One of the more surprising details of consulting firm KPMG’s recent woes was that it had been using a ‘set distribution’ or ranking approach to assessing employees’ performance. Professional services companies are increasingly embracing more flexible ways to manage how employees are faring, but why?
As employer branding exercises go, February was not a great month for consulting company KPMG. Chairman Bill Michael announced he would step aside after telling employees to “stop moaning” about working conditions during lockdown in a virtual staff meeting, and later telling staff that he thought unconscious bias training was “complete and utter crap” and did not improve diversity and inclusion.
But amid the furore around Michael’s comments and the reaction of staff – who told him to ‘check his privilege’ and accused his ignorance of employees’ mental health issues of being ‘heart-breaking’ – was the revelation that the company still used a ‘set distribution’ model of ranking employees’ performance.
This model, also known as ‘forced ranking’, ‘rank and yank’ or a ‘vitality curve’, is one that was famously championed by General Electric chief executive Jack Welch. In it, employees are plotted across a distribution curve of high performers, mid-ranking performers and finally a tranche of under-performers. The premise is that there is a ‘normal’ curve of distribution where there is a small number of under-performers, lots of people in the middle, and a small number of star performers who are commensurately rewarded. Employees in the bottom 10% are placed on performance improvement plans or in the worst cases dismissed for not meeting standards.
The thinking is that by forcing employees into these rankings, they will compete against each other and so everyone’s performance will improve. Despite its ruthless nature, it had until recently been a relatively common approach to managing performance. In a competitive area such as law, where there are obvious metrics such as billable hours, such an approach could factor into both promotion and bonus decisions.
Microsoft ditched it in 2013, saying it wanted to focus more on “teamwork and collaboration”, while Accenture dropped it in 2016, along with annual performance appraisals. Amazon has also used it in the past, but it has definitely lost favour in recent years as organisations have sought more inclusive and collaborative cultures.
For its part, KPMG says it is working hard to move away from this model. A spokesperson said: “Our performance management framework was introduced in 2018 and is designed to create a continuous feedback culture, build meaningful relationships for performance managers and colleagues and support high performance. A set distribution is a commonly used approach designed to support a high quality, high performance culture. Each year we aim to evolve our approach to performance development and we have decided to move away from a fixed distribution model and allow much greater flexibility.”
So what’s wrong with plotting people along a performance curve? One of the chief issues is bias: if one individual manager is tasked with ranking their team from high to low, there will likely be implicit biases for those that they get on with or whose style of working is one they favour. “It’s hard to know whether that data is genuinely reliable because it’s subjective,” says Roly Walter, founder of Appraisd, a performance management system. “You can try to remove that subjectivity by introducing sales targets or things that are objectively measurable, but if you’re measuring things closely associated with someone’s personality, it becomes tricky.”
The forced nature of the distribution curve is also problematic (and also assumes all organisations have a normal, bell-shaped curve). Someone who happens to fall into the bottom 10% according to that team’s chosen metrics is not necessarily a poor performer, their colleagues have simply scored higher.
“It’s arguably the polar opposite of where performance management has been going for the past three or four years,” says Ian Lee-Emery, CEO of talent management technology company Head Light. “You’re trying to overlay a mathematical model onto people’s performance. That then drives bonus allocations, which is not without its issues as people feel it in their pay packet. It can be a real downer in terms of employee motivation.”
In the legal sector, he acknowledges that if lawyers are dropping billing days or are not producing the right level of work for their day rate, this can be a problem. But that does not necessarily mean this approach works well in this environment. “The world doesn’t fit into a bell curve, so if we assume that managers have high performing teams, you’d have to skew the curve to the left as there would be a larger proportion of high performers.”
“No-one wants to be branded an under-performer, so you get a halo effect at one end and a ‘non-halo’ at the other. Is performing least well not good enough, and if so what do you do about it? If it’s not perceived as fair and transparent there is a potential risk of tribunal.”
Indeed, the impact on employee engagement of this type of approach all comes down to perceptions of fairness. It doesn’t matter what your favoured method of measuring employees’ performance is as long as they consider it to be transparent, fair and unbiased.
“In life this sort of judgement happens all the time – the top people go here, bottom people there. But rarely is it described in such a bare and harsh way,” says Elva Ainsworth, a transformation consultant. “The impact will depend on trust in and the fairness of the process. If people see that it’s done on objective measures, that they’ve had a chance to have their say, there’s a chance it will succeed. Any hint of ‘only the boss can have a say’ or lack of trust in the system and it’s a disaster.”
This is especially true at a time when most of the workforce is likely to be working from home, and anxiety is heightened anyway due to the ongoing pandemic. She adds: “It gets in the way of truly open conversations around performance and emotional safety, which is what we’re trying to work towards.” The shift to home working and changing business priorities has arguably moved the goalposts as to what should be measured, while the high-fliers of pre-Covid days may struggle to excel in the ‘new normal’.
Teresa Boughey, founder of HR consultancy Jungle HR, says: “Covid has really broken so many traditions, not just ways of working but the importance of being viewed as a whole person and what you can bring to an organisation. Those employees that were successful in an office environment may have dropped [in terms of rankings], while those who are more self-reliant and operate well in a different environment have thrived. Any organisation sticking to a forced ranking measurement would not be able to track that change.”
She adds that this type of performance management framework tends to rate one or two things highly (such as new business or billable hours), so can be one-dimensional. “It creates this idea that we have an ideal, high-performing employee and that everyone should be modelled on that path,” she explains.
“But as a leader you can’t be good at everything, and if you put everyone through a process like this you’re churning out clones. It means people will be reluctant to take ownership of other areas they need to develop.” Over-achieving in one area doesn’t necessarily translate to business success, either – someone could be putting in all the hours but going against the company’s values or disengaging clients in the process, which could lead to problems in the longer term.
‘Arbitrary and unfair’
A review of approaches to performance management by the CIPD in 2016 said there was “good evidence” to ditch forced or guided distributions, saying “employees typically view the practice as arbitrary, not a reflection of actual performance and unfair. Even when applied with perfect consistency, its research found, employees see this type of approach as “showing a lack of interest in and a lack of appreciation or value for the work they’ve done”. In areas where performance can be more uniformly measured, it adds, the differentiation between employees is more likely to be artificial and so perceived as unfair.
Jonny Gifford, senior adviser for organisational behaviour at the CIPD, argues that companies need to clarify what it is they’re trying to achieve with the approach they have chosen. “Are you trying to inform administrative decisions about pay, to decide whether to promote someone or renew their contract? Or are you trying to help people improve in their roles? These are two different aims,” he says. When supporting employees to develop their skills or overcome issues on a day-to-day basis, strengths-based performance conversations work best, but discussions around reward should arguably be kept separate. “Monitoring how people are doing helps them learn and stay focused on their goals so periodic discussions are really important,” he adds. “But then you have a conversation about pay at the end of the year. You can say ‘this is the reward available to our team and this is what you’ll get’ without labelling it as ‘above average’ or ‘below average’.”
Walter from Appraisd agrees: “You can split it out into component parts, you don’t have to do all these things at the same time. You can have an appraisal process that happens once a year because it’s good to reflect, but managers should also be giving continuous feedback, behaving more like a coach than an assessor – this is not an MOT.” One strategy is to timetable career discussions at one point in the year, wellbeing at another, and discussions about competency and skills at another, he adds.
There are potential legal ramifications for employers that misapply a forced ranking system. According to Esther Langdon, an associate at law firm Vedder Price, strict action resulting from placing someone in the bottom tier could mean dismissal processes are not carried out fairly. “In the US, employers have used this approach in the past as justification to get rid of the bottom layer. Here in the UK, employees have a right not to be unfairly dismissed after two years’ service and if they do it must be for not meeting capability and performance demands,” she explains. A tribunal would want to see if an organisation had followed Acas guidance on asking someone to improve before dismissal, and that due process had been applied. “There’s nothing wrong with robust appraisals, done in the right way. But ranking creates a cut-throat culture,” she adds.
One way to mitigate any potential bias in performance management is to draw data and feedback from multiple sources. Boughey advocates ‘calibration sessions’, where an individual’s performance is discussed with a group of relevant people who see the employee in a different operating environment. “The line manager sometimes has limited access to what that person is actually doing,” she explains, “so you get the views of others and calibrate a view of how they’re doing. Then you’re clear about what the results mean for their development and how they will be supported.” Peer reviews can also help to build a truer picture: what value does this person add to the team? How do clients rate them?
Ainsworth adds: “At these times when our emotional needs are heightened, the emphasis should be on honest conversations, gathering information from relevant sources.” She now sees elements of 360-degree feedback integrated into the appraisal process more and more, with differing levels of detail and modes of giving that feedback. “The aim is to guide a conversation when you may not have seen that member of staff for months, or in some cases perhaps not even met them at all. Performance management falls down where it is ‘done to’ people and they don’t feed in.”
It’s important that – in introducing sources of feedback into the performance management discussion – there isn’t undue emphasis on what one person thinks. Lee-Emery adds: “One of the things wrong with 360-degree feedback in the past is undue emphasis on what the manager thinks. Peer-level feedback is often thought of as an add on and has been underplayed.” Boughey says that “self-calibration” is crucial, too; asking the person how they feel they have represented a client rather than telling them they did a bad job, and helping them to explore what they could have done better.
Gifford believes that, from a governance point of view, HR teams will increasingly be expected to have a strong grasp of performance management approaches that work well for their firm. As firms slowly set aside the distractions created by the pandemic and begin to focus on increasing productivity, there will be renewed questions around how they understand performance, how they measure it, whether they should apply standard scales and who does the ratings. “Performance will be on organisations’ radar more and more – we need to equip boards and HR leaders with the right insights into how performance works, and how they prioritise different approaches at different steps along the way,” he says.
But the focus should always be on supporting, rather than scoring, an employee, Gifford concludes. “The main thing is that we help people see where to access support, to overcome the challenges they face. Above all else the focus should be on improving people’s performance.”