The Problem with Processing at Light Speed
We’ve come so far with being able to work with data loads, figure out what we target to learn and then be processing those workflows at break-neck speed. It’s really quiet remarkable how quickly information can be processed, especially once you’ve identified the processes and rules you want to apply.
Of course the old addage, "garbage in, garbage out" still applies as well.
Take the Microsoft "oops," for example. In the last day or so, they issued copyright take-down demands to YouTube for a whole host of videos. The only problem is, it was a case of moving too quickly with too few checks and balances.
Here’s a look at what went down (or had to come down, as the case may be).
Yes, they’re correcting it, and acted quickly to do so, targeting the actual information bits they were concerned about.
But the fact is, someone was going through big data – unstructured, high-velocity, high-volume information – and making rules and decisions, then acting on them without vetting those decisions.
This is one of the challenges we all face as we work with more and more information. We also have to create new ways to pay attention. We have to create rules and processes and wisdom about working with these information stores and making sure that the inferences, assumptions and learnings we take from the processing are valid.
Not only that, but we have to make sure that the actions that come from these analysis bits are valid courses of action as well. I think this is a critical responsibility as we work with information. As data professionals, we have to be the "yeah, but…" people. We have to help point out the fact that whatever information or guidance you’ve found in the information is based on certain assumptions, certain rules and considerations. From there, the results have to be checked as well. So not just the processing, but what you’re doing with the information. It may be that the rules you’re applying do indeed work, but the way you apply the results just doesn’t make sense.
You know, like the Microsoft oops. The data is good (the license keys were being given away) but the action taken (remove the videos outright) didn’t match up.
This is really easy to do in a business environment too. You make assumptions based on the findings, but the findings weren’t found based on the information that would support your assumptions and actions. Read that 10 times fast…
We have to be the voice of reason on the data use, as well as the data protection and processing.