5 Data Analytics Takeaways from SHRM13 to Help You Go From “I Think to I Know”
If you have been following our writing on InternMatch you know we are big advocates of using data and analytics to improve university hiring, and other HR practices. We have written posts on why your next HR hire should be a data scientist as well as talked about how Prudential uses data to build a more diverse intern program.
I was excited to see that Data and Analytics was one of the major topics at SHRM this year and am thrilled to share some key takeaways from the conference on this topic. These insights were gleamed from a number of sessions but one of my favorites was done by Cathy Missildine, from Intellectual Capital Consulting:
1.) Embracing Data, Means Embracing Change
10 years ago Facebook didn’t exist. Now companies like Starbucks are making hires on Instagram. Data-driven decision-making is powerful because it allows you to identify the new technologies that truly drive your goals, and ignore the rest of the noise. However, data is only as good as the goals you set. It can easily tell you whether attending Career Fairs, posting positions on InternMatch, or creating an internal referral program is providing you with the lowest cost per hire, but it cannot tell you if cost per hire is the best goal for your team. This is why it is critical to constantly re-evaluate, update, and change the core goals you are tracking to fit your company’s changing top-level needs. Ideally you should stick to 2-3 core goals and ask yourself at least once a month if those goals are still relevant and correctly prioritized.
2.) Start with small wins
Like any skill, some data projects are easy, while others are incredibly complex. If you or your company is new to using data within HR, don’t try to build Rome in a day. There is plenty of low hanging fruit to start with. For example, you might want to start by building a funnel for your interview process that tracks candidates by source, conversion % from application to interview and conversion % from interview to hire. This is valuable information as it quickly shows which sources are sending you the most qualified candidates and it is relatively easy to setup.
Once you master campaigns like this you can start to look at deeper problems, like building and testing hypotheses’ for the high turnover rate at your call center!
3.) Use Data to Tell A Story
While data is cool, in it’s raw form it is certainly not sexy. The power in data is that it helps you go from saying “I think” to “I know,” which is incredibly powerful when trying to convince your team, CFO, or CEO to make changes or give you increased budget for new campaigns. When making your case or sharing new insights from data-driven campaigns, don’t just send people a big Excel file, instead summarize your insights into a slide deck that tells a clear story. A perfect slide deck will focus on how the changes you recommend will drive top level business goals and will include an “Aha Slide” that sums up in clear language why making the changes you are recommending are valuable. For example, if you find that you get twice the number of hires at half the cost, by using Facebook ads vs. attending Stanford’s career fair, your aha slide would say, “We can save $30,000 while increasing quality of hire.” Enough said.
4.) Correlation does not equal causation.
When you start getting into the higher levels of data science best practices you are going to start using regression models to determine how impactful variables are on your end goal. For example if you have a high turnover rate at your call center turnover you might hypothesize that this is due to a reduction in your number of training sessions, a lower number of years of experience required on your job posting, or because you changed your benefits package.
In any standard regression model, you will ultimately solve for R-squared, a variable that tells you how correlated your variables are to the change you are testing against. The closer R-squared is to 1, the stronger the relationship is between your variable and your hypothesized result. That said, the golden rule of data models is that correlation does not equal causation so even if you get a highly correlated r-squared it is critical that you give all your results a gut-check analysis to see that they make HR and business sense as well.
5. ) You don’t need a PhD
It is important to note that if you are interested in becoming the data guru on your HR team, you shouldn’t be intimidated. Any HR generalist with a reasonable amount of comfort with math and Excel can learn the skills needed to start creating significant value for your team. Still not comfortable? Thanks to new online education sites like Udemy and Coursera you can learn how to work with data bases or do high-level Excel number crunching for free or cheap all from the comfort of your home! (Heck maybe your company will even pay for it).
For example, calculating something like average cost per hire by university is as simple as formulating total cost of your campus visit (airfare + hotel + career fair + schwag) and dividing this by total hires that came from that visit. Tracking this for every school over a couple years will easily allow you as a campus recruiter to know which schools are worth attending and which should be cut from the list.
As we enter a time period where we become more and more inundated by new innovations and distractions, it becomes increasingly important to be able to manage and understand all these inputs. This is especially true for HR teams who have lean budgets and need to make hard decisions about where to invest their scarce dollars and time. Data analysis will help you choose between continuing to put dollars into Twitter or launching a new mobile site. It will help you choose between your winning sources and those that just feel good because you think they are sending you a lot of candidates. And thankfully in today’s age more and more companies are producing tools to make managing and drawing meaningful conclusions from data easier than ever in the past. It is time for you to stop talking about getting serious with data and train someone on your team or make a hire so that can outcompete your competitors for talent in the quickly evolving HR landscape.