fbpx

Calibration in Performance Evaluation

Renan Araújo

tecnicas-de-marketing

What is it?

Calibration is the penultimate phase of a performance evaluation.

In it, managers (typically from a department or set of related departments) meet to discuss the performance of their subordinates and agree on the performance evaluations carried out.

What is it used for?

Calibration serves precisely to calibrate the hands of managers. Managers may have biases or simply sets of experiences and backgrounds that make them more or less “heavy-handed” in their evaluations. Calibration meetings serve precisely so that managers, under the tutelage of the HR area, arrive at common evaluation weights and measures among themselves.

Calibration of performance appraisals serves to increase the accuracy of the assessments made by managers. Making sure that these appraisals are fair and accurate is critically important. The calibration process helps ensure that all employees have been assessed against the same criteria.

The calibration (or calibration) process ensures that:

  • The evaluation of the employee’s past performance is carried out in a fair and objective way in relation to other employees of similar function/position/department;
  • Managers apply similar standards and criteria for all employees;

The collective discussion among managers allows them to gain new insights into the performance of their subordinates and to reduce potential biases they may have. The peer discussion (the manager with other managers) brings transparency to the process, drawing attention to the tendencies of this or that manager to evaluate their employees in a too lenient or too critical way. Managers become accountable to each other for the evaluations of all employees discussed.

How does it work?

During a calibration meeting, managers discuss with facts, examples and data their visions in relation to those they lead. Typically, some discussions take place in relative terms. A more heavy-handed manager (Manager A) may have rated her subordinate

(subordinate A) as “Meets Expectations” when another manager (Manager B) rated her subordinate (subordinate B) as “Above Expectations”. However, Manager A understands that Team Leader A outperforms Team B, and a discussion based on examples, results and behavior follows until an agreement is reached.

Generally, we try to gather in the same meeting managers from related areas. For example, in a Qulture.Rocks client company, the meetings were separated in HR and Finance, Support, Technology and Strategy, and finally Product and Design.

After a “slice” is made by area, a slice is made by seniority of the employees to be discussed. In a company that has simple positions, the managers of the HR and Finance areas can begin by discussing the evaluations of supervisors and analysts. Subsequently, the directors discuss the manager’s evaluations, and so on. The idea is that the employees who will be actively discussed by their managers always leave the room.

In performance appraisal software like Qulture.Rocks , HR can leave calibration groups pre-configured so that discussions can be started from a click. We call these “calibration segments”.

In the hypothetical example below, we would have one calibration segment with all the supervisors in the company, who would be calibrated by the managers, and another committee with the four managers in the company, who would be calibrated by the two directors:

Checklist for preparing a calibration meeting

Here are some critical points you should ensure if you are organising a performance evaluation:

  1. Schedule meetings well in advance and publish a schedule with key dates for all managers;
  2. Communicate and train managers on what is expected of them in terms of participation and prior preparation;
  3. Maximise the personal presence of managers at meetings. Remote participation (video or call) is far worse and less productive;
  4. Ensure that all managers have already carried out their appraisals in the performance appraisal system so that these can be discussed;
  5. If an optimal distribution of appraisals (such as a 20/70/10 curve) is used for northern purposes, it should be communicated to managers at the beginning of the appraisal process;

The HR area should have data and tools to help conduct the meetings. Some suggestions are averages of the evaluations given by each manager, which may accentuate discrepancies in rules/criteria (for example, on average Manager A evaluated her subordinates worse than Manager B, in addition to averages per committee being held (for example, an average of the managers of the HR + Financial areas in relation to the other managers of the company).

The effect of cognitive biases on calibration

At Google, all participants in a calibration meeting are given a sheet of paper listing and explaining the key cognitive biases that can affect our judgement in performance

appraisals. This help is taken by Laszlo Bock, the company’s former VP of People Operations, as key to increasing the objectivity of calibration discussions.

Cognitive biases are shortcuts that our brains take to make quick judgments about our environment. They were very useful when we were semi-wild beings and needed to defend ourselves quickly from the threats of nature (ever noticed how quickly our body reacts to a scare? How adrenaline, which serves to prepare us for an eventual escape, is discharged in a millisecond? Well, yes.) However, these reactions are no longer useful nowadays because they get in the way of our rational judgements. It is therefore essential to know about them and be aware of their effects.

Some of the main cognitive biases are:

Recency bias

  • Mistake: we tend to remember only what is fresh in our memory.
  • Antidote: force yourself to think about evidence from the whole period.

Halo” or “horns” effect

  • Mistake: we tend to ignore evidence to the contrary when we have our memory scarred by good or bad things the person has done.
  • Antidote: wonder if the evidence does not point to a trend reversal.

Fundamental attribution error

  • Mistake: we tend to attribute results (or the lack of results) to a person’s competence or effort, when often those results were determined by exogenous issues.
  • Antidote: question whether the results were actually produced by the person (if there is a cause and effect relationship).

Confirmation bias

  • Mistake: we seek to filter and “find” evidence that confirms our prejudices (hearing only what we want).
  • Antidote: ask yourself: what is my bias here? What do I want to confirm unconsciously?

In the video below, which has been used by some of our clients as the opening of calibration meetings, see an introduction on key biases, much like the one distributed on Google:

What is performance calibration not?

Calibration of performance appraisals – what we have discussed so far – should not be confused with a succession planning meeting, which is called People Chess. In fact, a big risk of the calibration process is that there is no clarity of what is – and what is not – being discussed there.

The purpose of the people’s chess meeting is to discuss who will be promoted to which position, to define who is more or less ready to assume more responsibilities, who is the natural successor of position A or B, and so on. In these succession planning meetings, much more complete aspects than just performance evaluations must be taken into account, such as the collaborator’s potential, his willingness to relocate, his willingness to take on new challenges, the existence of a successor for him, among other factors.

Performance calibration, on the other hand, only discusses the criteria being evaluated in the assessment, which are usually some variation of results, on one axis, and behaviours, on another axis.

If you want to receive more content about new HR practices and performance management, subscribe to our Newsletter and stay on top of new blog posts!