Data, the Ultimate Storyteller

by Becky Tasker

posted on 09-06-2016

I was recently tasked with putting together a “best of the best” report for the Asia Pacific region. Best of the best reports are analyses of websites’ average performance, as well as the ranking of the top 20 percent of performers in specific countries and industries on various key performance metrics. Thanks to the Adobe Digital Insights data repository, I had access to a tremendous amount of data. I’d led four of these reports for other regions, so the enormity of the data set wasn’t new to me. However, the region was, and I was excited to explore and learn about its various cultural and regional distinctions.

As an analyst, the challenge is to translate data into actionable findings for your audience. Data can show you the numbers. That’s the science. But crafting the story behind the numbers—that’s the art and where the real work of a good data analyst begins.

The strong set of data available to us at Adobe Digital Insights allows us to come up with real-world insights that marketers can take back to their own companies. Here are my thoughts on how to tackle enormously robust data, unique cultural frameworks, and, sometimes, your own sense of being small.

Dealing with—and Benefitting from—Extremely Robust Data

For the 2015 Asia Pacific Best of the Best, we pulled aggregated and anonymized information from more than 100 billion visits to 3,000 Asia Pacific websites for all of 2014 and 2015. One of the first challenges I ran into was the sheer size of the data. Our normal go-to tools couldn’t handle a report of this magnitude. The data didn’t fit into our framework and we had no choice but to expand methodologies.

As analysts, we’re tasked with both producing meaningful, useful insights, and providing direction and strategic guidance so our audience can come up with their own insights and strategy. For example, looking at smartphone visits by country, I saw that in Japan, an average of 38 percent of total website traffic comes from smartphones, compared to only 28 percent in the U.S. What’s more, for the top 20 visited websites, nearly 60 percent of reported traffic comes through a smartphone. Clearly, not only were companies in Japan leading in this area, they were way ahead of the curve.

Although the size of the data set precluded me from giving specific examples of how top performers were making the experience engaging or attracting traffic, I was able to provide marketers with the data they needed to draw conclusions and make judgement calls about how to apply the results.

Takeaway: In the face of robust data, don’t let tools limit your findings. Search out and identify alternative methodologies to get at the data. Finally, be sure to qualify your story line and speak to it appropriately.

Look to Your Own Experiences to Guide Insight and Recommendations

Analysts are often expected to know every bit of data and everything that goes into it. The truth is, that’s impossible. Sometimes it’s helpful to take a step back when deciding what’s most important to include and what should be left for readers to determine and investigate on their own. How do you make these decisions? In this case, I drew on my experience as Director of Retail and Email Strategy to determine the kinds of things that were important to marketers. Since the data showed that consumers were shifting to smartphones, I knew a discussion about the winners, or even those underperforming in this space, would be interesting and meaningful.

Knowing from the data that companies were successfully attracting traffic on smartphones, I wanted to understand what that meant for marketers further down the funnel. It could be anything from a retention metric on a website (are visitors leaving or staying?), to an engagement metric (how many pages are they interacting with?), all the way to a conversion metric. It mattered not only where traffic was coming from, but also how visitors were interacting with the website’s content.

Takeaway: Taking a step back from the data and looking to your own experiences can help highlight those recommendations that are important and meaningful.

Discovering the “Why” Behind the Data

Although I didn’t know exactly what Japanese companies were doing, I knew the data was important. I also knew that companies in India were showing increased activity. Yet, interestingly, the data showed India’s smartphone traffic at 28 percent—low compared to the others. The data was right, but to explain it I needed input from someone more familiar with the Asia Pacific market. From discussions with my director, I learned that India as a country has bandwidth issues that cause them to lag behind their counterparts. This was an “a-ha” moment for me. Knowing about the bandwidth issues and looking at some of the other metrics, I was able to tell the story.

Takeaway: With underlying issues explained, perspective changes—helping you to identify strategies your audience can use. It would have been easy for me to leave it at a statement that that India is underperforming. Instead, I chose to try to discover the why behind India’s lackluster performance. In doing so, I was able to discover what was really happening, and that helped tell a more complete story.

Helping Your Audience Create Individual Strategies Using a Macro-Level View

As marketers, making sense of macro-level reports often requires examining internal data. An increase in one metric for company A might be positive, while the same increase for company B can be negative. Marketers often find macro-level views to be applicable and good guidance. But it requires identifying where a company’s breakdowns are occurring within its own purchase funnel and optimizing toward those breakdowns. At the end of the day, it comes down to identifying individual strategies based on external input of your own data.

Analysts decide how to present the data set to arrive at insights that tell a story. But revealing good insights sometimes means cutting a data set two different ways. If you want to provide insights that align with information that’s useful, strategize beforehand. We create a starting point for what we’d like to reveal using a pre-analysis form. Then, we look at the data to see what it reveals and how it matches our expectations. If it doesn’t match, that’s an insight too, and we include it because it’s unexpected.

There were some surprises, which were the direct result of looking at data differently.

For example, the automotive industry ranked lower on some engagement metrics within the Asia Pacific region than other industries. This finding might just reflect a U.S.-based company trying unsuccessfully to market to people in a different geography. However, instead of looking at the data vertically, I looked at it horizontally, highlighting additional differences and similarities, making it possible to pick up unseen macro patterns.

Takeaway: If you know your audience, a pre-analysis can help you more effectively craft what you want to say, build the report, and identify those insights that would be most revealing to your audience. Consider looking at the data in different ways to reveal unexpected patterns.

In Sum

How do you tackle a robust data set and turn your findings into translatable insights for your audience?
In this instance, in the face of incredible amounts of data, we were careful not to let our tools limit our findings. We also made sure to qualify our story line with data and speak to it appropriately. I sought out expertise when needed, and talked through many of the issues I couldn’t explain.

When I finally had the a-ha moment—when I identified the resources I had available and applied the input—I was able to more effectively deliver meaningful recommendations. A pre-analysis helped craft the report, identifying insights that would be useful to our audience. Finally, discovering the “why” behind the data was a big part of understanding what was happening behind the numbers. And like any great analysis, the explanation makes all the difference in telling the story.

Topics: Analytics