With the rise of Big Data, companies have been scrambling to collect as much actionable data as possible, both externally and internally. As noted in a recent YDT infographic, 64 percent of surveyed corporations have invested or are planning to invest in big data; and, that percentage is even higher among Fortune 100 companies. This is a good first step, of course – companies cannot properly analyze data if they don’t have it – but sometimes the focus gets caught on the data collection and not enough on the getting the right analytics and execution strategy. Or, as one recent Harvard Business Review article put it, analytics can end up becoming “an unsolved puzzle with the pieces flung all over the floor”. There is work to be done – for companies to make better use of their data so that it can drive business decisions and strategy. Granted, the corporate situation is different for each company, thus making the question “What should I do with this data?” difficult to answer with a concrete list of simple options. However, there are some analytics failures that are particularly common, and it is good practice to evaluate whether your company is falling into any of these common mistakes.
Mistake 1. Not knowing the cleanliness of the data
It is important to know the cleanliness of the source data and how much it will need to be manipulated. This comes in two forms, chiefly. The first is trying to gauge the amount of time it will take to get the data into a usable format; this is more important from a time management or project management perspective. From a data analyst’s perspective, it is also important to know the quality of the data being dealt with. If the data itself is not usable or valuable, then even the best analytics strategies will be powerless to make informative action plans.
Mistake 2. Assuming correlation is important
We all learned in basic psychology the old maxim that “correlation does not equal causation”. The problem is that the human brain is still wired to look for patterns, and to seek causes to observed effects. This of course is important – we want to dig to the root of why the data acts the way they do. It is still worth mentioning that human instinct is to jump to conclusions. If website traffic is up, are we sure it is because of that new PR push or redesigned website, or are there other potential factors (perhaps even seasonal)? This sounds like a basic example, and it is, but for the enterprise IT manager, it is still easy to make such a fundamental error in reasoning.
Related: Virtualization Manager, helps administrators manage resources, capacity and performance in the virtual environment
Mistake 3. Not having the right analytics team
Having the right team in place is another big issue for companies, and this one is trickier than #2 on this list. One important component to an analytics team is making it interdisciplinary – combining forces from engineering, management, data science, and perhaps more. This might sound like it will make the team size oppressively large, but it is helpful to have a well-rounded perspective on the data and how it should be used towards business objectives. Collaboration among these different bodies is another key, so for that reason and others, proper leadership is an important aspect to this analytics team. A good leader for such a team would need to be, not just technically competent, but skilled at gathering various opinions while keeping in mind the overall business vision or goal.
While these three do not comprise the total of all possible analytics mistakes, they would go a long way towards helping any company keep a well-running analytics program. While your company might not approach the level of an InsightSquared, Paxata, or Trifacta – three big data companies that were among the most recommended in 2015 – keeping these potential pitfalls in mind will greatly help any enterprise IT company make better use of their data.