How many Key Performance Indicators (KPIs) can you have before none are actually ‘key’? Using theory and practice, Paul Rogers tries to answer this question.
Picture the scene. You’re at the first contract review meeting for a major services contract. The supplier’s report runs to 47 pages. There are 26 separate KPIs, each with a target, weighting, calculation method and traffic-light status.
The meeting room has been booked for two hours, but everyone knows that is optimistic. The first batch of performance-related KPIs takes 45 minutes to discuss. People are glancing casually at their watches, working out how much time is left for the remaining 20 KPIs.
We’ve all been in versions of that meeting. The instinct is understandable. We want accountability, transparency and continuous improvement. However, what we’ve got is frustration.
Part of the cause is that the quality of the conversation is compromised by the sheer number of metrics. We have KPIs for operational performance, timeliness, work quality, any rework, data quality, reporting and customer service.
And that is before we add in sustainability, workplace health and safety, and social procurement metrics. So, are they all ‘key’, or are some more key than others?
Cognitive Load Theory
There is a story about Albert Einstein that claims he reached for a notebook when asked for his telephone number.
“Don’t you know your own telephone number?” asked the other person. Einstein replied, “Why should I memorise something that I can so easily find in a book?”
The implication is that the great man deliberately kept trivial facts out of his mind to preserve mental space for important ideas. While the story may be apocryphal, the idea that we have a finite working memory is supported by research.
Cognitive Load Theory explores the architecture of human cognition and specifically the severe limitations of working memory when dealing with novel or complex information.
According to the theory, our working memory can typically hold and manipulate only about four chunks of novel information at once. This means that more chunks of information create ‘cognitive load’.
The theory posits three types of cognitive load. Using the contract performance review meeting as an example:
- The intrinsic cognitive load is the inherent complexity of the task, such as understanding the interdependencies in a complex outcome-based services contract
- The extraneous cognitive load is the load imposed by poor presentation, redundant matrices, inconsistent definitions or too many data points
- The germane cognitive load is the mental effort needed to make connections and understand root causes and generate insights or improvements
The point is that if the sum of the first two cognitive loads exceeds our working memory, then there is no capacity left for the important part: the germane cognitive load of identifying root causes and securing improvement, which is the ultimate goal, isn’t it?
The impact of AI
Artificial intelligence complicates this picture rather than giving us a simple answer.
Used well, AI can reduce cognitive load by summarising long reports, spotting exceptions, drafting meeting notes and turning raw data into clearer options for human judgement.
In that sense, it can remove some extraneous load and free people to focus on the germane work of asking better questions, identifying causes and making decisions.
However, AI can also increase cognitive load if it produces more analysis, dashboards, recommendations and follow-up actions than people can reasonably absorb.
The productivity gain may lift throughput, but it may also raise expectations about how much knowledge work can be processed at the same time.
The real test is not whether AI makes more information available. It’s whether it helps people see what matters faster. In contract management, that means AI should support the discipline of choosing the vital few measures, not become a faster way to generate another 20 KPIs.
From theory to practice
The simple truth is that contracts are managed by people, not spreadsheets. Our brains have genuine limits on simultaneously tracking, interpreting and acting on many performance streams while retaining capacity for judgment, relationships and creative problem-solving.
The contract review becomes a compliance exercise rather than a vehicle for delivering real value.
Let’s take the two-hour meeting. Assuming no small talk or breaks, there are fewer than five minutes to review each KPI. That’s not realistic, is it?
Fitting everything important into a workable number
Another apocryphal quote attributed to Albert Einstein is, “Not everything that can be measured matters and not everything that matters can be measured”.
Most contracts carry expectations across core performance, reporting, workplace health and safety, and social procurement goals. Visibility in all areas is reasonable. The question is whether all dimensions are equally important. Are they all key?
In some cases, the answer is no. A practical target is six to ten KPIs for the contract’s primary purpose. These are the vital few that should directly influence payments, extensions or escalations. Supporting dimensions can also be measured but as a secondary suite of indicators.
Returning to the services contract, this might include overall user satisfaction, critical system uptime affecting operations, cost performance versus benchmarks, consolidated measures of responsiveness and resolution for priority issues, examples of delivered innovation, and a couple of leading safety indicators where predictive.
That’s seven KPIs. Detailed WHS obligations and compliance sit in a dedicated management plan with audits, supplemented by one or two leading indicators, such as near-miss reporting.
The remaining useful indicators become Secondary Performance Indicators (SPIs). Reported in aggregated or exception-based formats via dashboards, they support diagnosis and conversation but do not automatically drive commercial remedies unless trends warrant it.
This tiered distinction preserves everyone’s mental capacity for high-value work while maintaining broad visibility.
In practice, SPIs provide context and early warnings without dominating reviews or commercial mechanisms.
A clean, visual dashboard that highlights the KPIs and shows secondary measures only on demand or by exception will reduce the cognitive load of a thirty-slide PowerPoint pack or a 40-row spreadsheet.
Why this matters beyond meetings
Imagine you could only have seven KPIs. You would be forced to focus on the contract’s essence and relegate other metrics to a secondary tier. Might that be a good thing?
The resulting conversations will be richer and more relevant to what really matters. The challenge is not setting a ceiling on the KPIs, but selecting which measures will help deliver the outcomes we seek.
Ten or fewer KPIs, carefully chosen, tiered with SPIs and supported by smart reporting, isn’t a compromise on rigour. It represents the professional standard for contracts designed to be actively managed and improved, rather than merely administered and audited.
If everything is key, nothing is. Good contract management starts by choosing what deserves attention.
About the author
Paul Rogers, Practice Manager at Landell, brings more than 40 years of experience to the table, having worked in more than 20 countries as a procurement professional, consultant and trainer.
An internationally-recognised expert in procurement and negotiation, Paul has worked on a diverse range of projects, from buying a billion dollars’ worth of gold in an off-market transaction to negotiating with an airline on behalf of a government that owned it.
He’s also the chair of the judges for PASA’s coveted Nigels Awards and is a fellow of CIPS (FCIPS).
