Big data holds huge potential. Most businesses have undertaken big data initiatives or plan to soon, but not everyone is successful. In fact, many of the companies that jumped on the bandwagon the fastest have already jumped off. Why? What causes well-funded, good-intentioned big data initiatives to fail? Here are six signs your initiatives are on the wrong track.
1. You’re Trying to Answer Specific Questions With Big Data
Big data has a way of producing more questions than it answers initially. If you need quick answers to relatively simple questions, regular analytics is a better choice.
Big data, unfortunately, isn’t about answering specific questions or dealing with limited, compartmentalized issues. Big data is broad, both in definition and in scope. While it’s imperative to have clearly set goals before adoption, it’s also important to step back and let the data show you what it has to say. If you’re looking for specific solutions to specific questions, then ordinary business intelligence and analytics are cheaper, easier, and deliver results much faster.
2. The Data is One Dimensional
Lots and lots of data isn’t necessarily big data. Big data is a lot of data that is disparate and comes from numerous sources. If you’re trying to use big data analytics to derive insight from a ton of one-dimensional data, it’s a bit like trying to go to Grandma’s house via an F-15. You’ll spend far more time, energy, and money getting where you need to go than you should. Unless you’re analyzing multiple data sets from different sources, it probably isn’t worth it.
3. Your Business Folks and Data Scientists Aren’t Communicating Well
If your big data plans are off track, it could be a failure to communicate. Make sure your data science team and business folks are speaking the same language.
You know the stereotypical geek on TV and in the movies who always gives an answer ten times too long and complicated for anyone else to understand? This bit is funny because there’s a bit of truth in it. Technical people love the technology and understand it intimately. They just have trouble boiling everything down to what non-techies need to hear. At least one member of your data science team needs to be able to speak business-ese so that what the data is trying to tell you doesn’t get lost in translation.
4. You’re Having Data Management Issues
Data ownership can get murky when you undertake a big data initiative. From the moment you begin the data transfer process, there are questions about who can access the data, who can use it for analytics, where it can be stored or transferred, and what third party access is allowed, such as cloud services and big data solutions vendors. This problem is easy to right with a well-planned, thoroughly-documented data management policy.
5. You’re Trying to Get Quick Answers From Big Data
An investment in big data is more like building a stock portfolio and less like day trading. You don’t see results overnight, no matter how much money and effort you pour in. Many businesses scrap their data projects when they don’t achieve measurable results in three to six months, but it can take a year or more for big data to begin yielding definable benefits. Hang in there for the long haul, and your efforts will pay off.
6. Your Organization Lacks the Necessary Skill Sets
Finally, big data initiatives sometimes fail because the qualified personnel wasn’t available. Big data scientists command a starting salary of $100,000 per year, and in some parts of the country and certain industries, this sum can easily soar to over $200,000, especially for a data scientist with a proven track record.
Yes, it is certainly possible to produce a data scientist in house, but you’ll need to add the tremendous learning curve to the time it takes for a big data initiative to produce results. A scientist developing their skills will need to be allowed the time to learn the technology, to learn the ins and outs of data science, and then to work through the plans to produce the actual results.
The one key element in all of these warning signs is that each one has a solution. It isn’t necessary to scrap your data plans if you see one of these signs. Just rethink what’s causing your lack of success and take steps to resolve the problem.