Analytical HR: 10 lessons learned on the use of metrics in People Management


Analytical HR: 10 lessons learned on the use of metrics in People Management

People Management techniques are increasingly based on metrics and data. See some experiences that RD had working like this!

HR has been reinventing itself, and many companies and professionals need to adjust to this new market context. In the same way that new technologies are making areas such as Marketing, Sales  and Customer Success  more strategic and intelligent , areas known for being “more traditional”, such as People Management, have also been improving and remodeling in recent years.

Maia Josebachvili, former VP of Strategy and People at Greenhouse Software, explains in her article HR Reinvented: The New People Teams  how the People Management area has migrated from a traditional approach to a more consultative, deliberative, holistic and strategic approach.

In this new context, the People Management departments have been defining their indicators so that they are linked to those of the business , thus seeking more clarity on the impact of the area and alignment with the company’s strategy.

It’s still very important to keep track of those metrics  more operational , such as absenteeism level, average time with the company, rate of labor claims, ratio between overtime and hours worked, among others. However, more robust and strategic metrics have been gaining ground, such as leadership assessment, performance management and employee engagement.

Here at RD, the People Operations team has the mission of supporting the processes and projects of the Talent Management (TM) area, as we call our HR, as well as measuring, monitoring and analyzing the KPIs  related to People Management, proposing suggestions for improvement from hiring to climate management and leadership.

Be clear about what needs to be measured

It is important to emphasize: be clear about what you need to measure , not what you want to measure . Sometimes, here at RD, we come across indicators that we thought were important to monitor, but which in fact were not aligned with the company’s objectives.

For example: at the beginning of 2018 we thought about defining a metric that would calculate the percentage of internal promotions per RDoer  (as we call ourselves here). However, the direction so far was to focus on external hiring. Even though it still encouraged internal movements in the DR, this indicator was not in line with the adopted strategy.

In a webinar , David Harcourt suggests asking the following questions when defining an indicator:

  • What are company executives worried about? (for example, is it with talent retention, with the succession plan or with leadership development?)
  • Where are we losing money? (when hiring non- fit talent  with the organization or with those who leave prematurely?)
  • What is our competitive advantage? (For example, in RD we have a strong professional development culture, so how do we define metrics that support and measure this advantage?)

Link the business indicators with those of People Management

Reinforcing the previous point, when defining the metrics, it is essential to know those that are monitored by the board, associating them with those of the area.

In 2018, RD focused heavily on the experience of its customers and partners and adopted NPS  as one of its main KPIs. Following this premise, the TM area defined  eNPS as its main key result , which is associated with  the company’s OKR .

At first, you won’t know what to do with the data…

…and everything is fine! When we are in the process of defining metrics and collecting data, we are not always able to make decisions right from the start. This happens because, many times, we don’t know how the numbers behave, what the trends are and what is a good or bad result for the organization.

Even so, it is important not to give up in the first moments and continue to monitor and build a history of these indicators. Benchmarking with other professionals working in the same area can be useful to get a “second opinion” and understand how the metric is defined and what can be done with it.

Here at RD, we monitor the eNPS month by month and, in the first few months, we didn’t know what to do with the data we collected. We did not understand the triggers we needed to be aware of and how to direct the improvement actions together with the area, mainly because it is an indicator of climate and engagement.

Something that was very rich at that time was adapting the focus group practices applied in other companies, so that we were able to go deeper into this diagnosis.

Over time, the numbers start to make sense and with that come project insights

Our premise, in our People Operations team , is that indicators justify project demands. These tend to become processes whose results are measured by indicators. We appreciate that every project in the area is linked to some problem we want to solve, which is evidenced by the metrics we calculate.

It is important to mention that different indicators can be associated, but they have the same cause and, therefore, the same possible solution. This was the case of the salary table revision project, carried out by our Total Rewards team , which is responsible for the compensation strategy and benefits package. The project was based on the results obtained in the Great Place to Work climate survey , on the feedback from interviews of dismissal and decline of proposals in the selection process.

But it’s not just with robust projects that day-to-day problems are solved. Our technical recruiting team managed to reduce the time to fill vacancies from 40 to 20 days, with just a change in the way we manage the hiring pipeline.

Prioritize the smallest projects and learn fast from mistakes

According to David Harcourt , when defining a project, it is important to assess the effort that will be expended and the added value, as shown in the figure below.

He does not recommend that we spend time on those projects that are high effort and add little value. On the other hand, those with high added value and low effort are called “unicorns”, very rare and hardly a reality in the business context.

David recommends starting with projects whose effort and added value are low, because they are modest, but generate impact on the organization and the need for little data, since this is the reality of most companies that start working with data analysis. That way, you get faster results and you can make mistakes and learn faster from them.

Don’t define the solution before understanding the problem

A mistake that is constantly made by organizations (and here in DR it is no different) is to define a solution before actually understanding which problem is at hand. An example of ours was the project to acquire a new management system for the selection process.

We had the belief that we needed to change software to solve the operational problems we faced on a daily basis. We started the search for a new supplier, but there was great difficulty in defining which was the ideal tool for our needs.

The main mistake in this case was that we arrived with the defined solution (in this case, acquiring new recruiting and selection software), but we did not take the time to carry out a diagnosis of the process and identify what the main pains in the area really were.

Because of this, we had rework and it was necessary to take two steps back and measure which were the indicators that were most impacted by process inefficiencies and how the project would contribute to improving those.

It sets out 4 steps to defining a problem:

  1. Establish the need for a solution : define the scope of the problem and what is the basic need we want to meet;
  2. Justify the need : does this problem affect the organization’s objectives?
  3. Contextualize the problem:  which context is this problem inserted in?
  4. Write the problem statement : complete description of what the problem actually consists of.

In all these steps, having clear metrics is essential to fully understand what needs to be resolved.

Action today does not solve the problem tomorrow

This happens especially in People Management, where a large part of the challenges involve longer and more complex processes, such as the organizational climate, training and level of employee satisfaction. We learned this after a lot of “breaking face” trying to improve indicators from one month to the next.

A clear example was the leadership assessment. Before we understood how the metric worked and what actually influenced it, we thought that if we trained the leadership team in one month, we would have results the very next month. This was not the case for the simple fact that this type of action had an impact in the medium term, as it was about people development.

A good practice for this type of indicator is to set semiannual or quarterly targets – not monthly. Thus, the area can have more time to diagnose the problem (following the steps suggested in the previous item) and work on solutions.

Combine the quantitative with the qualitative

It is quite common for us to try to investigate some phenomenon or calculate some indicator through surveys with collaborators. Here at RD, some of the ones we carry out are the eNPS, the assessment of the 1-1s with the leadership , the managers’ satisfaction with the hirings, among others.

In them, we seek to work with quantitative data (used to calculate KPIs) and with qualitative data . They complement each other and serve as a basis for analysis and investigation, after all, it’s no use knowing that your KPI is below the target if you don’t have feedback from employees to understand what’s behind the number.

Ron Ashkenas, in the Do You Need All That Data? , cites that “neither quantitative nor qualitative data tell the whole story. For example, to make good decisions about products and pricing, we need to know not only what is being sold to whom, but also why some products are selling more than others.”

Use and abuse categories in qualitative data

It is often difficult to work with qualitative data because of the subjectivity behind it. However, a good practice is to categorize these data, so that it is possible to verify how often the themes are repeated.

Here at RD, as we work with several different sources (eNPS, 1-1s rating, GPTW, checkout interviews etc.), we have created unified categories that are used in all surveys, thus enabling a better understanding and comparison between these sources.

David Harcourt, in the same webinar mentioned above, suggests that we work with this “unstructured” data as it is an excellent source of insights and often reveals hitherto unknown issues. He proposes several techniques for text analysis, such as:

  • Count and word cloud ;
  • Cluster analysis;
  • Feelings analysis;
  • Extraction of themes and topics.

These techniques are essential when there is a high volume of comments and feedback from a survey, but not enough time to read each one.

Communicate, Communicate, Communicate!

I dare say that this is one of the main lessons that we have here in the TM team, communicating what we are doing, whether for the management or for all RDoers. At RD, we value transparency regarding business metrics and strategies. This is an important approach if the company wants to build a collaborative environment that is open to change.

What happened (and sometimes still happens, because nobody is perfect) was that we stopped presenting these numbers and, mainly, the initiatives we took to improve the work environment. Of course, it is necessary to know what to communicate, when and how to do it, so that people have access to the information they need, but that the confidentiality of responses is not breached.

A good practice we adopted was to communicate the overall results on corporate TV so that everyone knows that yes, answering that 5-minute form is important and we are looking carefully at the results. Another practice adopted was the creation of dashboards for leadership, showing the results by area, thus enabling the construction of actionable according to the reality of each one.


Working with metrics and data still scares a lot of talent management teams out there. But this is an increasingly common practice in companies and is the result of this revolution that is still taking shape in the Brazilian market.

As Maia Josebachvili mentions in her article, successful companies “will be those that are reinventing their HRs to become more deliberate, strategic and holistic People Management teams”. Human Resources, People Management, Talent Management (or any other name chosen) professionals need to readjust to this new management model if they want to keep up with this market movement.

To make this area more strategic, intelligently analyzing metrics and indicators is essential. But don’t panic, you’re going to make a lot of mistakes along the way. The important thing is to learn from each one of them and move forward.

It is not difficult to start this movement, but persistence, patience and consistency are required to continue working with data in people management, because unlike what happens in other areas, the ROI  is never so clear and the results do not come in the short term. However, data-based decision making is much more assertive and effective.

Do you want to delve into the topic? I highly recommend the post that talks about working with Business Intelligence in recruitment and selection. And you, do you have any learning to share with us about how to apply metrics in the People Management area? We would be happy to read your comments!


People Management techniques are increasingly based on metrics and data. See some experiences that RD had working like this! HR has been reinventing itself, and many companies and professionals need to adjust to this new market context. In the same way that new technologies are making areas such as Marketing, Sales  and Customer Success  more strategic and intelligent , areas known for being…