Performance Measures: Are They More Trouble Than They Are Worth?
Some business practices are, apparently, so obvious and above reproach that they become entrenched and taken for granted. This appears to be the case with performance measurement. The adage that 'what gets measured gets done' is repeated so often that it practically has the status of holy writ.
I have been as guilty as anyone of promoting the use of performance measures in relation to both organisational and individual performance. However, with increasing evidence of the perverse and unintended consequences of performance measures I have also had a growing sense of unease. I have, therefore, been reassured to find that others - experts in different fields - have been experiencing similar and even broader ranging concerns. To give readers a cross-section of views on this important topic I want to share the perspectives of two of these commentators. One takes a macro (societal) view, the other a more micro (intra-organisational) view. Both point to reasons why we should take a hard look at how we use performance measures.
The perverse and unintended consequences of performance measurement
In her 2002 BBC Reith lectures, Baroness Onora O'Neill, a political philosopher and ethicist, was critical of what she referred to as the modern and fashionable quest for 'accountability' and its associated panoply of performance measures. (1)
Focusing on the provision of public services (police, hospitals, universities, schools, etc) O'Neill highlighted numerous problems that have arisen from an overzealous application of performance measures. She was able to identify many instances in which the result of applying those measures was the opposite of what was intended.
The following issues are extracted from her lectures to illustrate the depth of the performance management malaise about which O'Neill was concerned.
- In the pursuit of accountability the vehicle has been more detailed (and at times prescriptive) regulation and control. Modes of public accountability have resulted that are internally incoherent.
- Underlying the ostensible aim of accountability to the public is the real requirement for accountability to regulators, government departments, funders and legal standards. "If we want a culture of public service, professionals and public servants must in the end be free to serve the public rather than their paymasters."
- Targets are set that must be achieved by applying processes that are inconsistent with the achievement of those targets. This cuts across desirable behaviour by inviting compromise and evasion.
- By imposing the wrong sorts of accountability professional practice becomes defensive: defensive medicine, defensive teaching, defensive policing, etc.
- Performance indicators are chosen for ease of data collection rather than because they accurately and validly measure the quality (impact) of performance.
- Managerially imposed targets create a bureaucratic process burden on those who have to comply. The costs of compliance (and of playing the 'league tables' game') are substantial and not just financial.
- Performance indicators are not only applied with a precision that is illusory but frequently create perverse incentives. (2) Those serving the public are rewarded for their diligence in responding to the incentives imposed on them rather than for pursuing the essence of what is required to be, for example, good doctors and policemen.
- Professionals are required to pursue performance targets that compromise their professional expertise and judgment about what is in the best interests of their patients or students. When incentives encourage arbitrary and unprofessional choices these can distort the proper aims of professional practice and damage professional pride and integrity. Related consequences are a growing cynicism and low morale.
O'Neill describes a situation in which the quest for greater accountability has resulted in the micro-performance management of complex institutions from the centre. This has, in O'Neill's estimation, resulted in over-complex and inadequate rather than good and effective governance. Consequently, public institutions are apparently less trusted and arguably less effective.
Myths on which performance measurement is built
Our second commentator is David Parmenter the self-styled 'King of the KPIs'. (3) An accountant by profession, Parmenter is a writer, facilitator and international presenter on the development of key performance indicators (KPIs). Given that Parmenter has worked with performance measures for over 20 years it is surely of significance that he has just posed the question 'should we abandon performance measures?' (4) While remaining committed to 'the higher purpose' of performance measures he observes that all too often performance measures cost organisations dearly. This is because, for example:
- they are 'gamed' so that executives can increase their pay;
- teams are encouraged to perform tasks that are contrary to their organisation's strategic direction; and
- valuable management and staff time is locked up in costly measurement and reporting regimes.
To get performance measures to work Parmenter says we need to challenge the myths on which they have been built. These are four of the six myths his article refers to.
- Most measures lead to better performance. According to Parmenter there is a dark side to every performance measure that is likely to provoke an unintended action and result in inferior performance. He estimates that more than half the performance measures in an organisation are likely to encourage unintended negative behaviour. As an example he refers to an Australian city rail service that penalised train drivers for late trains. Responding to this 'incentive' drivers began skipping stations (i.e. failing to pick up passengers) to achieve on-time performance. In a private sector example a fast food manager was striving to achieve an award for zero wastage of chicken. He required that chicken not be cooked until it was ordered. The resultant long waiting times caused a significant loss of customers in the following weeks.
- All measures can work successfully in any organisation, at any time. In fact, for measures to deliver their potential there needs to be a positive 'context of measurement'. To help identify the extent of risk in organisations Parmenter offers a Dysfunctional Performance Measures Checklist. A number of these relate to the extent to which the organisational reward structure (both formal and informal) is tied to key performance indicators.
- By tying KPIs to pay, you will increase performance. Parmenter refers to the common belief that the primary driver for staff is money and thus one needs incentives in order to get great performance. He acknowledges this may be the case for employees sitting on the lower rungs of Maslow's hierarchy of needs but it does not apply to management and staff further up that hierarchy. Recognition, respect, and self-actualisation are far more important drivers once basic needs are met. "When KPIs are linked to pay, they create key political indicators, which will be manipulated to enhance the probability of a larger bonus."
- You need performance measures in order to drive performance. "Performance measures don't drive performance - they support and enforce the positive environment that already exists." Parmenter concludes that if an organisation has dysfunctional performance measures it would function much better without them.
O'Neill's 2002 Reith lectures describe a well intended but failing system of performance management of public institutions. In the quest for increased accountability greater control from the centre and the imposition of an expensive system relying on extrinsic (externally imposed and often coercive) motivation has reduced accountability. This has happened because of the harm done to trust and intrinsic (internalised) motivation. In the 10 years or so since those lectures were delivered it is arguable that this problem has become even worse and not just in the UK.
While Parmenter's review is more directed to the organizationally level his conclusions are not inconsistent with O'Neill's concerns. He drives home the point that performance measures should be servants not an end in themselves. They will be useful to the extent that an organisation already has a positive climate for their use and that other forms of performance evaluation are also applied (a point also emphasized by O'Neill).
These are important messages for boards charged with oversight of organisational and executive performance. Superficially, the notion of systematically setting, measuring and rewarding performance is unimpeachable. History suggests, however, that we have fallen well short in our ability to do this successfully. Performance measures are difficult to design and apply validly and coherently to complex organisations and those who work in them. As Parmenter suggests, "the greatest danger of performance management is dysfunctional behaviour". In this regard the greatest risk may lie in linking performance measures to remuneration. Such links undoubtedly create incentives for gaming that can ultimately lead to organisational meltdown (think Enron among many others!).
(1) Onora O'Neill (2002). A Question of Trust. The BBC Reith Lectures 2002. Cambridge University Press.
(2) A classic example of this was the response of one British hospital to the demand to reduce emergency department waiting times. Nurses prevented patients from being admitted and triaged until doctors could see them.
(4) David Parmenter (2013). Should We Abandon Performance Measures? Cutter IT Journal, Vol. 26, No. 1