Yesterday’s post drew a comment from Vera Rozanova, MEng (Hons), MBA, FCIPS Chartered, FCILT, that went further than the post had.
Her observation was that the measurement system itself determines the boundaries of what an organization can discover. We tend to think of analytics as a window onto reality, she wrote. In practice it is also a filter. The risk is having perfectly accurate data that makes the wrong questions easy and the important questions impossible — which she named as one of the most important forms of strategic blind spot in modern organizations.
My reply was that this is exactly the deeper implication. The danger is not necessarily inaccurate data. It is accurate data inside an incomplete field of vision. Which creates a problem: if the measurement system determines what questions can be answered, how do we recognize the important question the system was never designed to answer? In 1998 that question was what time of day the orders came in. In 2026 the trace has again reached the boundary of the instrument. Sometimes the next step is not better analysis. It is realizing you need a different window.
This post takes up where that exchange left off, because business management has been working on Vera’s problem for thirty years, the two most successful attempts both stopped one level short of it, and one of the two authors has said so himself.
Two answers to the same complaint
Kaplan and Norton’s Balanced Scorecard began from the observation that financial measures were necessary but insufficient. An organization judged on its financials alone was being judged through a window too narrow to show what produced them. So they widened it: customer, internal process, learning and growth. A colleague, Luis Lima, put the book in front of me in the same period as the defence engagement, which is why the two have sat next to each other in my thinking ever since.
John Elkington’s Triple Bottom Line, a term he coined in 1994 and built out in Cannibals With Forks: The Triple Bottom Line of 21st Century Business (Capstone Publishing, Oxford, 1997), made a structurally similar move on a different axis. Financial performance alone was too narrow a definition of performance. Social and environmental results had to enter the picture.
Both are answers to the same complaint. The frame is too narrow. Add dimensions.
Neither addresses the filter.
The recall
In June 2018, in the Harvard Business Review, Elkington did something almost nobody in management does. He issued what he called a management concept recall on his own idea, observing that he could not think of a single prior case of a concept being recalled by the person who invented it.
His reason is the part that matters here. The Triple Bottom Line had failed to displace the single bottom line. Built to force a rethinking of how business creates value, it had instead been absorbed as an accounting tool — a way of balancing trade-offs rather than a reason to do anything differently. Organizations added the dimensions and kept the behavior. Progress across the available options, he noted, had not been benchmarked against real-world impacts.
That is not a critic’s charge. It is the author’s own assessment, published under his name, twenty-four years on.
What the record already carried
I have been working with Elkington’s material since 2008, and the record is dated, so the sequence can be checked rather than asserted.
In April 2008 I published a forty-seven-page white paper, The Greening of Procurement, through CATAAlliance, built around Cannibals With Forks. That September I delivered the same material as a two-day seminar for the Ontario Institute of the Purchasing Management Association of Canada — a session an attendee described that October as one where the room worked through what the personal and corporate impact of the material actually was.
Three things in those documents belong next to the 2018 recall.
The mechanism. The seminar recorded that in the public sector the Triple Bottom Line had become the dominant approach within Full Cost Accounting methodology, that expert opinion on its reliability as a way of quantifying impact was mixed, and that the impact of sustainability programs was therefore usually confined to traditional management consulting principles. Written in September 2008: the frame had become accounting, the accounting could not reliably quantify, and behavior defaulted back to the traditional model.
The measurement. The white paper’s section titled A Preliminary Scorecard? published the EcoMarkets 2007 Survey — 692 procurement professionals across Canada and the United States, representing organizations with more than five billion dollars in combined annual spend. Ninety-one percent said they considered environmental factors at least occasionally. The survey’s own conclusion was that the translation of policy into practice was mixed and incomplete. Asked what actually drives purchases, respondents put product performance, durability and price above environmental considerations. Sixty percent would not pay any premium for environmental advantage.
The diagnosis. Writing about the borrowed aphorism at the heart of Elkington’s title — is it progress if a cannibal uses a fork — I documented what I saw as the weakness in how the frame was being taken up: doing something because it is supposedly good for you without really quantifying what the good actually is.
I want to be exact about what that does and does not amount to. Elkington’s recall and my 2008 position are not identical complaints. His is that a frame built to change capitalism was reduced to accounting. Mine was that the good was being asserted rather than quantified, with survey data showing the distance between stated policy and actual spend. What they share is the underlying failure mode: adding a dimension to the frame is not the same as knowing what that dimension determines. He arrived at it from inside the framework he built. I arrived at it from procurement, and the documents carry the date.
Why widening does not reach the problem
Adding dimensions still assumes you can name the right dimensions in advance.
That assumption does its work before the measuring starts, which is why it stays invisible. Every scorecard, balanced or otherwise, is a list of things somebody decided were the things to watch. The measurement can then be flawless and the reporting accurate to two decimal places, while the condition that determines the outcome sits entirely outside the list — not through carelessness, but because nobody thought to put it there.
This is Vera’s point, and it is why the filter matters more than the frame. A wider frame admits more light. It does not change what the instrument is built to select for. And breadth makes the selection harder to notice, not easier: two measures invite the suspicion that something is missing, while twenty measures across four perspectives, each with an owner and a target, produce the opposite feeling. Comprehensiveness reads as coverage.
The same white paper contains the proof, on the money axis
The last section of that 2008 paper was titled The Financial Disconnect, and it has nothing to do with sustainability.
Citing a 2006 Aberdeen study, it reported that of the 11.9 percent in average identified savings presented by purchasing departments, only 3.2 percent was actually booked by finance — a difference of 73 percent between identification and realization. Fewer than one in five CFOs considered the work of CPOs and their staffs to have a very positive impact on competitiveness.
Look at the shape of that. Purchasing’s measurement was not wrong. The savings were identified accurately, inside purchasing’s own frame, against purchasing’s own definitions. And roughly three quarters of the value died at a boundary that appeared on nobody’s scorecard, because the condition determining whether a saving became real was what finance would agree to book — a handoff, not a metric.
Accurate data. Incomplete field of vision. The determining condition one step outside the instrument.
What a scorecard would not have found
In 1998 I was brought into a defence maintenance operation with a next-day delivery rate of fifty-one percent. Every conventional explanation was available: carriers, warehouse throughput, supplier performance, order accuracy.
You could have built an excellent Balanced Scorecard around that operation. Delivery performance, technician productivity, inventory turns, cost per line, supplier scorecards, cycle times, learning and development. Every measure legitimate. Every measure accurate.
None of them would have surfaced what time of day the technicians submitted their orders.
Time of day was not a performance dimension. It was nobody’s KPI, in that operation or any other I have seen since. It became visible only by walking backward from a specific failure — this order, this day, this technician — until the chain reached a point where the instrument had nothing left to say.
That point was four o’clock in the afternoon. Procurement cannot do anything with an order it has not received.
The difference is direction, not width
Trace backward from a failure and you find the cause. Map forward toward an outcome and you find failures of unknown origin after the fact.
A scorecard is a forward map. It requires the dimensions to be chosen before the failure occurs, which means it can only find what somebody already suspected was worth watching. A backward trace makes no such demand. You do not have to guess which variable matters, because the failure has already selected it. The work is to keep walking until you reach the boundary, and the boundary is the finding.
There is a name for what sits on the other side of that boundary, and the 2008 paper used it. Reviewing the Stern Review of 2006, it recorded Stern’s description of climate change as an economic externality — a transaction producing an effect on an uninvolved third party, an innocent bystander who was never part of the exchange.
That is the structural problem with widening. A frame with three bottom lines instead of one is still drawn around the reporting organization. The externality is, by definition, on the other side of the line.
Which is what the more recent posts in this thread have been tracing. Electric vehicles score well on the environmental line; trace backward to how the battery minerals are mined and you arrive at people who appear on nobody’s scorecard. Western fleets electrify, and the displaced vehicles are traced forward into countries that have only recently gained affordable transport. Two ports rank last globally on efficiency, so automation is warranted — trace backward from the strike and you reach longshoremen whose livelihoods are the cost of the improvement.
In each case the frame is not merely missing a dimension. It has an edge, and the cost has crossed it.
Why any of this could be checked
There is a reason I can put a September 2008 seminar slide next to a June 2018 recall and a 2026 comment about analytics as a filter.
Nothing written in 2008 was written to answer a 2026 question. Nobody in that room was thinking about AI-driven analytics, because there was nothing yet to think about. The EcoMarkets numbers were published because the survey had just landed. The Financial Disconnect section was written because CFOs and CPOs were talking past each other that year. The material was recorded when it happened, by someone who did not know what it would later be needed for.
That is the whole difference between a contemporaneous record and a retrospective account. A retrospective account is assembled once the answer is known, which means it can only contain what the answer needed. A contemporaneous record contains what was true at the time, including the parts that turned out to matter for reasons nobody could have anticipated — and including the parts that turned out to be wrong, which is how you can tell it was not tidied afterward.
Diligent recording in real time is not archiving. It is the only way to have evidence that was not selected by the conclusion.
What this does not claim
It does not make the scorecard wrong, and it is not a replacement.
Continuous measurement and backward tracing answer different questions on different occasions. A scorecard runs all the time and reports whether the things you decided to watch are moving. A trace runs after something has failed and reports why — including, often, that the thing which failed was never on the list.
An organization needs both. What it cannot afford is to mistake the first for the second, and to conclude from a full dashboard that its field of vision is complete.
Elkington’s recall is the most honest evidence available that a wider frame does not, by itself, change what an organization can see. He built the wider frame, and he is the one who said so.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
-30-
Related
From Triple Bottom Lines to Balanced Scorecards and the Backward Trace: What Does Your Window Really Show?
Posted on September 7, 2026
0
Yesterday’s post drew a comment from Vera Rozanova, MEng (Hons), MBA, FCIPS Chartered, FCILT, that went further than the post had.
Her observation was that the measurement system itself determines the boundaries of what an organization can discover. We tend to think of analytics as a window onto reality, she wrote. In practice it is also a filter. The risk is having perfectly accurate data that makes the wrong questions easy and the important questions impossible — which she named as one of the most important forms of strategic blind spot in modern organizations.
My reply was that this is exactly the deeper implication. The danger is not necessarily inaccurate data. It is accurate data inside an incomplete field of vision. Which creates a problem: if the measurement system determines what questions can be answered, how do we recognize the important question the system was never designed to answer? In 1998 that question was what time of day the orders came in. In 2026 the trace has again reached the boundary of the instrument. Sometimes the next step is not better analysis. It is realizing you need a different window.
This post takes up where that exchange left off, because business management has been working on Vera’s problem for thirty years, the two most successful attempts both stopped one level short of it, and one of the two authors has said so himself.
Two answers to the same complaint
Kaplan and Norton’s Balanced Scorecard began from the observation that financial measures were necessary but insufficient. An organization judged on its financials alone was being judged through a window too narrow to show what produced them. So they widened it: customer, internal process, learning and growth. A colleague, Luis Lima, put the book in front of me in the same period as the defence engagement, which is why the two have sat next to each other in my thinking ever since.
John Elkington’s Triple Bottom Line, a term he coined in 1994 and built out in Cannibals With Forks: The Triple Bottom Line of 21st Century Business (Capstone Publishing, Oxford, 1997), made a structurally similar move on a different axis. Financial performance alone was too narrow a definition of performance. Social and environmental results had to enter the picture.
Both are answers to the same complaint. The frame is too narrow. Add dimensions.
Neither addresses the filter.
The recall
In June 2018, in the Harvard Business Review, Elkington did something almost nobody in management does. He issued what he called a management concept recall on his own idea, observing that he could not think of a single prior case of a concept being recalled by the person who invented it.
His reason is the part that matters here. The Triple Bottom Line had failed to displace the single bottom line. Built to force a rethinking of how business creates value, it had instead been absorbed as an accounting tool — a way of balancing trade-offs rather than a reason to do anything differently. Organizations added the dimensions and kept the behavior. Progress across the available options, he noted, had not been benchmarked against real-world impacts.
That is not a critic’s charge. It is the author’s own assessment, published under his name, twenty-four years on.
What the record already carried
I have been working with Elkington’s material since 2008, and the record is dated, so the sequence can be checked rather than asserted.
In April 2008 I published a forty-seven-page white paper, The Greening of Procurement, through CATAAlliance, built around Cannibals With Forks. That September I delivered the same material as a two-day seminar for the Ontario Institute of the Purchasing Management Association of Canada — a session an attendee described that October as one where the room worked through what the personal and corporate impact of the material actually was.
Three things in those documents belong next to the 2018 recall.
The mechanism. The seminar recorded that in the public sector the Triple Bottom Line had become the dominant approach within Full Cost Accounting methodology, that expert opinion on its reliability as a way of quantifying impact was mixed, and that the impact of sustainability programs was therefore usually confined to traditional management consulting principles. Written in September 2008: the frame had become accounting, the accounting could not reliably quantify, and behavior defaulted back to the traditional model.
The measurement. The white paper’s section titled A Preliminary Scorecard? published the EcoMarkets 2007 Survey — 692 procurement professionals across Canada and the United States, representing organizations with more than five billion dollars in combined annual spend. Ninety-one percent said they considered environmental factors at least occasionally. The survey’s own conclusion was that the translation of policy into practice was mixed and incomplete. Asked what actually drives purchases, respondents put product performance, durability and price above environmental considerations. Sixty percent would not pay any premium for environmental advantage.
The diagnosis. Writing about the borrowed aphorism at the heart of Elkington’s title — is it progress if a cannibal uses a fork — I documented what I saw as the weakness in how the frame was being taken up: doing something because it is supposedly good for you without really quantifying what the good actually is.
I want to be exact about what that does and does not amount to. Elkington’s recall and my 2008 position are not identical complaints. His is that a frame built to change capitalism was reduced to accounting. Mine was that the good was being asserted rather than quantified, with survey data showing the distance between stated policy and actual spend. What they share is the underlying failure mode: adding a dimension to the frame is not the same as knowing what that dimension determines. He arrived at it from inside the framework he built. I arrived at it from procurement, and the documents carry the date.
Why widening does not reach the problem
Adding dimensions still assumes you can name the right dimensions in advance.
That assumption does its work before the measuring starts, which is why it stays invisible. Every scorecard, balanced or otherwise, is a list of things somebody decided were the things to watch. The measurement can then be flawless and the reporting accurate to two decimal places, while the condition that determines the outcome sits entirely outside the list — not through carelessness, but because nobody thought to put it there.
This is Vera’s point, and it is why the filter matters more than the frame. A wider frame admits more light. It does not change what the instrument is built to select for. And breadth makes the selection harder to notice, not easier: two measures invite the suspicion that something is missing, while twenty measures across four perspectives, each with an owner and a target, produce the opposite feeling. Comprehensiveness reads as coverage.
The same white paper contains the proof, on the money axis
The last section of that 2008 paper was titled The Financial Disconnect, and it has nothing to do with sustainability.
Citing a 2006 Aberdeen study, it reported that of the 11.9 percent in average identified savings presented by purchasing departments, only 3.2 percent was actually booked by finance — a difference of 73 percent between identification and realization. Fewer than one in five CFOs considered the work of CPOs and their staffs to have a very positive impact on competitiveness.
Look at the shape of that. Purchasing’s measurement was not wrong. The savings were identified accurately, inside purchasing’s own frame, against purchasing’s own definitions. And roughly three quarters of the value died at a boundary that appeared on nobody’s scorecard, because the condition determining whether a saving became real was what finance would agree to book — a handoff, not a metric.
Accurate data. Incomplete field of vision. The determining condition one step outside the instrument.
What a scorecard would not have found
In 1998 I was brought into a defence maintenance operation with a next-day delivery rate of fifty-one percent. Every conventional explanation was available: carriers, warehouse throughput, supplier performance, order accuracy.
You could have built an excellent Balanced Scorecard around that operation. Delivery performance, technician productivity, inventory turns, cost per line, supplier scorecards, cycle times, learning and development. Every measure legitimate. Every measure accurate.
None of them would have surfaced what time of day the technicians submitted their orders.
Time of day was not a performance dimension. It was nobody’s KPI, in that operation or any other I have seen since. It became visible only by walking backward from a specific failure — this order, this day, this technician — until the chain reached a point where the instrument had nothing left to say.
That point was four o’clock in the afternoon. Procurement cannot do anything with an order it has not received.
The difference is direction, not width
Trace backward from a failure and you find the cause. Map forward toward an outcome and you find failures of unknown origin after the fact.
A scorecard is a forward map. It requires the dimensions to be chosen before the failure occurs, which means it can only find what somebody already suspected was worth watching. A backward trace makes no such demand. You do not have to guess which variable matters, because the failure has already selected it. The work is to keep walking until you reach the boundary, and the boundary is the finding.
There is a name for what sits on the other side of that boundary, and the 2008 paper used it. Reviewing the Stern Review of 2006, it recorded Stern’s description of climate change as an economic externality — a transaction producing an effect on an uninvolved third party, an innocent bystander who was never part of the exchange.
That is the structural problem with widening. A frame with three bottom lines instead of one is still drawn around the reporting organization. The externality is, by definition, on the other side of the line.
Which is what the more recent posts in this thread have been tracing. Electric vehicles score well on the environmental line; trace backward to how the battery minerals are mined and you arrive at people who appear on nobody’s scorecard. Western fleets electrify, and the displaced vehicles are traced forward into countries that have only recently gained affordable transport. Two ports rank last globally on efficiency, so automation is warranted — trace backward from the strike and you reach longshoremen whose livelihoods are the cost of the improvement.
In each case the frame is not merely missing a dimension. It has an edge, and the cost has crossed it.
Why any of this could be checked
There is a reason I can put a September 2008 seminar slide next to a June 2018 recall and a 2026 comment about analytics as a filter.
Nothing written in 2008 was written to answer a 2026 question. Nobody in that room was thinking about AI-driven analytics, because there was nothing yet to think about. The EcoMarkets numbers were published because the survey had just landed. The Financial Disconnect section was written because CFOs and CPOs were talking past each other that year. The material was recorded when it happened, by someone who did not know what it would later be needed for.
That is the whole difference between a contemporaneous record and a retrospective account. A retrospective account is assembled once the answer is known, which means it can only contain what the answer needed. A contemporaneous record contains what was true at the time, including the parts that turned out to matter for reasons nobody could have anticipated — and including the parts that turned out to be wrong, which is how you can tell it was not tidied afterward.
Diligent recording in real time is not archiving. It is the only way to have evidence that was not selected by the conclusion.
What this does not claim
It does not make the scorecard wrong, and it is not a replacement.
Continuous measurement and backward tracing answer different questions on different occasions. A scorecard runs all the time and reports whether the things you decided to watch are moving. A trace runs after something has failed and reports why — including, often, that the thing which failed was never on the list.
An organization needs both. What it cannot afford is to mistake the first for the second, and to conclude from a full dashboard that its field of vision is complete.
Elkington’s recall is the most honest evidence available that a wider frame does not, by itself, change what an organization can see. He built the wider frame, and he is the one who said so.
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
-30-
Share this:
Like this:
Related