Anatomy of an Agile HR Analytics Project

August 10, 2020 | 5 minute read
Lucie Trepanier
Product Marketing Director
Text Size 100%:

This may seem familiar: You spend half your week getting data and updating spreadsheets, half your week trying to make sense of said data so you can make confident decisions, and half your week doing your actual day job. Yeah, the math doesn’t work at all. Let’s fix that, with a practical method you can use on repeat.

Let’s start with the first part—all the time you spend getting data, and managing and updating spreadsheets. Of course you know it’s a bad idea on a lot of levels.

  • The data gets stale fast.
  • It’s a security risk.
  • You’re always missing some part of the picture.
  • It’s taking too much of your time.
  • You might not catch errors.
  • And so on …

Subscribe to the Oracle Analytics Advantage blog and get the latest posts sent to your inbox

But, it’s how you’ve always done it. And lately it’s gotten even worse because of seismic changes to your business, like flipping to a workforce that’s working from home; or employees intensely concerned about health and safety procedures; Zoom replacing in-person meetings; employees managing work and kids at home; how to scale back up into a (possibly) new business model; and how to be the best possible partner to each part of your business.

That’s a lot of important factors to be juggling, and you don’t have time to wait for some big data warehousing and analytics project from IT to solve it for you. Which means, you’re doing it as best you can by pushing your applications’ reporting to the limit, and using spreadsheets.

It doesn’t have to be that way. Spreadsheets are great—I’m a huge fan myself—but they’re the wrong solution to manage data. So, promise #1: We can fix the “getting and managing” data problem.

Next comes analyzing the data and working with all your stakeholders. List out how many steps you have to take to get to a report, then to get that report into the multiple formats that your different stakeholders need. Also, list out how much rework you have to do each time report users change the question they were asking, or refine it based on your first analysis. Finally, count how many applications (including Excel, PowerPoint and email) you use as part of this process.

Again, it doesn’t have to be this way. So, promise #2: We can fix the “it’s too slow and takes too many tools” problems, and promise #3: We can fix the “many people always asking for different things” problem as well.

Remember those three promises, I’ll get back to those in a moment. But I promised you a practical method, and a practical method you shall get.

Four Parts of an Agile Analytics Project

1. Pick a question worth answering

Pick a question you’re trying to answer right now and can’t, or not easily. That’s the simplest place to start. It could be a metric or key performance indicator (KPI) that you have to calculate but doesn’t come straight out of your human capital management (HCM) application. Something where you have to blend a lot of data. This could involve ever-changing “why” types of questions. Why did these top performers leave; or why are these recruiters successful? Or it could be a “what if” scenario you’re trying to model.

Write it down, be as detailed as you can, and specify what metric or KPI is core to the question.

2. Quantify the value of answering the question

Now ask yourself why you are doing this. What impact does answering the question better or faster have on the business? Who uses the analysis? The more value, the easier you’ll be able to get support from your internal stakeholders if you need them to do something to help you. 

3. List the data sources

List out the data you need to analyze to answer the question. List out additional data that you’d like to get but cannot.

4. List your roadblocks

Where do you get stuck? What takes the most time? How long does it take? What hoops are you jumping through to do this? Think in terms of three possible problem areas: data, analytics, and people.

Once you do this quick exercise, write it out. It could look like this:

“Every week I have to provide the CHRO and EVP Sales a report on sales rep participation and linearity, and include an analysis of how rep participation impacts attainment and revenue. An increase in rep participation of X% leads to an average increase in revenue of Y%. Lately, I’m being asked to analyze why certain reps are performing better—is it their territory, is it their customer industry, is it their at-home setup—because this has a direct impact on revenue. To do this analysis, I have to blend employee data, employee location, sales data, customer data, and conduct employee interviews.

“The data changes constantly so I have to refresh it weekly if not daily. I don’t have direct access to data outside of HCM so I get spreadsheets. I’ve recently gotten some with fields missing and have to wait days for refresh. I don’t have good ways to visualize where employees are, or conduct root cause analyses on different assumptions, or run scenarios. Multiple people want to view and use the results, but since I’m not allowed to email spreadsheets with sensitive information, I have to do screenshots, then hide data, and put together briefing PowerPoint presentations, which I email.”

Eventually you’ll notice a pattern: The questions will keep changing (attrition one week, recruiters the next, benefits the following), but the roadblocks will most often be the same—problems getting and managing data, problems analyzing it, and problems getting stakeholders the insights they need. Solve that root problem and you can answer any question worth answering, with speed and agility.

  1. Roadblock: getting and managing data

Solution: Leverage all data securely, with no administrative overhead, and drive collaboration around a single source of truth with a sharable and secure workspace using Autonomous Data Warehouse. Then connect and load data with Data Flows and connect to Oracle applications with smart SaaS Connectors.

  1. Roadblock: analyzing data

Solution: Uncover deep insights fast with machine learning and augmented analytics. This means you boost your effectiveness with capabilities like automated explanation of results, predictive analytics for one-click trends and forecasts, and a single environment for all your analytics work.

  1. Roadblock: too many people need too many different things

Solution: Act with speed and agility without losing control; give your stakeholders the self-service freedom they want, but within a secure framework.

If you put all three together, you get Oracle’s data management and analytics solution for HR.

If you’ve made it this far, you probably have three questions:

  1. What does this look like? Check out this Guided Quick Tour packed with videos that show Oracle’s data management and analytics solution for HR in action. You’ll see how you can securely leverage all data, uncover deep insights fast, and act with speed and agility—all in one secure, scalable, governed, and autonomous solution. Most slides have an embedded video (just click it to play), and many also include customer proof points
  2. Who is doing this? Check out this video where Luca Ascolese from Generali talks about how data and analytics can help businesses hire the right people and predict employee attrition.
  3. How can I try it with my own data and questions? Try the interactive demo or sign up for a workshop.

Once you’ve done this once, you can use this methodology repeatedly (pick a question, define its value, select the data, identify your roadblocks) on top of Oracle’s data management and analytics solution for HR, to go after more and bigger fish to fry, reports to create, or worthwhile questions to answer.

To learn more about analytics, visit, and follow us on Twitter @OracleAnalytics.

Lucie Trepanier

Product Marketing Director

Lucie is an experienced technology marketing professional with a current focus on payment and cost-related products in Oracle's construction and engineering global business unit. At Oracle, her work in product marketing is centered on customer engagement and creating the right content to support our customers.

She has over 20 years experience in engineering, product management, marketing, project management, and sales. She holds a BS in Electrical Engineering and an MBA, from McGill University, Montreal, Canada, and can be found most nights at the dance studio, prepping for her next ballroom competition.

Previous Post

Embedding Oracle Analytics Dynamically - Part I

Philippe Lions | 5 min read

Next Post

Using Session and Repository Variables in Oracle Analytics

Guest Author | 6 min read