The impact of performance variation on service delivery (…and how to reduce it)
Increasingly over the last few years macroeconomic conditions have put pressure on Government budgets and profit margins of private firms.
Key to survival and success for organisations, more than ever before, is effective performance management, improved productivity and delivering superior value to customers.
An essential foundation of success is an understanding of performance variation in organisations and an active programme of activity to control it.
Ford Motor Co. saved $886m over 4 years as a result of an integrated performance variation programme (Kwiecien and Wolford, 2001).
This article is not arguing for enforcing absolute standardisation and formal unbreakable rules across a service — which may in themselves reduce variation but often at the cost of low innovation, poor engagement and ultimately higher failure demand and higher costs.
This article instead argues and demonstrates that an understanding of service performance variation, can boost overall performance, reduce costs and increase customer satisfaction.
Managing processes and performance by relying on the average level of performance (“the mean”), without understanding the variance in performance, results in;
· Sub-optimal quality
· Increased costs and inefficient service delivery
· Poorer customer experience and retention
· Missed opportunities for learning and sharing of best practice from the “star performers”
Despite this, most performance management at an individual, team, group and unit level focuses on mean level of performance.
Sometimes reporting of the range will occur, however this measure of dispersion is sensitive to extreme values and does not tell the overall story of performance.
Questions for leaders:
· Do you know how performance of your teams is dispersed from the average?
· Can you truly justify why teams with similar systems, processes and work perform differently?
· Have you calculated the impact or performance gains from addressing this issue?
‘Unforced’ or non-value variation can account for up to 10% of service delivery costs — this is cost that serves no value to the organisation or the customer.
Therefore it is crucial that organisations understand and pro-actively manage variation in service delivery — and this doesn’t mean performance managing people via “forced distribution” and focusing on the top/bottom 10%.
Team leaders and Unit managers may intuitively look at variation, but this practice in itself is subject to variation, is inconsistent and not undertaken systematically.
Whilst not empirically perfect the normal distribution or bell curve is a good way to visualise the impact of variation.
Most people are familiar with the normal distribution, and it is used quite extensively because it occurs naturally in many situations in business (e.g. selection tests) or in the real world (e.g. height, weight) and its shape is perfectly described by two parameters (mean and standard deviation).
The impact of variation is illustrated in the attached by the red, blue and green shapes — all are normal distributions, and all have the same average (µ=0) (DN: the green is the “standard normal”). Yet the mean hides the fact that some customers are getting “minus 5” service or superstar performers are delivering “+4/5 service”.
You could at this point argue that, what is the problem, if the average is the same, the overall cost is the same?
From a first level financial position yes:
(10 x -5) + (10 x 5) = 0
(20 x 0) = 0
However depending on a specific process, think about the impact to the first 10 customers receiving “-5 service”?
Frustration, dissatisfaction and probably a requirement for re-work all leading indicators of additional costs.
Alternatively what is the impact of “+5 service” managed by the mean only?
Well if a small number of people are delivering this, they may not be recognised for their “hyper performance” leading to under recognition and future greater turnover of these high performers.
The team will also be likely to suffer from the free rider effect — team members taking it easy because a small group are “carrying the load”, this is a particular risk of team based targets and rewards.
There is also a negative cost associated with delivering at “+5” if there wasn’t the need to deliver at that level, you could reduce the investment in each unit of service to reduce the overall cost to serve (for lean aficionado’s this is “over-processing”).
The call-centre for Working Active (WA) a specialist health care insurance provider, employs 20 customer agents who are responsible for processing insurance claims.
Sarah the manager of the call-centre has a KPI of 100 claims per day the level of customer demand is also 100, based on a work-study she has been given enough supply based on 5 claims per day (100/5 = 20FTE), luckily for Sarah her 20 staff are on average completing 100 claims per day.
However, now due to budget constraints she needs to make efficiencies, so she decides to look at the distribution of performance across WA:
What this tells Sarah is that 10 members of staff have over a sustained period of time averaged less than 5 claims per day, luckily she has 6 members of staff who are delivering either 7 or 10 claims per day.
Sarah knows that she has a couple of people who have recently been trained, but she is unsure why 8 other members of staff are well below the average.
Sarah calculates that if out of the 10 members of staff that are currently delivering less than 5 claims per day, 8 increase their performance to 5 claims per day and the other 2 continue to deliver 1 claim (due to their training schedule), she can increase production to 113 claims per day.
Easy to say, right?
Alternatively, Sarah can reduce her level of FTE from 20 to approximately 17.5 — almost £70,000 per year whilst maintaining service levels at 100 claims per day.
This is the “size of the prize” of identifying and addressing performance variation.
So ok, if you accept that, what can be done?
Performer analysis –
The purpose of this initiative is to identify the top 20% of performers (individuals, teams and sites) across multiple metrics, care needs to be taken to identify the performers who are there through skill and application, not circumstance (e.g. right place, right time) and perform comparable functions.
Using this population of high performers, you should identify:
· Shared characteristics (% of new hires, level and type of training, quality of line management) — make sure that you can isolate the variables that truly have an impact and not just make judgements based on heuristics (e.g. tenure is not always the best indicator of performance)
· Local processes, to what extent are high performers compliant with organisation processes (what could change at an organisation level as a result)?
What good practice have they adopted that enables superior performance?
How can that performance be codified and replicated in other sites (e.g. via checklists, training, coaching and shadowing)?
This process can and should be replicated with the bottom 20% of performers, and arguably a potentially more fruitful group is those that consistently perform at an “average” and/or around the “baseline” — this will provide a useful way of comparing performance and the drivers of variation (e.g. adoption of a process).
Prioritise areas of attention –
Organisations have limited change capacity therefore focus should be on variation across the key metrics for the firm, for an underwriter it could be number of decisions per day, level of complaints or for an airline dispatch handlers % of on time flights.
Another way to identify the areas is focusing on:
1. Highest volume tasks
2. Highest (unit) cost tasks
3. Tasks with strategic impact (e.g. with legal, environmental or political risks)
Build, replicate and share best practice –
Does your organisation have a culture that incentivises individuals to hoard sources of excellence or to share and be a “corporate leader”?
Most organisations are in the camp of the former. Managers are often competing against their peers and having a source of competitive advantage is something worth holding on to for their own individual gain.
Questions to consider;
· How easy is it for best practice to spread across the organisation ( known as “knowledge diffusion”)?
· How easy is it for colleagues to find and follow the ‘right practice’ (“how to do something”)?
· How quickly does a good, new idea spread?
People Capability Analysis –
Good service delivery, needs good people.
A frequent symptom of performance variation, is a variation in the skills, knowledge and experience of staff. An effective way to boost overall performance and reduce variation, is to complete an analysis of the capabilities that are required for successful performance of tasks, using these indicators you can assess the current capability of staff and develop a plan to boost the key capability areas. This can be time consuming and resource intensive, however the payoff is invariably worth it (in terms of overall performance, productivity and staff engagement), to maximise the value of this exercise you can prioritise the roles that you focus on using the below matrix.
Effective Performance Management –
Where individuals/ teams or sites have underperformed against their peers for a sustained period of time, with limited differences in context (economic conditions, staff experience, trade union relations) then focused performance management should be considered.
· Are there truly mitigating factors that result in a location being in the bottom quartile over a sustained period?
· Are local managers confident and capable enough to have difficult conversations (given that the level of performance has been ‘accepted’ for a long period of time)?
Benchmarking across comparator organisations (e.g. by product or market) –
If your best performers are the average in your market internal variation analysis won’t be sufficient. However, understanding the performance of other organisations is a useful way to improve your own outcomes.
In a Government context this means looking across other Government departments and with foreign Governments who undertake similar work and sharing best practice.
In a private sector context this means reviewing the economic performance of your peers in your own sector and across comparable sectors.
Variation from ignorance or variation from ineptitude? –
As complexity of services increases over time, depending solely on the capabilities and recall of individuals leads to greater variation in performance (Gawande, 2010).
As service practitioners are required to deal with ever increasing complex systems, more diverse customers and a wider product range, the potential for two types of error increase:
Errors of ignorance — mistakes we make because we don’t know enough
Errors of ineptitude — mistakes we make because we don’t make proper use of what we know.
· What steps have you taken to reduce the complexity of a service for your practitioners (checklists, job design, effective Knowledge Management etc.)?
How can you improve and or standardise the diagnosis of customer needs by practitioners (what problems does the customer have and how can the organisation solve them) — to prevent nugatory work?
This article doesn’t provide a comprehensive view of variation and the tools and techniques to resolve it, however hopefully it helps you to understand its impact and the value of taking an active interest in variation.