If you ask a group of big data analysts what are some of the top professional challenges they face, you can easily imagine which challenges are consistently shortlisted. One is the amount of time and effort needed to get data ready to analyze before anyone gets a chance to actually start analyzing it. That’s time spent ingesting the data, cleaning it, normalizing it, and so on.
One high ranking challenge is data sharing. This goes to the heart of why big data analytics is so critical in today’s connected environment. Increasingly, you now gain valuable insights by connecting the dots, i.e., connecting relevant information items, that are widely dispersed among departments, customers, supply chain partners, and elsewhere. So while you can only analyze the data you have, you also don’t control or have easy access to much of the data you need. That makes the data owners’ willingness to share data a key success factor in any successful big data initiative.
Willing to share is good; motivated to share is better. Because analysts often don’t know what they don’t know (which is, after all, the point of analysis), they may not know what data exists or what to ask for. A motivated data owner, who’s probably more familiar with the data than the analyst is, may point the analyst in the right direction.
How are people sharing their data?
So Why Won’t Data Owners Share? Anyone who has dealt with this problem can probably recite many of the reasons data owners give for not sharing. Here are a few we’ve heard:
“How’s this going to help me? Why should I care?”
“If everyone knows what I know, why do they need me?”
“I don’t want other people messing with my data.”
“There’s stuff in my data that might embarrass me.”
“Data sharing is not my job, and takes me away from things that are.”
“Sharing data is technically hard.”
It is human nature for people to resist doing more work if they perceive the effort to be difficult with little value to them personally or, worse, actually threatens their livelihoods or reputation. This is especially true when it involves information that might reflect negatively on their performance and could actually hurt their performance. Organizations need to consider the following factors to determine how to make sharing data successful.
Wrong Ways to Get People to Share Data. As in most areas of life, there are some less than effective , ways to encourage and motivate people to share their data. Let’s start with some of the not so good:
Just tell them to do it. Few people like being ordered to do anything. This goes double for data owners getting orders from people they don’t report to or who aren’t their customers. Even if there is a formal reporting or customer relationship, most humans will naturally want to rebel against this kind of behavior (often in ways that are not apparent).
Don’t involve them in decision-making. If you’re asking someone for their data, that person is probably a professional knowledge worker just like you are. So asking nicely might not be enough. It may come off as patronizing, especially if you don’t show them respect as professionals by asking for their opinions or any context you might not be aware of.
Don’t give back any value. “You gotta give in order to get” is a core maxim of marketing that applies very well when it comes to marketing your data requests. Show data owners how your insights will reward them, such as by revealing new ways the data owners:
Give people more work to do but not more resources with which to do it. Just increasing people’s workload without additional resources can backfire in multiple ways. For one, it provides an easy excuse for failure — they can easily point to more pressing priorities. For another, it sends a message that the extra work is not critical, because if it were, then it be adequately resourced.
Right Ways to Get People to Share Data. You can start by doing the reverse of all the points just mentioned. But here are some other ways to do this right:
Create a safe sharing environment. If people are unwilling to share data, it may be a sign of bigger problems, such as a very political culture where everyone plays a zero sum game (i.e., I only win if you lose and vice versa). If people are going to share data nicely with others they need to know it won’t be used against them. A good place to start is by not setting traps for people in general, where punishment often follows compliance.
Make this a collaborative effort. Asking people to contribute, and giving them key roles in which to contribute, will create loyalty to the big data initiative (beyond loyalty to their own functional silo). So they’ll more likely be motivated, not just willing, contributors.
Make sharing easier with the right tools. This echoes the “more resources” point above — except emphasize that some resources (like tools to ingest, clean, and normalize data) make data sharing easier for the data owner too, not just the analyst. Data sharing will happen faster, with less hassle, and less disruption of the data owners’ other work.
These all seem like simple precepts to follow, and they are. And if more organizations would actually follow them there would be much less resistance and big data analytics would be much more efficient and effective for many more organizations.
See how people sharing their data