01/13/2021
Look around an organization in 2021 and you likely see some examples of the shimmering panoply of 21st century technology. You may see more flexible, agile teams, and people working seamlessly across time zones and continents. There is still hierarchy, but it’s less suffocating than in the past.
But there is one obvious legacy of the last century that often remains untouched and out-of-date: the performance review process. It has been written about, researched, and debated for years — but it’s still broken and has become a political and bureaucratic battlefield. Part of the reason it doesn’t work, in our view, is that all the power still rests with the boss. Additionally, the end-of-year performance review often becomes a negotiation around compensation and promotion rather than a true development discussion.
The future of the performance review is especially relevant as companies adopt new ways of working. Companies from Spotify and Amazon to ING Bank and Haier are using agile, self-organizing team-based structures. In these models, teams and small organizational units have both autonomy and accountability for results. There is no formal boss, so the traditional boss-employee performance management system is no longer applicable.
Whether your company is considering a team-based structure or not, we believe there are ways to reinvent performance management and make it more valuable and less boss-centric. The opportunity to create a socially-based feedback system feels even more urgent during the COVID-19 crisis, since many people are working remotely and without the same level of daily interactions with managers.
Here’s how such a system generally works. Instead of a performance review created and discussed solely with the boss once a year, the employee receives a larger amount of feedback (often 50 or more instances over the course of a year) from colleagues. This feedback captures the view of their performance and behavioral attitude from the people the employee has been working with on a regular basis. Because the system generates a larger amount of feedback, each individual is less subject to the potentially biased view of a single person: the boss. Of course, the boss doesn’t disappear, they are still the person aggregating the feedback and summarizing it in a review that’s part of the performance evaluation process. But in this model, the boss is not the only person providing that feedback to the employee – colleagues, peers, and even other managers offer timely input.
To implement such a model, managers need to consider six key questions:
The logical place to start is with other members of the team. In agile practices, teams meet to review their work in retrospective meetings and offer feedback to each other. In those settings, the team, rather than the individual, is generally the focus, but it is an important template for how performance analysis and development can be shared across more people.
Another source is an employee’s individual network in the company, especially those they collaborate with frequently. Google has a simple peer selection practice for this purpose. Googlers suggest a list of colleagues to their manager and their manager has to then approve that list after considering whether the selected people are close enough to the employee and their work to provide substantive feedback. Googlers select both peers, senior managers, and more junior colleagues in order to capture several perspectives. All feedback has the same weight. That information is analyzed and aggregated by the manager and discussed in regular check -in meetings as well as in a performance summary and discussed with the employee at the end of the year for career development and compensation decisions.
Feedback should generally be open, direct, and transparent – or at least that is the usual aspiration. But what if the culture of the company is not mature enough to sustain an open system when feedback can come from any peer or collaborator? In such cultures, people fear creating conflict or damaging relationships by delivering negative feedback. If this is true, such a system won’t be effective or help people develop.
We can again turn to Google for an example of how this can work — in its peer-performance feedback system, it shows its employees anonymous peer review content. Only managers have access to see the reviewers.
Netflix encourages employees to give blunt feedback and use “radical candor.” When Netflix stopped doing formal, annual performance reviews, they instituted peer reviews. They kept them fairly simple: People were asked to identify things that colleagues should stop, start, or continue. In the beginning Netflix used an anonymous software system, but over time shifted to direct feedback, and employees held their 360s face-to-face with guidelines on how to use candor and be frank.
There is no one-size-fits-all answer on anonymity. If the organizational culture is not ready for a free and frank exchange of feedback among colleagues, it is probably better to go for an anonymous system. The hope is to avoid what we have seen in multiple companies embracing 360° feedback over the last 20 years: a bell curve skewed toward the highest rate. With such skewed bell curves, the results are worthless.
Timing is important when delivering feedback. But there are some rules of thumb when moving away from the boss-centric performance review process.
First, consider cadence. Requests for feedback should come at a specific cadence tailored to how an organization works – that could be a Sprint (in agile) or by quarter.
Then consider volume. Moving to a social feedback system requires managing a large volume of feedback on a single person; but that feedback needs to come from a wide range of colleagues and not from a focused, inner circle. A prompted system puts guardrails in place to ensure that managers get large and similar volumes of feedback for each employee.
Praxis Precision Medicines, a Boston based biotech company, issues requests for feedback at the end of each month with a target of around 100 pieces of feedback for a single person in a year. This is aided by Microsoft Office Analytics, which identifies the people with whom the employee interacted most . The system sends automatic requests for feedback to 10 of those collaborators.
In a team-based organization where teams are self-administering and hierarchy is flat, logically all feedback would be equal. It wouldn’t matter if the feedback came from the line manager, other team members, the Product Owner, or the company’s CEO. With equal weight and volume of feedback (achieved with a prompted system), this approach is trying to move to the “wisdom of the crowd,” where independent pieces of input are statistically combined to form an accurate, aggregate picture of performance. In other words, if 100 people provide a consistent negative feedback on one aspect of performance or behavior, such feedback will most likely reflect the sentiment of the overall company. This reflects a belief that the “wisdom of the crowd” is more accurate than the judgement of an individual, as demonstrated by several psychological and social studies.
ENEL, a global utility company, has adopted a social feedback system through a proprietary, internal platform. Employees can request feedback from anyone with whom they have shared a job activity at any time. Peers can freely decide whether to evaluate the performance of their colleagues. All feedback has equal weight and is aggregated in monthly, quarterly and annual reports which are visible to the employee and managers at any point in time.
The feedback categories don’t actually require a significant departure from traditional models. Usually some are on performance, while others focus on behaviors that reflect the company’s aspired-to goals, culture, and values.
At Praxis Precision Medicine, they use three categories for performance (the “What”) and four categories for values and behaviors (the “How”). Feedback providers don’t need to offer input on all categories, only on the ones observed directly during the period.
Since truly social feedback generates a large number of pieces of feedback on any single team member, the feedback itself needs to be simple and quick. It should take no more than a few minutes. In some companies, every single team member is requested to provide 3-4 pieces of feedback every few weeks, at the end of a project Sprint.
Based on our experience with traditional performance management systems, to prepare and deliver a single review takes 2-3 hours on average. This is not feasible in a social feedback system targeting hundred(s) of pieces of feedback per person and asking each person to provide the same number. So, the simplicity of the feedback is more in line with how we think about social media — a quick like/dislike, binomial feedback on each relevant and observed category.
The employees of ENEL use their internal platform or app to provide such feedback. A user can click on each of the feedback categories they have observed and then has two choices on a simple binomial scale: appreciation or opportunity for improvement. Giving feedback takes on average no more than 5 minutes.
Load older comments...
Loading comments...
You've Been Timed Out
Please login to continue