Most companies know that data analysis will play a big role in their future. But moving into that future is a leap that defies the best efforts of many. According to various surveys, up to 85% of business data projects either fail or run into serious trouble. Here is a quick guide to avoiding the pitfalls and optimizing your chances of success.
One basic principle is simple. Don’t try to hit a grand slam before you have players on base. Time and again, the failures come from trying to do too much too quickly.
A typical company today collects vast amounts of customer data in its IT systems. The company has been told — correctly — that the potential for using this data is largely untapped. This induces company leaders to think they’re sitting on a gold mine, and they just need to start digging. CEOs recently have been very generous in approving multimillion-dollar data projects, as well as being overly optimistic about results. The road ahead often turns out to be more complicated than it seemed.
The sheer abundance of data is overwhelming to begin with. Much of it can’t easily be accessed or correlated with other data, because it exists in various raw forms in different systems. So the task is not at all like mining a rich vein of gold. It’s more like looking for nuggets in a hundred-acre junkyard.
Furthermore, data itself is only a means to a possible end. There are many, many different business goals that the data within a large dataset could help you to pursue. If you hire a data professional and dive into a big project without a well-targeted agenda, you are making the classic “Ready, Fire, Aim” mistake. Time and money will be wasted while you root around the junkyard trying to do everything everywhere all at once.
Massive projects with long timeframes face another set of risks. Data analysis is a relatively new field. The technologies keep evolving. The market is also emergent. Data vendors come and go. Your business goals may change or evolve, too, which means that multiple moving parts are in play. And a big “everything” project that takes a long time to finish doesn’t fit this situation. You risk winding up with systems full of yesterday’s data tools, aimed at yesterday’s targets, unsupported by a vendor who is no longer around.
Finally, consider the people factor. Are your key people truly on board with being data-driven? Unless they are, even the best data tools won’t be much use. Officers will continue to make marketing decisions — or lending decisions, or whatever — based on their own favorite mixtures of human analysis, personal judgment and gut instinct.
The solution to all these problems is to proceed one building-block step at a time. Instead of swinging for the fences, get runners into scoring position and start chipping away:
The idea you must share is that data analytics will help them build on their skills, not undermine or replace them. What if — by generating and consulting relevant data — they could get insights that enable them to drive revenue gains of 2X percent or even more?
Such things are quite possible with smart, targeted use of data. So don’t try to swallow the whale, lest it swallow you. Just go for the insights you need now … and keep going.