How to design key performance indicators

In the broadest sense, a KPI provides the most important performance information that enables organisations or their stakeholders to understand whether the organisation is on track or not. KPIs serve to reduce the complex nature of organisational performance to a small number of key indicators in order to make it more digestible for us.

This is the same approach we use in our daily lives. For example, when you go to your doctor he might measure blood pressure, cholesterol levels, heart rate and your body mass index as key indicators of your health. With KPIs we are trying to do the same in our organisations.

KPIs should be clearly linked to the strategy, i.e. the things that matter the most. Once you have agreed, defined and mapped your strategic objectives you can design KPIs to track progress and gain relevant insights to help manage and improve performance. KPIs have to provide you with answers to your most important questions.

Only when the senior management team has agreed on the content of their strategy should they progress to the design of metrics. The purpose of metrics and related targets is to monitor progress toward the achievement of strategic goals, and ultimately the delivery of the organisational strategy. Metrics are not an end in themselves. 

To heighten the likelihood that metrics or are both meaningful and relevant, it is useful to articulate the questions you want to have an answer to before you develop any KPIs. So called key performance questions (KPQs) should be developed that capture precisely what managers need to know for the delivery of strategic objectives. 

KPQs steer the KPI designers into asking: “what is the best data and management information that we need to collect in order to help us answer our most important and unanswered questions?” Starting with KPQs ensures that all subsequently designed performance indicators are relevant.

An in-depth explanation of KPQs can be found in the API management white paper: What are Key Performance Questions?

It is also important to stress that strategic measures are different from those required to monitor operational performance. Too many performance frameworks and scorecards confuse the two. 

While with operational measures, it is desirable to get closer and closer to ‘real time’ measurement, this is not required for strategic measurement. Strategic measures are rarely monitored day-by-day, and certainly not hour-by-hour. Strategic measures are more about monitoring progress toward achieving a new and different destination (as opposed to just doing things better), and they don’t change that often. 

Most organisations have a preference for choosing quantitative metrics. This is not surprising as quantitative data is easier to collect and to translate into meaningful metrics. It is generally seen that having numbers can make performance levels easier to understand and numbers generally produce a less subjective result. This means that most companies are more comfortable with analyzing quantitative metrics as this can be done using existing spreadsheet applications.

However, it is important to balance numeric data with qualitative (non-numeric) assessment of performance, as this can be a powerful way to highlight issues that are important to customers and stakeholders.

Consider a quantitative customer satisfaction survey that asks how happy your customers are with the service – using a 5 point scale. The benefit is that you are getting numeric results that you can compare against other departments and month-by-month. The downside is that you don’t know what it is they do or don’t like.

Now consider a qualitative survey asking your customers about feedback, without any numeric answers. The benefit is that you are getting an insight into what they really like or dislike. The downside is that this qualitative feedback is harder to analyse and very difficult to compare. Many companies now make sure they collect both at the same time to get the numbers that are easy to compare and analyse as well as the rich insights from the qualitative data.

We all want objective metrics that give us unambiguous insights into company performance. Many believe that quantitative measures are more objective while qualitative measure are more subjective, and with it less valuable. This is a myth that is important to tackle. There are two important observations to make:

1. Quantitative measures usually have some degree of subjectivity
   As we know many reported financial numbers require judgement and contain estimates. Also fully subjective assessments can be translated into numbers. For example, when we rate our customer experience this is our subjective assessment which we can then translate into a rating or number (e.g. 5-star, 80 out of 100, etc.)
2. Subjective assessments are important and valuable
   What we see as ‘objective’ measures can also often be very narrow and specific measures with limited insights (e.g. how many service visits we have made, how many widgets were produced, how much cash we have in the bank).
   
   Subjective assessments allow us to take into account the experience and subject matter knowledge of the person making the assessment (e.g. a manager making an assessment of the IT support, a mystery shopper scoring overall service experience, etc).
   
   Subjective assessments also allow context to be taken into account. For example the objective sales figures may have exceeded targets and prior years but may still be subjectively assessed as poor because since the targets were set a major competitor ceased trading.

Again, companies with well-crafted KPI frameworks would have a balance of subjective and objective metrics.

KPIs are often classified into leading or lagging indicators. The idea is that lagging indicators are output, outcome or result indicators that show how well you are performing today based on the actions of e.g. the last quarter or the past. Just think of financial performance, which is generally a lagging indicator.

A leading indicator on the other hand is seen as a measure of something that leads to better results, outputs or outcomes in the future. Leading indicators are therefore often seen as the more important measures for the management team of a company. Customer satisfaction is often seen as a leading indicator for future customer loyalty and future financial results.

The complication to this definition is that some KPIs can be both – leading or lagging indicators – depending on the point of view. Think of ‘operational effectiveness’ – this can be a leading indicator for customer satisfaction or a lagging indicator for improved training and staff engagement. The classification of leading vs lagging might not be perfect, but the general idea is important: we need to measure both:

• whether we are achieving the end results today, as well as
• the things we believe will help us achieve those results in the future.

While on the one hand there is a need for measures which provide companies with the answer to their unique business questions on the other hand there is a need to be able to compare information across the organisation or to other organisations. While the former calls for unique KPIs the latter calls for generic organisational or industry benchmarks.

The problem with using generic KPIs is that even though data is available to compare against others, it is not always clear whether the data was calculated in the same way and it can be a challenge to understand the underlying reasons for differences in the numbers.

Here again, companies need to strike a balance. In most industries there will be some industry metrics that are useful to track because they provide useful comparisons and may also be tracked by stakeholders. Many of the results indicators can be measured using generic industry KPIs, while many of the leading indicators are more difficult to measure using generic metrics.

To be useful for aggregation, comparison and best-practice sharing, measures should be commonly defined organisation-wide. Typically this is an early and difficult challenge as it is not unusual to find that performance is measured in many different ways across the enterprise. Using the indicator template discussed below and storing the definitions in a central repository should help create common definitions.

Metrics should be actionable. Measures that are nice to know but do not trigger step-change performance improvement typically have no place in strategic performance management systems.

For instance if an organisation has an objective to retain talent and its well-defined, enterprise-wide metric shows that strategically critical employees are walking out the door, then this should trigger an intervention.

Simply put, we have a strategic objective, the measure indicates we are failing to meet that goal and so we do something about it. This represents the most basic, and oldest, premise of performance management – turning strategy into action.

For more details on how to develop a KPI see also the KPI design template and the KPI decision template. For actual examples of financial and non-financial KPIs browse the KPI library.

Further reading into designing KPIs is available through the eBooks and articles below.

These resources are available to ICAEW members and ACA students.