Machine Learning Times
Machine Learning Times
AI Success Depends On How You Choose This One Number
 Originally published in Forbes, March 25, 2024. To do...
Elon Musk Predicts Artificial General Intelligence In 2 Years. Here’s Why That’s Hype
 Originally published in Forbes, April 10, 2024 When OpenAI’s...
Survey: Machine Learning Projects Still Routinely Fail to Deploy
 Originally published in KDnuggets. Eric Siegel highlights the chronic...
Three Best Practices for Unilever’s Global Analytics Initiatives
    This article from Morgan Vawter, Global Vice...

10 years ago
Talent Analytics Isn’t Enough



Was 2013 the Year of Talent Analytics?

I think it’s safe to say that “2013 was the year of talent analytics”. Global discussions about talent analytics, as a concept, have certainly been great for our brand, and have happily led to new customers and connections.

However, when we review our client work, it’s clear that the industry is just at the very beginning of something much bigger. Many firms are now collecting better talent and workforce information.  Others are moving upwards in analytics maturity, as measured by firms such as Bersin by Deloitte.  But most have yet to transcend what Dr. Eric Siegel calls “descriptive” analytics.

Reporting last year’s employment attrition rates isn’t enough. Those employees have left your company. Such a report of past activity doesn’t connect with business objectives, which always exist in the future. Basic talent reporting elements, which sometimes pass for analytics, are just a stepping-stone to greater things.

Analytics Goal:  Measurable Business ROI

The goal in any analytics initiative is to deliver intelligence that leads to tangible business results. Not HR results, business results. For example, you might predict next year’s attrition rates and create a table of who is likely to leave, so the firm can train, intervene, and retain. Better still, the goal is to model the “top performers”, and predictively hire job candidates who match the model.

Wanted: Performance Data

HR systems deliver employee data like hire-date, salary, performance review ratings, attrition codes, competency ratings, training, and the like. But that’s only a start. None of these variables (no matter how fancy your analytics models) indicate the dimensions of value that the employee brings to the company.

Only the business unit can deliver performance data; real business performance metrics like sales quota achieved, number of calls handled in a call center, customer service scores, cash drawer shortages for tellers, how many lines of code a software developer writes, number of orders processed, and other KPIs (key performance indicators) directly measured and valued by the business.

To predict outcomes the business cares about, analytics models need input from both sides of the equation. Many in HR are focused on solving attrition challenges, without including data about performance of the people leaving.  Don’t try to solve attrition challenges without including business performance information.

We feel it’s too bad so many analysts and thought leaders relegate predictive analytics into a far-off future for HR, something only to be attempted after HR has moved through rigid data collection and software implementation steps. Companies that do this are leaving significant ROI (today) on the table.

A Million Dollars Won’t Buy Analytics ROI

Some analysts and thought leaders recommend that HR defers predictive initiatives for years, until they have fully exploited historical reporting and implemented several stages of “perfect” HR data infrastructure.

A more lean, rational approach enables the organization to deliver results from the early stages, using the data that you have now, while also building your infrastructure.  Anyone in the credit scoring industry will tell you that a little prediction goes a long way. In low-information scenarios, marginal improvements can be quite large.

It’s OK (normal) if all of the data is not yet available in a high-availability, HR system or redundant Hadoop cluster. It’s OK if there are three different ID formats across a dozen databases. There is an analytics (ETL) specialist who specifically does this work. And it’s OK if all of the variables are not yet available. Guess what – nobody’s data is, or ever will be perfect. The data journey is just that, a journey.

It is not necessary to buy a million-dollar software platform to see a million dollars worth of business ROI. In fact, you’ll spend that million dollars more wisely if you start predictive work now.

More than any amount of money, your analytics ROI will come from the experience of predicting with your data.

Benefits of an Incremental Approach

Only once analysts start experimenting with the data that you have, will you learn what you really need. Predictive Analysis and data gathering are a positive feedback loop, a loop that needs to learn.  It should be started early and looped often. As more data is available, algorithms will include them – if the variables are useful.

There are additional benefits of this incremental approach.  With incremental predictive talent analytics success, the organization can build needed political and financial capital for more infrastructure.

Talent Analytics … Needs to be Predictive

Keep working on your HRIS platform, but to stay competitive, your business needs answers now. Your business can’t wait for that system to be fully implemented.

It’s 2014. Technology, data and analytics are at a great crossroads. Talent and analytics vendors should provide analytics solutions beyond reporting, beyond gathering historical information, beyond recommending a new HR data silo.

Talent analytics, the concept, isn’t enough.

2014?  It’s the year of Predictive Talent Analytics™.

Greta Roberts is the co-founder & CEO of Talent Analytics, Corp. and the foremost expert on predictive talent analytics. Follow her on twitter @gretaroberts.
Originally published at

Leave a Reply