Calculating and reporting key performance indicators (KPIs) is a routine practice in the energy and chemicals industry. It is key to helping decision makers act, and helping managers oversee performance. Digitalisation tools like business intelligence (BI) and dashboards are helping to increase the visibility and timeliness of KPIs, break down organisational silos and drive consistency across organisations. However, this is not enough in itself.
The challenges of poor underlying data, difficulties of setting targets and the interactions between conflicting KPIs mean that, all too often, KPIs are not effective in delivering any true improvement in performance. At the same time, digitalisation is changing organisational responsibilities, and predictive analytics are changing information from being about the past, to being about the future.
This article outlines the basic principles and objectives of KPIs with regards to achieving operational excellence, discusses some of the root causes for ineffective KPIs, and then explains how various digitalisation technologies and approaches can transform the effectiveness of KPIs in business models.
Purpose and principles of KPIs
KPIs are an important ingredient in achieving operational excellence. KBC believes the two facets of operational excellence are:
- Making better decisions, faster.
- Perfect execution of those decisions, every time.
Within the first, leading KPIs provide timely decision support information to support people to make the right decisions in a timely way.
In the second area, lagging KPIs measure the effectiveness of the entire process (both decision making and execution) to allow fine tuning or modification to the work processes, physical processes or organisation to take place.
Given that leading KPIs are intended to support decision making, it is essential to understand how people make decisions. In particular, how people make decisions in environments where there is limited time and limited information – such as most energy and chemicals facilities.
In this type of environment, decisions are typically made through recognition primed decision (RPD) making, which is fast, natural and intuitive, and effective when deployed by experts. The typical process is as follows:
- Situational awareness:
- Based on visual information about the situation, the expert makes a mental ‘pattern match’ to determine what is happening.
- The expert interprets the meaning of the event, and then makes a mental simulation of what will happen next.
- Identify options:
- Determine the root cause of the issue.
- Search for solutions, typically from experience.
- As solutions are found, check if it will suffice to solve the problem.
- Implement the first ‘sufficient’ solution, then stop searching for more solutions.
- Data visualisation such as colour coding changes the KPI from a number on a screen somewhere to something highlighting that action may be required.
- Data aggregation, to bring together data from multiple sources can allow KPIs to be viewed in comparison to each other, and makes decision makers aware of situations outside their direct silo. This can make decision making more holistic without needing meetings or calls to exchange information.
Leading KPIs play an important role in this process by providing a top level of situational awareness, by flagging up and highlighting issues, and potentially providing a simple overview of all the possible issues that need to be considered.
KPIs can help provide situational awareness by explicitly identifying the most important factors rather than hoping it will be spotted amongst the weeds of many data points.
Current simple digitalisation technologies such as data integration and dashboards greatly enhance the power of situational awareness KPIs:
KPIs can be applied on a number of different timescales and at all levels in an organisation. Figure 1 illustrates one of the decision timescales.
Figure 1. The digitally wise asset decision cycle.
So, for instance, top level business financial results (lagging KPI) are reported quarterly or annually, whereas operational decision making takes place on a timescale of days, hours or minutes. One of the ways digitalisation adds value to an organisation is compressing the decision timescales, thus reducing losses due to uncertainty. This is key to understanding how digitalisation changes KPI management, as will be discussed later.
Having established the theory behind KPIs, this article will now examine some of the things that make KPIs ineffective.
KPIs, and the situational awareness they bring, are only as good as the data being used to calculate them. It is commonplace for experienced operators to ignore or reject data when they ‘know’ (either rightly or wrongly) a certain meter might be wrong. The entire system is dependent on the inputs being correct, and being able to detect and highlight errors in the data.
To make KPIs effective requires some sort of measurable target, or target range, to be set. There are various interlocking issues that make target setting a challenge:
- Motivational risks: set a target too high and people become demotivated and complacent; put a target on someone that is not within their realm of control and they will feel persecuted. Even targets that are appropriate can be perceived as top down micromanagement rather than useful ‘situational awareness’ to empower the decision maker.
- Conflicting targets: particularly in the energy and chemicals industries there are many trade-offs between competing priorities, for instance energy efficiency improvements may compromise process yield, pushing up throughput can compromise reliability and maintenance costs. Targets set in isolation can drive narrow improvements that make things worse from a holistic perspective.
- Volatile external environment: what is best today may not be best tomorrow, as operating strategies can change, and the relative value of different objections is not fixed. Targets need to be smart enough to adapt to changing circumstances.
- Keeping the key in KPI: in very complex environments, there are a lot of things to pay attention to. As such, there is a general tendency to have too many KPIs on too many dashboards, which starts to erode the very purpose of the system. Research shows that human cognitive capacity can only cope with approximately seven things at any one time, so large numbers of KPIs, however important, are cognitively difficult to process.
Lack of action
Many KPIs are regarded with a ‘so what?’ attitude by the users. Correct data and targets do not necessarily result in corrective action. Training is the common method of solving this with simple steps to solve the typical out of target KPIs. However, the method is limited when trying to consider the complex environments with multiple variables acting on the KPI, this results in either additional time and effort spent in troubleshooting the issue or disregard of the KPI as a lower priority when compared to KPIs that return to ‘normal’ when the known action is applied.
This is part one of a two-part article. Part two will be available shortly.
Written by Duncan Micklem, KBC (A Yokogawa Company), USA.
Read the article online at: https://www.hydrocarbonengineering.com/special-reports/20052019/transforming-kpi-effectiveness-part-one/