Getting Started with People Analytics

If your team has been focussed on dealing with everday general HR, workforce analytics can feel like a confusing and complicated place.

Delivering useful HR reports from a standing start relies on delivering something of value.

Work out your priorities

Identifying the benefits of developing your HR analytics capabilities is key to success.

Are you looking to answer requests from the Board or your Senior Leadership Team? Do you need to support key transformation programmes, such as cost reductions or company acquisitions? Do you want to better understand your organisation's diversity and make-up?

What data do you have access to?

Knowing what you want is only half the equation. The equally important other half is knowing what you can have.

Starting with a more readily available dataset, or one that is less complex to work with, may prove a better starting point.

Consider the data sources you have access to, such as your HR database, payroll, sales, and others. Which of these do you have or can obtain, both as a one-off and on-going. And how complex will the data be to work with. Will you need to report across multiple data sources, or create complex queries to manipulate your data? How up to date are your data sources, and how accurate are they? Are there gaps you need to consider, for example how much historical data do you need?

Work to create a balance between what would be most useful, against what is easiest to build.

Look for the pitfalls

Keep in mind some key considerations when planning your analytics. For example, think about how you will provide the reports with the required frequency. Will you need a weekly or monthly refresh? Critically, consider your data security and privacy responsibilities carefully. HR teams have access to a lot of sensitive data that must be treated with care.

Also think about how you will check your reports. You should be able to validate your numbers, perhaps by cross-checking the calculations or spot-checking the numbers for a subset of your data. Look for common problems such as how missing data affects your calculations. Often, missing data is coalesced to zero, which can dramatically bring down your averages.

Dates can similarly prove problematic. Missing dates are often represented with special dates such as 01/01/1900, 01/01/1800 or even 01/01/1753. If your employees average age is 268, this might be the reason! Other problems can arise if some of your data uses day/month/year, whilst other data uses month/day/year. Or if your reporting platform expects a different format.

Plan the technology

You will need to consider how to extract your data, how it can be transformed, where it will be stored and how you will present (or visualise) your data.

For report creation, Excel is the default starting point for many organisations. Whilst Excel is a great tool for reviewing and investigating data, dedicated reporting applications such as Power BI and Tableau offer much greater flexibility.

If you need to handle large datasets, it may be more efficient to develop your reports with a reduced subset of your data, then publish your report with the full set. It is often better to avoid having to export then import your data; if you can, import your data into your reporting platform directly from source. Reporting platforms built on Microsoft Azure or Amazon's AWS can provide the foundation for an enterprise reporting solution that grows with you.

Consider the next steps

Once you have your base report, consider how to deliver the information. Exporting to PDF is an obvious choice, but most reporting tools are able to create interactive reports that allow your customers to investigate and drill into the data most relevant to them. More advanced options include creating interactive websites and apps.

Regular reports will need their data refreshing to keep the information current. This process can be automated if you are able to connect your reports directly to your data sources. Cloud reporting platforms built on Azure and AWS can be used to automate the process of accessing and aggregating your data.

More advanced reporting techniques can help convey and explain your data. Once you have your basic presentation, consider how different visualisation styles can better communicate different types of data. More advanced options include developing what-if style reports to allow your users to explore various scenarios and extend into predictive analytics.

Helen Ramsey is responsible for strategy and operations at viewpointbi.com. Her focus is data and technology within HR. She writes regular articles on people data and analytics.

Like this article? 
Please share and follow us!

Share on Twitter
Follow @viewpointbi

To find out more, please contact us...