#databeast

There has been much discussion about the huge amounts of data gathered in any business. The data-driven culture has been described by many, including myself, as being a major goal to achieve for any organization or team of any size. But, how does an organization go from digital piles of data to a meaningful data-driven culture? How can a team garner insight from years worth of disparate data?

The bridge from collected data to meaningful data can be a challenge to build, and an even greater challenge to cross. More often businesses find themselves having to cross the data bridge at the same time they’re building it. It is a huge undertaking that should not be done without forethought and planning.

The larger datasets of today tend not to fit completely in such a well structured relational database. Other ways must and have been developed that allow analysis of data in different ways that veer off of the typical relational structure format. Instead of imposing a structure on the data, these more disparate datasets must be analyzed with the idea of finding inherent, implied, or deep-rooted relations that may otherwise go unnoticed in a preconceived structure. Relations and relevancy can come from the data itself, which can assist in providing insight useful to a business. However, this can open the door to seeing trends in data that may not actually exist or may be inaccurate. For example, the Google Flu Trends (GFT) project was based on the idea that people who had the flu or exhibited symptoms would search for information about the flu. Turned out that user habits changed over time. Many users who had almost no direct immediate connection to the flu were searching for information about it for other purposes, thereby throwing off the GFT statistics. Be aware of what data you are seeking and be mindful of jumping to conclusions about it.

We must remember that data is still only a part of the system that captured it and must be analyzed with respect to its origin. Was the system that was used to capture a given dataset optimized towards a certain goal, notion, or outcome? Does the stance from which the analysis takes place allow for objectivity when looking at the data? Objectivity is not always something necessary when analyzing data. But, one should be aware of the position of the point of view of the analyzer as compared to the data and its origin. Be sure that the analysis does not impart a bias so that real tendencies in the data can come to light and be accounted for in the interpretation. At the same time, do not over complicate matters. Be mindful to see what is actually there. If you look at something long enough, you will see in it what you want to see and lose sight of what is actually there. For example, a sales manager may sit for the better part of a day looking at sales numbers, seeing a general downward trend, and looking to find what could bring those sales numbers back up. If he looks at the numbers long enough, he might come to a conclusion that it is a hopeless situation, especially if the trend has been happening for an extended period of time. Taking a break or looking at the numbers from a different perspective, or even a friendly call or two to a competitor or colleague may bring to light a general trend in the industry itself. A general downward trend in the industry, when applied to the numbers the sales manager is looking at, could actually indicate that the sales team is performing much better than the competition. In fact, the sales manager may even notice that team may have an edge over the competition, a viewpoint otherwise missed without taking that step back to study the current industry trends. Therefore, with a slight change in perspective and taking a small action step, the sales manager sees different results from the same dataset. The industry trend bias on the data was not being accounted for in the sales manager’s analysis until the extra step was taken to look into the industry itself.

Furthermore, with so much data being captured and the growth of the data-driven culture across industries, more and more people do get access to data and analytics. This is a good thing that can bring multiple points of view onto multitudes of datasets and could generate a lot of insight and points of view, both complimentary and contrary. However, without action, insight garnered from any analysis is simply mindful chatter. To gain the most benefit from any analysis, action steps must be organized and taken for growth of any team, organization, or company.

Taming the data beast usually takes a team effort in the current data-saturated society and business world. Allowing discussions and analysis from multiple viewpoints towards the same data is a good way to hammer and balance out bias. It should be noted that an analyzer’s bias may be partly the reason for a particular insight from the dataset. Data and business coaches can help muddle through such discussions and come up with action steps to execute to reach the goals of the analysis and the business team, in general. Keep in mind, that insight from data leads to real-life benefit only when proper action is taken on it. And, once action is taken, results must be monitored and periodically reviewed to keep a pulse on it. As any industry or business matures, grows, and evolves, data relevant to it (even “old” data) must be monitored and analyzed to adjust for the ebbs and flows of any business. Steps to tame the data beast should be carried out in an organized, deliberate, mindful way to gain the most advantage from the data-driven culture. 

Better data. Better world.