This year, I received a lot of emails from software vendors about products promising to make back-end data management tools almost irrelevant. The idea is to create a data visualization and analysis tool that's simple enough for any business manager or worker to use and that comes with prebuilt data connectors to link up to a range of data stores -- relational databases, NoSQL systems, Hadoop clusters -- without requiring heavy involvement from IT staffers or data engineers.
It's an interesting proposition that's worth considering. If software from the likes of Tableau, Tibco and DataHero really are as simple to connect to disparate data sources as they claim, it could be a boon to business users. But companies need to put those claims to the test and make sure their internal processes are ready to handle the simplified connectivity, before investing in such tools.
The potential benefits are obvious. The only reason a business would want to spend a lot of time managing data is to make it useful. But data management challenges can be huge, often requiring intensive efforts to set up data pipelines just to get relevant information into an analytics tool. The pitch of the vendors pushing the connectors in self-service analytics tools is clear: What if you didn't need to worry about all that work?
Obviously, no matter what tool you go with, you're still going to need to set up some kind of data store. But in a world where front-end applications seamlessly connect to any type of store through data connectors, the particulars of the back-end systems might not matter quite so much. This would limit the amount of time it takes to prepare data for business users and let them put it to use more quickly.
But at the same time, there are some obvious risks. When self-service business intelligence and data visualization tools are implemented by individual departments and business units, it can create data silos. These silos often make it difficult or impossible to scale data analysis practices that work well throughout an organization. It's also not hard to imagine a scenario in which different departments are duplicating efforts on BI and analytics tasks that could be done more efficiently by a centralized team.
And then there's the issue of interpretability. A marketer or sales rep might be able to quickly and easily create a report and derive what he thinks is an important insight by using light-touch analytics and visualization software -- but if he has no formal training in data analysis, it's hard to know for sure if the findings are statistically significant or data sources have been combined in ways that maintain the accuracy and consistency of the data.
So, depending on how you look at things, it's either a time for exploring new possibilities or exercising extra diligence. There are plenty of vendors promising that their software's data connectors will make these types of links seamless, and some of them might even deliver on that promise. But even in the cases where they do, there are still likely to be governance and scalability issues that no piece of software can solve.
To avoid those problems, businesses need to be sure that any tool they implement fits in with their broader data strategy. That strategy needs to come from, or at least be approved by, senior executives, and it should address things such as who gets to access particular data sources and for what purposes, and how BI and analytics results can be used. Organizations will find they that have no shortage of options for helping business users become more data-driven -- but there are some policies and procedures to iron out first.
Another option for simplifying analytics: self-service tools for data preparation
Business users don't always embrace new data analysis tools
In creating a data-driven culture, it's what you do, not what you buy that counts
How a data analysis tool helps fight drought