vintage picture of group around an open book

How to define Key Performance Indicators (KPIs)

Table of Contents
    Add a header to begin generating the table of contents

    What is KPI definition?

    KPI Definition is the process of clearly breaking down the data, calculations and exceptions we need to know to be able to reliably reproduce the production of the metric being defined. Effective KPI definitions also support the auditing of management information processes, training of new staff and the creation of user guides and process cheatsheets.

    Why key performance indicators need to be defined

    KPIs, or Key Performance Indicator, exist to help us measure performance, identify opportunities, to improve and to track progress. They are the central tool in any kinds of performance management.

    For our metrics to be effective you probably already know that it's essential for them to be measurable, for there to be a good mix of leading indicators and lagging indicators, all of which should be carefully linked to our organizational strategic goals (using KPI Trees) and presented using a well designed set of reports and KPI dashboards.

    The hidden problem with most metrics

    The problem is, all that design effort will be for zero benefit if the measure definition is poor, incorrect or even non-existent. Poor key performance indicator definition can lead to terrible personal and professional outcomes.

    There is a simple tool that you can put into action today that will dramatically improve the quality of your KPIs and add a stack of professional credibility to your reputation, that tool is structured KPI definition using the ROKS KPI Canvas toolkit.

    The ROKS Canvas approach will sound very obvious and straight-forward when explained, but based on twenty five years of consulting experience, the author of this article can safely say that almost no organizations that have not implemented the ROKS method will have clear, thorough key performance indicator definitions in place.

    A hidden issue in almost every business is the ‘common sense’ problem. Almost all organizations use descriptive names for their KPIs and measures. This can be very dangerous as the human mind is great at making assumptions based on the name. As soon as we see the name of the KPI we start thinking 'It's obvious, everyone understands that!', even if we do so subconsciously. This can tempt managers to assess performance by comparing dissimilar KPIs based on very similar sounding (or even identical) names, where the underlying calculations are profoundly different.

    A KPI definition horror story

    KPI definition problems example - Utilisation

    Let’s take an example. There is a popular operational measure called ‘utilisation’. Utilisation is a way to measure the proportion of the available employee time was used for billed work to the client. Utilisation is used across a wide range of industries, but particularly in professional services firms.

    A KPI definition of Utilisation

    So, here’s a plain English definition:

    ‘Utilisation is the ratio of billed hours to available hours for an individual.’

    That seems pretty simple, doesn’t it? But let’s look at some of the questions that this does not answer:

    • For 'available hours' do you mean nominal, or actual, and do you include overtime?
    • What about holidays or sickness absence, is that zero utilisation?
      By individual do you mean only individuals who are billable, or do you include support staff?
    • Do you include mandatory commitments, such as safety training, in the available hours?
    • What about sales? This is clearly a value-add activity but is not billable.
    • What happens if you bill at a much reduced hourly rate? Does that utilisation count the same as full-rate work?
    • What do you do with staff who have dual roles, only one of which is a billable position?
    • Clearly there is plenty of scope for misunderstanding here. It’s pretty rare to find a plain English definition that does not give rise to lots of similar questions.

    It looks like our initial definition has lots of 'holes' in it. How can we make sure that our definitions don't suffer from the same issues?

    How to define your KPIs in 10 simple steps

    Coming up with clear simple definitions doesn't have to be complex. Download the ROKS KPI Canvas (for free) and complete each section by answering each of the questions below....

    Sample ROKS KPI Canvas - KPI Definition for Safety metric

    Typical time to create first simple definition: 20 minutes

    Things you will need

    ROKS KPI Definition Canvas (free download)
    Subject matter experts from your organisation on systems or process being measured

    Method Steps

    Step 1

    Give your KPI a unique name

    It is essential that management decisions are never made by comparing different measures that share an identical name.

    For example, if you have several measures that relate to efficiency, make sure you give them distinctive names such as 'Line 7 Bottling Efficiency', rather than just 'Efficiency'. The key thing is that there’s only one measure for with each name.

    Step 2

    Record purpose and description

    Why should we measure this?

    Sometimes we lose sight of why we are measuring something. This is particularly common in very top-down businesses where the boss maybe asks for some specific data - analysts are often too scared, or beaten down, to ask what it’s going to be used for. You will have thought carefully about the purpose of each measure in the Shortlisting step (a powerful way of making sure we don't end up with too many key performance indicators), so this should not be too hard.

    Step 3

    Identify your reporting customers - [Optional]

    Who will use this metric?

    We produce reports to help in decision-making. If you are a one-person business, then this is a very easy question to answer. If there are more people in your business, you need to have a clear understanding of who uses the data. That way, if a question comes up about how we measure or report something we can talk to our KPI customer and discuss it with them, instead of just trying to guess what they want. If there’s no clear customer or decision being made on the back of a KPI you should seriously think about chopping it.

    Step 4

    Identify data sources precisely

    Where will the data come from?

    One of the most common sources of errors in performance metrics comes from holding similar data in multiple places. If you have lots of spreadsheets holding data that is similar, but maybe slightly different in content or scope, then you have a real risk of mistakes creeping in. This is particularly true if you have more than one person involved in producing KPI reports. This field needs to go down to painful levels of detail.


    Rather than just saying ‘From the production performance spreadsheet’, you should specify the name of the server the document lives on, the folder, the precise sheet name, the tab in the spreadsheet and the column-row range. Your test for whether you have done this properly is whether someone who hasn’t used this data before can successfully navigate to exactly the right data by following the instructions shown under Data Sources. 

    Step 5

    Record KPI formula or calculation

    If there’s any calculation, how is the measure calculated?

    What is and is not included in the values used?

    It’s important to be crystal clear about the calculation used, and what is and is not included in the calculation.

    Step 6

    Document production resources - [Optional]

    What resources are needed to produce the KPI and reports?

    Sometimes you will need input from other teams, a particular individual to provide or analyse data. It’s not uncommon to need data from a particular system, or piece of software, that only a specific person is able to extract. Knowing who, or what, you need to produce the data enables you to plan and spot dependencies that may cause problems.

    Step 7

    Set and record targets

    What score do we want to achieve? (If we know at this stage)

    Targets can get pulled out of thin air and can sometimes change over time. Record what target we are trying to achieve so that we have a record. If it changes, fine, but let’s record the new and the old, so we can see how things have changed.

    If you want to get into much more depth on targets and incentives, check out 'GAMED: Why targets and incentives fail and how to fix them'

    Step 8

    Describe the benefits or purpose of the target

    What will achieving the target deliver?

    If you have nice round targets (10%, 50% or 100% improvement) ask ‘Why that particular figure?’. The most powerful and compelling targets are ones that are linked to outcomes.

    Backup your targets with reasons, like this…

    • If we achieve a turnover of $200k we can become a government approved supplier.
    • If we undershoot our budget by $11,500 we can invest in a new espresso machine.
    • If we cut downtime by 14% we can all have an extra day of vacation this year.

    Using outcome-linked targets in your KPI definition can make them much more meaningful. Notices this format is very close to Objective Key Result (OKR) approach.

    Step 9

    Estimate production cost - [Optional]

    What is the cost of implementation and production?

    KPIs don’t come for free. Unless they are a 10 for Ease of Measurement/Availability, there’s time and effort involved in collecting and collating KPI data.

    Step 10

    Problems and Errors

    What are the known issues with production & accuracy?

    Pretty much all KPIs have issues. It’s one of the most common reasons I hear for not bothering to measure something - ‘We could try and measure that, but it’s pointless because…’.

    All KPIs are flawed to some degree. They crunch question is; ‘Is the information delivered by the flawed KPI better than no information at all?’.

    If the answer to that question is ‘Yes’, then we need to start collecting that data but be very clear and open about the problems and limitations of that KPI. This section is where we record, as honestly as possible, what the known problems and errors with the KPI and its source data are.

    By keeping this record up-to-date we encourage ourselves to put the right level of trust in the KPI and also to try and fix the known problems.

    Your pre-designed KPI definition solution - the ROKS KPI Canvas

    There is quite a lot to remember here. So here's the good news, you don't have to. All of these are covered in the free printable ROKS KPI Canvas . Just fill in the download form below to download your printable PDF template!

    Free ROKS KPI Definition Canvas Download

    The ROKS KPI Canvas serves as a structured memory jogger to help you navigate through the key questions you need to answer in order to fully define your KPIs.

    Here is what a blank ROKS KPI Canvas looks like:

    ROKS KPI Definition Canvas

    You can download a nice crisp, print-ready, PDF of the ROKS KPI Canvas here...

    Sign up and get this free download!

    Guide to using the ROKS KPI Canvas tool

    Each of these questions needs to be answered for each KPI and measure in the organisation. Tedious? Yes. Important? Also, yes. The definitions should be held in a single managed location so that multiple definitions don’t coexist. This will avoid confusion and arguments.

    The benefits of this kind of KPI definition are:

    • It forces your organisation to clarify and discuss the KPI definitions.
    • Any weakness or uncertainty around a KPI is written down and ‘out in the open’.
    • You avoid having similar sounding (or identical) KPIs or measures that are actually calculated in different ways.
    • It is a reference document for people who are unclear or uncertain how a measure is calculated.

    Setting up KPI definitions well is a bit like flossing regularly. Most people agree you should do it, but very few people do it regularly, or well.

    Copyright and usage

    The ROKS KPI Canvas is covered under the Creative Commons Attribution-NonCommercial-ShareAlike License. This license lets others remix, tweak, and build upon our work non-commercially, as long as you credit us and license their new creations under the identical terms. Full details live here:

    Whether you fill it in using the forms capability of Adobe Acrobat Pro, print it out - covering it in Post-Its and ink or recreate it in the software of your choice, it should help you properly define each and every one of your shortlisted KPIs. (But only if you use it of course).

    Why are some bits of the ROKS Express Canvas ‘Optional’?

    As this method is focused on a wide range of business sizes, not all the steps in the method make sense for all sizes of organisation. If you are a sole trader, you really are not going to have to worry too much about who the ‘Customer’ for the KPI is, it’s clearly going to be you. Similarly, it’s fairly obvious what the ‘Production resource’ is going to be in a one-man-show too.

    Paper form or spreadsheet?

    Whilst the ROKS KPI Canvas is shown as a sheet, one that’s designed to be printed off and scribbled on as you define your KPIs, it’s the questions and headings that are important. You can easily build an Excel template, with each section as a heading, but I prefer to start with a paper sheet, then stuff the definitions into a spreadsheet once they are worked out. That’s just my preference, if you are comfortable going straight to the spreadsheet and entering the definitions - that’s fine - it will save you some time.

    Looking for a jump start on your KPI definitions?

    If you need a library of over 400 KPIs with explanations, worked example and basic definitions, check out Getting Started with KPIs.

    For topic-specific KPI Trees, including definitions, worked example and notes for the most useful KPIs, check out our Premium KPI Packs.

    How to share your KPI definitions

    Sometimes you need to explain a KPI to new team members, as part of training or as a refresher for existing team members. Although the ROKS KPI Canvas can do this job pretty well, there is a tweaked format, called the KPI Cheat Sheet, which focuses more on the background and examples side of things. You can download a KPI Cheat Sheet from this page.

    4 simple examples of KPI definition

    Here are some common KPI examples. The full definition will depend on your specific situation and intent - for example data sources will be entirely dependent on your data collection approach and systems - so these are only 'starter' definitions and will need refining for your specific situation.

    Here are some common KPI examples. The full definition will depend on your specific situation and intent - for example data sources will be entirely dependent on your data collection approach and systems - so these are only 'starter' definitions and will need refining for your specific situation.

    KPI Definition - Frequently Asked Questions

    KPI stands for 'Key Performance Indicator'. 'Key' means that the metric is important to someone, 'Performance Indicator' is simply a figure that shows us how we are doing at that thing. Put another way, it is a quantifiable measure of performance for something important. Check out this short video for more detail:

    This guide is focussed on defining KPIs once they have been selected. If you are looking to select your KPIs, start with our Ultimate KPI Guide for a structured approach to doing this.

    KPI definition can be tough and time-consuming to implement. It is possible to gamify the definition process by counting the completed fields on each KPI definition column in your KPI definition spreadsheet and adding a heat map and completion score to to the sheet to give an 'at a glance' assessment of our definition performance over time.

    There are two principles we need to apply to decide if a KPI is 'good'. Firstly measure the most important things, but also focus on the measures that practical and possible. This is why we need to make sure we selected our KPIs using the KPI Tree approach followed by the Shortlisting method, where we rate each of our KPIs based on 'importance' and 'ease of measurement'.

    The exact 'right' number of KPIs is something that is unique to your business. There are some predictable factors to consider though, these include:

    • Size and complexity of what you are trying to measure
    • You KPI strategy, taking into account the measurement maturity of the organization
    • The availability of data, ease and cost of collation

    Because of the wide range of potential situations, it's not possible to come up with a single figure. As guidance, an organisation with low metric maturity should typically aim to implement no more than five KPIs in their first implementation wave. The key message is that it's not about how many KPIs we have but selecting the right key performance indicators.

    SMART is a five point checklist based on the work of Locke and Latham and really a goal-setting method, not a KPI definition tool, so in truth there is no such thing as a SMART KPI. SMART stands for Specific, Measurable, Assignable, Realistic and Time-related in the original version proposed by George T. Doran in the early 80s. SMART has some value to add when it comes to designing targets, but it's quite possible for targets to meet the SMART criteria and still encourage dysfunctional behaviour. Learn about a better method for designing targets and key results using the ROKET-DS approach.

    Yes! You can download a complete pack of essential finance KPIs here, along with the KPI Tree used to select them.

    For web marketing KPIs, check out our Premium Web Marketing KPI pack. For more support developing bespoke marketing KPIs, get in touch.

    This is a big topic, but good report and dashboard design, following sound visual design principles is the place to start.

    A leading measure is one that gives advance (leading) information related to our intended outcome. This type of measure is prized because it allows action to be taken in advance of the outcome we care about. Example: Customer satisfaction - A low customer feedback score gives you a chance to intervene, potentially recover that customer and protect their repeat custom.

    A lagging measure tells you what happened after the thing you care about. Example, 'closed accounts' might be a lagging kpi for customer satisfaction (or dissatisfaction). Once the customer closes an account it is likely to be much harder to win them round to using our services again.

    There are lots of strong opinions on leading and lagging indicators, but the truth is that it's impossible to run a business just on lagging or solely on leading measures. Most organisations manage their performance measurement using a mixture of both.

    The best approach for selecting the right KPIs is using the 'Results Orientated KPI System' (ROKS Enterprise). This approach allows anyone to turn their strategic objectives into a 'longlist' of candidate KPIs then select your shortlist of KPIs based on importance and ease of measurement.

    Just defining KPIs is not going to deliver improvement. Those KPIs need to feed well-run meetings in the form of KPI dashboards or reports.

    'Metric' refers to all types of performance measure, a 'KPI' is a type of metric that allows us to measure the amount of 'something' that will significantly impact your business or decision-making.

    If you need a library of over 400 KPIs with explanations, worked example and basic definitions, check out Getting Started with KPIs.
    For topic-specific KPI Trees, including definitions, worked example and notes for the most useful KPIs, check out our Premium KPI Packs.