Advanced Data Analysis — the Process Behind the Insight

One of our most recent Adobe Digital Insights (ADI) reports focused on the evolution of digital advertising. The scope of the project was broad, including many different sources and the impact of different advertising types — search, display, video, creative resource, web traffic, e-commerce, online video, and TV Everywhere (TVE), for instance — on business. Our strategy was to create a holistic view and then delve into larger questions, develop themes, and build a story.

Five Tips for Working Through Advanced Data Analysis
With any analysis, particularly one containing new data, you start with the basics — metrics every business unit needs to understand, such as year-over-year growth or market share, before exploring new questions. Once the baseline is established, I’ve found that following the lead — rather than burying it — helps uncover meaningful insights. Here are five tips — a best-in-class “how to” for understanding the process behind the insight — for working through advanced data analysis.

1. Understand the Strengths and Weaknesses of Your Data.
All data has strengths and weaknesses, and becoming familiar with them is essential. Spend extra time talking to experts — both technical and business — to not only reveal assets and liabilities, but also determine what the data does and doesn’t measure or capture. Data from marketing technology, for instance, measures what it was built for and not necessarily the entire market. Gain a sense in the beginning by working with the data to resolve simple questions — such as share of traffic by device — and comparing those answers to other sources to see whether they make sense. Not everything you carry out in this step will make the final analysis, but it will build invaluable knowledge and experience and prevent skewed numbers from entering the process.

2. Become a Polyglot: Learn at Least Three Languages to Work With Data More Easily.
Working with such a wide range of data means using different technology stacks. We used standard SQL (web traffic), Hadoop (raw video advertising data), Splunk (TVE event logs), R (customer retention), and even Excel (yes, it still plays a role). Rounding out your skills across a variety of technologies allows for better integration of data from technology silos. Learning the difference between ‘top,’ ‘limit,’ and ‘head’ seems to be a minor point, but being able to seamlessly flow from one language to another is a critical skill.

3. Explore Relationships Across the Data — in the Right Order.
Telling the whole story around data involves finding a common thread and building a narrative. Understanding how one fact relates to another is a logical step in data analysis, but the key to maintaining structure is to prioritize in the right order. I’ve found it best to use a three-step method of understanding the basics first, then looking at relationships across data sets, and finally identifying new relationships.

At ADI, we work to discover insights not previously reported upon. In the United States, for example, we found that desktop-formatted ads were completed less than half the time when viewed on mobile devices — effectively doubling the cost of reaching that audience. This idea led to exploring the creative executions used in video ads and identifying what worked and what didn’t.

4. Review Academic Journals to Find What’s Topical in the Academic World.
Another often-untapped resource for analysts is the world of academia. Explore research and journals, and if you find something interesting, reach out to the author or authors. Talking to academics can lead you to a new line of questioning. Based on research surrounding the increased “cost of attention” in advertising, we compared the cost of digital advertising to inflation (ad cost is rising much faster) and looked back in history to see when the introduction of technology changed how consumers interacted with content and, ultimately, affected the cost of ad delivery (1997 seems to be a tipping point).

5. Aim for Continuity.
People make decisions based on relative changes rather than absolute numbers. For example, “10,000” doesn’t reveal much, but “an increase of 10 percent” does. To uncover those meaningful comparisons and trends, you need to have baseline data. However, as industries evolve, research methods do too. If you’re changing how things have been done in the past, be respectful of what came before and do your best to build from a common foundation before venturing into new territory. That said, uncovering strategic insights — the “I didn’t know what I needed to know” facts of life — often comes from looking at different perspectives.

In Conclusion
Fleshing out threads of information that will uncover meaningful insights in complex data requires a strategic process. As an analyst, you need to develop (through repetitive work and exploration) the ability to understand relationships between data sets. True insights come from looking for context, exploring academic research, developing themes, and following the lead. Don’t be beholden to preconceived notions of hypotheses — instead, let the data take you to the story.

The post Advanced Data Analysis — the Process Behind the Insight appeared first on Digital Marketing Blog by Adobe.

from Digital Marketing Blog by Adobe https://blogs.adobe.com/digitalmarketing/adobe-digital-insights/advanced-data-analysis-process-behind-insight/

via Tumblr http://euro3plast-fr.tumblr.com/post/159407960214