Making data work for talent pros
by SwoopTalent, on October 17, 2017
Big data! Small data! Pudgy data! Kind of average middle sized data! I want to talk about big data for talent, but that's already kind of taboo. Sad face.
What's the deal with "big data" anyway?
The biggest deal is that the term "big data" already makes most eyes roll right across the room. It's already achieved despised buzzword status. But that's OK, we can call it what we like. The important part is that the dataset we need:
Includes both structured and unstructured data
Is in disparate datasets
Exists both inside and outside the enterprise
Is almost never completely accurate or precise
Has no single data dictionary or definitions
Contains relationships, opinions and contradictions
Does not have universal unique identifiers
So when I'm talking about "big data for talent", that's what I mean. As far as datasets go, if you get it right it will be a little tougher to work with than your average spreadsheet. But it will let you unearth a lot more insights.
Why can't we talk about big data with talent data?
Well, THIS is embarrassing. We're an industry full of smart, motivated people, right? Yet, we are apparently "set to fail the big data challenge". There's also "no such thing as big data in HR" (good luck telling Humanyze, my team or even LinkedIn that). Oh, and we're also not treating people as humans. And that's only the START of the objections you can find when you start looking.
Althought I want to say PFFFT to all these naysayers, I confess: there's something serious here. Talent analytics involves humans and human behavior is complex to say the least. That means analysis of it is challenging, and predictions are tough. If you want to dig into data in complex systems, Scientific American's "Big Data Needs a Big Theory" is for you!
The trouble is, we don't have a unified, conceptual framework for addressing questions of complexity. We don't know what kind of data we need, nor how much, or what critical questions we should be asking. “Big data” without a “big theory” to go with it loses much of its potency and usefulness, potentially generating new unintended consequences.
Talent questions aren't the most complex system, of course. But the wider talent data ecosystem does get complex. Embrace that! One of the bigger risks we face is falling back to the simple, easy to manage, Excel style data we have always used. BUT...if we don't make that mistake:
There's a lot of potential upside if we get the data right
McKinsey gave three solid (and pretty easy) examples of how analysis on a broader dataset can uncover insights. Recruiting Daily pointed out that recruiting data is a valuable asset. And 41% of organizations invest in data-driven initiatives to "change the way we organize operations". OK, that last one raises a few alarm bells, but it's still a BUDGET for us in there! Lots of potential here, folks.
What do we have to do to get there?
There are places we need to focus our energy to make talent data work as well as possible. Here's a short list of mine:
Curation. Data curation is the integration, management and publishing of data. Curation gives you a clean, comprehensive dataset to consume how you need it. How nice does that sound? It can help a LOT: “organizations that offer users access to a curated catalog of internal and external data will realize twice the business value from analytics than those that do not.”
Democratization. Make talent data available and accessible to the "average" end user. Yes, there's a lot baked in there, with security, coherence, comprehensibility and so on. Worth it.
Connection. We gotta get rid of our data silos, people!
Visualization. The single best way to make data accessible to non-nerds is in good visualizations. Counterpoint: The single best way to lose their interest altogether is with sh*tty ones, so take care
Hypotheses and Intuition. I'm a bit old-school here, but many of the best insights on talent started with a hypothesis or intuition. From the hypothesis, analysts sought evidence in the data. Sure, they often proved their original hypothesis to be WAY off the mark. So what? These hypotheses lead to discovery of patterns and correlations and all that goodness.
Let's do this!
As Bill Schmarzo points out in a pretty great "harsh message for big data vendors", the #1 problem is still "determining how to get value from big data".
And in talent, that's our GOOD news. A lot of excellent practitioners have been leading the way on this for many years. New ideas will emerge as new data sources come into play, but they will evolve from what we have. We know what questions to ask of the data, at least for a good start.
And if that doesn't work, at the very least you can learn how to call B.S when you need to!