KPI management, benchmarks, and Artificial Intelligence (AI) do not always explain your organization’s unique challenges. For decades, organizations have invested significant resources to improve their ability to understand their own performance.
Tools such as market research, analytics, and, more recently, AI have taken on a central role in decision-making processes.
This evolution has brought undeniable benefits. Today, we can consolidate large volumes of information, identify patterns, track results in real time, and support decisions with a level of speed and precision that would have been unthinkable just a few years ago.
However, there is one question that, while obvious, rarely comes up when discussing KPIs, performance, or Artificial Intelligence:
- An insurance company does not face the same challenges as a retailer.
- A fintech does not operate under the same conditions as a manufacturer.
- A Shared Services Center faces completely different challenges than those found in decentralized or departmentalized structures.
So why do so many organizations continue to use virtually the same KPIs, benchmarks, and references to evaluate their performance?
The question might sound purely provocative, but it deserves careful thought.
Over recent years, organizations have adopted benchmarks, best practices, and reference models as tools to accelerate their growth. This is a healthy and necessary practice.
The problem arises when these references are no longer used as comparative tools for evaluation and instead begin to be treated as universal truths.
When the Benchmark Becomes the Final Goal
You have likely been in a meeting where questions like these were asked:
- “What is the market benchmark?”
- “What is the industry average?”
- “How are leading companies measuring this?”
These are legitimate questions. The issue arises when the follow-up question is never asked:
Do these indicators actually explain our operational challenges and help us set goals that drive meaningful results and market growth?
Case Study: The Trap of Unsustainable Growth
As a case in point, the strategic leadership of a grocery retail chain recently hired executive directors focused on growth and expansion. Benchmarking was conducted, and the project was launched.
A year later, multiple stores had opened, and revenue and growth indicators looked positive.
The expansion was deemed a success, at least according to the metrics, benchmarks, and marketing reports. Yet, the directors were let go at the end of the year.
This decision was made because the company ended the year in the red with increased debt, leading to no profit distribution and serious doubts about the financial sustainability of the business.
Benchmarks help organizations learn from successful experiences, mitigate risks, and design strategies. For this reason, they remain an extremely valuable tool.
However, a benchmark represents an industry reference point, not a template to copy or an off-the-shelf strategy. Benchmarks are simply data points that, if misused, can create significant problems.
Contextualizing Data to Operational Reality
For instance, consider a study indicating that 30% of accounting reconciliations in a given sector are still performed manually.
What does this information actually mean? It depends.
For a business processing a few hundred entries a month, it might not mean much. For a retailer processing thousands of daily transactions, it could mean the exact opposite.
In this scenario, the relevant question is not: “Am I performing better or worse than the benchmark?”
The right questions should be: “Does this benchmark reflect the operational reality I need to manage? Does it help me better understand my market?”
Benchmark data can complement analysis, but it can never replace a deep understanding of your own operations.
The NPS Case: The Right Indicator for the Wrong Questions
Perhaps one of the best examples of this debate is the Net Promoter Score (NPS). In recent years, it has become common to see articles, videos, and posts declaring the “death of NPS.”
Is the metric itself actually the problem? Most likely not.
NPS remains highly effective at answering the specific question it was designed for: “How likely is a customer to recommend your company?”
Read more: Customer satisfaction: How is it monitored?
The Risk of Distorting a Metric’s Core Purpose
The issue arises when we try to force it into answering questions it was never intended to address, such as:
- “On a scale of 0 to 10, how would you rate our overall performance and attentiveness?”
- “Considering the effort made by our team, how would you rate the service received?”
- “On a scale of 0 to 10, how would you rate the customer service representative?” (a question often asked when the service outcome was favorable to the customer).
The Limitation Lies in the Application, Not the Tool
At that point, we are no longer evaluating NPS. We are distorting a metric to get a simple answer for a complex evaluation, seeking results it was never meant to deliver.
The limitation does not rest with the indicator, but with how we choose to apply it. The same holds true for many corporate metrics.
They remain valuable tools provided we clearly understand what they measure, what their limits are, and which aspects of reality fall outside their scope.
Continue reading: What is QMS and How It Can Transform Your Company
Artificial Intelligence: The Challenge of Asking the Right Questions
Artificial Intelligence has introduced an extraordinary capacity to analyze data and uncover patterns.
However, there is a crucial nuance. AI learns from the specific data, metrics, and references we feed it, which is a fundamental principle in discussions around data quality for AI.
In other words, AI will not automatically correct or critique a narrow worldview. It simply operates within it.
If the initial prompt or question is wrong and the chosen indicators fail to capture the organization’s real challenges, there is a high probability that the resulting insights will also be inadequate or inaccurate for sound decision-making.
Technology can detect patterns with incredible precision, yet it still depends on our ability to decide which patterns actually matter.
Perhaps that is why one of the most critical questions in the AI era is not: “What can AI answer?”
The right question to ask is: “Are we asking the right questions?”
What We See in Practice: Why Technology Alone Is Not Enough
This discussion moves beyond theory when we look at real-world organizational transformation projects.
It is common to see companies equipped with:
- Sophisticated dashboards
- Dozens of indicators
- Automated processes
- High levels of digitalization
And yet, they continue to struggle with issues such as:
- Rework
- Low operational predictability
- Integration hurdles across departments
- Decision-making bottlenecks
- Misalignment between strategy and execution
- Hidden costs
- Customer complaints
In these cases, the problem rarely lies within the technology itself.
More often than not, the issue stems from how processes are defined before automation, which metrics are selected to track results, and, above all, a lack of clarity regarding the factors that truly drive organizational performance.
How to Align Technology, Processes, and Governance
For this reason, in initiatives led by Xcellence, implementing integrated management and automation solutions is typically preceded by a thorough analysis of the processes, governance, and measurement mechanisms that sustain the operation.
Platforms like SoftExpert Suite enable organizations to integrate information, structure workflows, automate controls, consolidate KPIs, and significantly expand operational visibility quickly and efficiently, ensuring they can answer the right questions.
Technology, therefore, is only part of the equation. Its effectiveness depends on structured KPI management, supported by clearly defined processes and monitoring criteria that reflect the specific needs of each business.
After all, automating a flawed process only allows an organization to execute mistakes faster.
Likewise, tracking irrelevant indicators simply makes observing non-critical issues more efficient.
Final Considerations: Are We Measuring the Right Things?
Metrics matter. Benchmarks matter. Artificial Intelligence will continue to grow in importance. The problem does not lie with these tools.
The challenge arises when we believe these tools alone can explain the full complexity of an organization. That is why one of the most vital questions for modern leaders remains surprisingly simple: are we measuring the right things?
What we choose to measure dictates what we are able to see. And what we see directly shapes the decisions we make.
Ultimately, the primary challenge in KPI management is not the volume of metrics your organization tracks, but whether those metrics accurately explain the problems you need to solve.
FAQ: Frequently Asked Questions About KPI Management
Here are some common questions regarding the topics covered in this article:
It is the structured process of defining, tracking, and acting upon performance metrics aligned with business strategy. It helps organizations connect day-to-day operations to long-term goals and supports data-driven decision-making.
A metric is any measurable data point within a business. A Key Performance Indicator (KPI), on the other hand, is a critical metric specifically selected to evaluate progress toward key strategic objectives.
Lagging indicators measure past outcomes, such as monthly revenue. Leading indicators are predictive metrics that signal future performance, such as customer satisfaction or lead generation.
Benchmarks represent average industry data rather than tailored strategies. Every company operates under unique circumstances. Treating a benchmark as a universal truth without considering operational context can lead to strategic missteps.
No, NPS is highly effective for measuring customer advocacy. The issue arises when organizations force it to answer complex performance questions it was never designed to address.
AI excels at detecting patterns, but it relies heavily on the data provided. If an organization uses incorrect indicators or poor data quality, AI will simply produce inaccurate or limited insights.
Technology merely speeds up execution. Automating flawed processes or tracking irrelevant indicators only ensures that errors are made and wrong metrics are monitored at a faster pace.
