Back in the day, data analysis
was nothing more than looking at spreadsheets and guessing what the numbers meant. But with increased competition, unlimited options, and an ever-changing marketplace, a higher priority on how we analyze data is needed, particularly in regards to value. The industry has shown this growing need with more than 75 percent  of companies investing or planning to invest in big data over the next two years -- a steep difference from the days of toiling over Excel sheets.
Ultimately, companies can no longer hype up their data with nothing to back it up. It’s much more important to understand the “whys” behind your data since it may tell an alternative story. Therefore, when shifting your interpretation and analysis from hype to value, three key considerations should remain top-of-mind.
1. Understand which data is relevant
Data analysis begins long before you start generating reports
. You can always gather more data, but not all of it is going to be useful to you. After all, if you're trying to measure the height of trees in a forest, it's not going to do you much good to count the number of leaves, right?
With that in mind, one of the best places to start is with understanding answers to key questions. What are you looking for? What are you hoping to learn? What do you want to measure? Those questions can help guide you in your decisions about what to track and analyze. However, understanding the relevancy of your data matters, too
Here’s an example: I recently worked on performance testing for a leading brand who was interested in minimizing the load time of their website. There are a lot of different factors in the load time of any site, so the first step was to understand what, specifically, they were trying to test. We discovered they were looking for information on the impact of loading an additional script on the page, which meant that rather than load time as a whole, the most relevant information was the difference between the load time with and without the script. Knowing that, it was much easier to gather data that was going to be relevant and useful to their purposes.
2. Know when to simplify the process
Even when you know what data you want to gather, it usually isn't going to be very useful to you in its raw form. The next step in getting value out of your data is deciding how to process it in a way that will make it easier to understand. Here, again, understanding your questions will be useful.
In the case of the performance testing, we understood that there could be a lot of variance within the load time that would be completely unrelated to the metrics the client was interested in. To filter out some of this variance, we decided to take an average over several iterations of loading the page, which would be much more representative than any individual instance. Sometimes more complicated analysis will be necessary -- based on the specific requirements -- but in this case, a simple average was enough to pull out useful information.
3. Recognize that analysis is in the eye of the beholder
After you’ve processed the data, the next step is deciding how to present the information in a way that will be useful for your purposes. While the processing step is mostly about statistics, the last step is about psychology; people can digest huge amounts of information very efficiently when it's presented in a visual report, but the format of a report can have a huge impact on interpretation.
For the performance testing, we decided that a stacked bar graph was the most effective way to present the information. This format offered a way to display the overall load time of the page, but also differentiated between the base loading time and the additional time caused by the new script. In other instances, a different graph type would be appropriate. For instance, line graphs are good at showing trends over time, while histograms are good for data sets that are easily broken into discrete intervals.
Though choosing a graph style may seem less important than gathering data or interpreting data, how it’s presented can change the “face” of your efforts. Knowing when to restructure your findings and considering which visualization method best depicts the data can be just as important as the analysis itself.
Moving beyond the hype
While many companies decide to implement analytics because they want to provide the best possible experience for their customers, it isn't enough to simply throw tracking code on a page; gaining useful insights out of the data you gather requires knowing what you want, figuring out how to get it, and presenting it in a way that people can understand intuitively.
Good principles of data analysis can guide everything from what data you choose to gather to how you choose to present it. And with numerous tools and services available in the industry today, it is much easier to understand the principles, relevancy, and interpretation of your data. In the end, data analysis doesn’t need to be all hype and no value -- it can substantial, thorough, and above all, contribute to goals of an organization.