4: Integrate the Evidence
Integrate words, numbers, and images. Data graphics should make use of many modes of data presentation.
Example: The patient monitor provides not only a snapshot, in the form of summary data (right part of the screen), but also a trend over time. Additional medical exams may help provide a more thorough diagnostic.
Principle 5: Documentation
It is essential to document the process, observations, relevant facts and data.
Example: Doctors keep a historical record of their patients exams results and other meaningful information. This provides context needed to understand the patient.
Principle 6: Content Counts Most of All
“Analytical presentations ultimately stand or fall depending on the quality, relevance, and integrity of their content.” (1)
The most effective way to create a compelling presentation is through clearly presented content. Pretty graphs can only take you so far.
“The purpose of an evidence presentation is to assist thinking, thus presentations should be constructed so as to assist with the fundamental intellectual tasks in reasoning about evidence: describe the data, make multivariate comparisons, understanding causality, integrating a diversity of evidence, and documenting the analysis. Thus the Grand Principal of analytical design: The principles of analytical design are derived from the principles of analytical thinking. Cognitive tasks are turned into principles of evidence presentation and design.” (2)
Conclusion
The key to taking action is to truly understand the primary goal and embark on a exploration derived from this goal. In this journey, you will ask questions and come up with a hypothesis.
Test your hypothesis and eventually find the answers you are looking for. Tableau software is a great for exploratory data analysis and data visualization, but always remember that creating pretty graphs is not enough to drive action.
(1), (2) – Edward R. Tufte (2006). Beautiful Evidence. Graphics Pr.