Hackett has moved from aggregate prescription to readiness-first context. The remaining question is whether readiness is measured against the organization that exists on paper — or the one that actually places the order
In October 2010 I covered Chris Sawchuk’s eWorld keynote on procurement agility, and said something there I have not had cause to revise. I have rarely been a fan of analyst firms, because their perspectives had historically failed to make a dent in the e-procurement failure rate. The Hackett Group I treated as a different case — an operation whose coverage ran broader than the cozy arrangements between industry pundits and the vendors they were supposed to be covering.
Sixteen years later, that assessment holds. Hackett benchmarks what organizations actually do, publishes the numbers, and repeats the exercise year after year. Very few firms in this industry can say that.
Which is exactly why the disagreement is worth revisiting.
2012: the supplier consolidation paper
In 2012 Hackett published The Benefits of Supplier Consolidation Extend Far Beyond Sourcing Savings — document CR_6000128, in its Procurement Executive Insight series. It drew on more than ten years of Hackett procurement benchmark data. Its analysis was restricted to indirect spending. It argued three benefits: lower purchased costs, lower procurement and supplier-management costs, and reduced noncompliance with greater stakeholder satisfaction. Its foundational proposition was that companies eventually buy too many things from too many different suppliers, and its analytical direction was consolidation-first, with provision for approved exceptions.
On 6 February 2014 I wrote that the paper was surprising for all the wrong reasons.
Not because supplier consolidation is wrong. Consolidation makes sense when strategically applied — complex products requiring specialized ongoing support, where the field of capable suppliers is naturally limited anyway. My objection was to the category it had chosen.
Why indirect spend was the wrong ground
Over a fourteen-year monitoring period, eleven years of it after the characteristics were identified, two recurring commodity behaviors emerged.
A Historic Flat Line commodity shows relatively stable pricing over an extended period and a narrow floor-to-ceiling price range. Direct materials and specialty products commonly — though not universally — behave this way. They account for roughly 80% of what an organization spends, but only 10% of the time its procurement function spends buying.
A Dynamic Flux commodity shows continuing price movement, often with a wide floor-to-ceiling range. Indirect materials, and MRO in particular, frequently — again, not universally — behave this way. They are the mirror image: 15 to 20% of the spend, and 90% of the time.
The finding was not that every direct or indirect commodity behaves identically. It was that the purchasing model has to follow the commodity’s observed behavior rather than its accounting classification.
Both were published on this blog in 2007 and first documented in Acres of Diamonds, which received limited direct circulation in 2004 before its public release in 2005 — still available for download.
That is the whole objection in one sentence: Hackett selected indirect spend — a heterogeneous portfolio containing many Dynamic Flux commodities — and drew a consolidation-first direction from aggregate benchmark correlations before distinguishing which commodities benefited from concentration and which required continuing access to a dynamic market. Apply a Historic Flat Line purchasing model indiscriminately to Dynamic Flux commodities, and the apparent volume-discount savings can disappear quickly, sometimes immediately. It is why buyers say they can beat the centrally negotiated price with one phone call to a local supplier. In the cases I cited in 2014, a government department had paid on average more than 23% above the going market price for its MRO products, and a major American retailer experienced a similar outcome with its rationalization initiative. In both, the correction ran the opposite way to the prescription: MRO cost performance improved by widening access to the market, not by narrowing the supply base. In the government case, that strategic expansion was the 1998 RAM work.
The methodological difference, stated plainly
Hackett’s paper demonstrated correlation. Organizations with fewer indirect suppliers also showed lower purchased costs, fewer procurement staff, greater automation.
Correlation within a benchmark sample is real information. It does not establish that reducing supplier numbers produced those results. There are usually other unidentified conditions that could account for some or all of the savings, such as clean data, clear procurement authority, sound category segmentation, and disciplined exception handling. Supplier count may therefore have been an indicator of organizational maturity rather than the independent cause of improved performance.
A benchmark tells you what high-performing organizations have in common. It does not tell you what your organization should do next. If benchmarks alone were sufficient, the outcome record across past generations of technology would look very different than it does.
Hackett benchmarked the visible configuration. I was modeling the behavior underneath it. That difference is the entire distance between the two positions in 2012 — and it is the distance that has been closing ever since.
What changed between 2012 and 2026
In 2017 Hackett reported that nearly 85% of procurement organizations expected digital transformation to fundamentally change service delivery, while only 32% had a formal digital strategy and only 25% had the resources and competencies in place. The gap between intent and capability was now Hackett’s own finding, in its own numbers.
In 2021 the Digital World Class® research quantified what separated the top performers: 25% lower cost, 2.5 times the procurement ROI. Still an enablement story, but a measured one.
In 2025 readiness itself became the prerequisite — strategy, governance, culture, technology, data architecture and AI enablement identified as the conditions for scaling. I wrote about that research in November 2025.
Then, in its 2026 Procurement Key Issues Study published on 17 March, Hackett stated that simply activating vendor-provided AI features is a missed opportunity, and that leading teams first assess where AI can drive measurable value in their specific organizational context before choosing between embedded capabilities, AI-native solutions or custom development. Its senior research director put it this way: the focus is shifting to redesigning processes, roles and decision-making so AI delivers measurable value rather than incremental efficiency.
Context before technology. Process and roles before deployment.
That is Phase 0™ in spirit, arrived at independently, from benchmark data rather than from a 1998 engagement — if not yet in executable form.
But look again at that last sentence, because it is also where the convergence stops. Processes, roles and decision-making are to be redesigned so that AI can deliver measurable value. One reading is generous: the technology alone was never going to be enough, and the organization has to change with it. The other reading is the one I would flag. The organization is being fitted to the technology. The direction of travel still runs from the tool to the operating model.
That is an equation-based approach — adjust the variables until the model produces the expected output. Phase 0™ runs the other way — a people/AI-centric, agent-based approach. Establish how the organization actually operates first, then decide what technology, if any, fits it. It is the difference between redesigning the work so the tool succeeds and selecting the tool the work can absorb.
The ARA™ RAM 2025™ assessment
These scores assess the cited public Hackett materials against four defined architectural dimensions. They do not assess Hackett’s undisclosed client methods, internal capabilities or intent.
- Problem recognition: 9/10 — whether the conditions that decide the outcome are named at all. Hackett names them explicitly, and measures them.
- Individual operating principles: 7.5/10 — whether each condition is stated in a form an organization could act on.
- Integrated diagnostic method: 6/10 — whether they are joined into a single procedure that can be run on a live decision. Hackett’s readiness framework goes further here than most.
- Full architecture: 4/10 — readiness assessment, independent challenge, decision provenance and human authority, operating as one system.
Hackett has reached readiness-first. The open question is whether readiness can be measured against an operating reality the organization has not yet discovered.
Readiness-first is not yet operating-reality-first
A readiness framework that relies primarily on documented processes, declared roles, reported performance and stated decision rights assesses the organization largely as it understands itself.
The 1998 engagement — and others like Virginia’s eVA initiative, which is still delivering value today — knew all of those things too. Documented process, service levels, performance data — all available, all accurate.
What none of it contained, in the 1998 case, was that technicians were holding orders until late afternoon to protect one metric. Nobody had written it down, because nobody had any reason to think it mattered. A readiness assessment confined to declared processes, formal roles and reported performance could have scored that organization highly and still missed the condition governing the outcome.
So the three questions now sit in sequence:
- Gartner: can the organization govern what the technology does?
- Hackett: is the organization ready to realize value from it?
- Hansen: does the organization understand the real system against which governance and readiness must be assessed?
The first two are being answered in public, by firms with far greater reach than mine. The third is the constant — the question that has run from RAM 1998 to ARA™ RAM 2025™ — and it is the one that decides whether the first two land.
On the 2014 disagreement
I have not softened my view of that paper. Its correlations may hold within its sample; the limitation was turning an aggregate relationship into a general operating prescription for a category that behaves dynamically.
But the firm that published it in 2012 is the same firm telling its clients in 2026 to assess their specific organizational context before selecting an approach. That is not a retreat. That is a research organization following its own evidence, and it is worth saying out loud.
A benchmark tells you what high performers have in common. It does not tell you what your organization should do next.
On Friday I’ll show what the method looks like in practice. The lab is free: Friday, September 25 at 9:30 AM Eastern — https://www.linkedin.com/events/7504567838724997120/
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
-30-
Related
Assessing Hackett’s Convergence With the Hansen Models™, 2012–2026
Posted on September 23, 2026
0
Hackett has moved from aggregate prescription to readiness-first context. The remaining question is whether readiness is measured against the organization that exists on paper — or the one that actually places the order
In October 2010 I covered Chris Sawchuk’s eWorld keynote on procurement agility, and said something there I have not had cause to revise. I have rarely been a fan of analyst firms, because their perspectives had historically failed to make a dent in the e-procurement failure rate. The Hackett Group I treated as a different case — an operation whose coverage ran broader than the cozy arrangements between industry pundits and the vendors they were supposed to be covering.
Sixteen years later, that assessment holds. Hackett benchmarks what organizations actually do, publishes the numbers, and repeats the exercise year after year. Very few firms in this industry can say that.
Which is exactly why the disagreement is worth revisiting.
2012: the supplier consolidation paper
In 2012 Hackett published The Benefits of Supplier Consolidation Extend Far Beyond Sourcing Savings — document CR_6000128, in its Procurement Executive Insight series. It drew on more than ten years of Hackett procurement benchmark data. Its analysis was restricted to indirect spending. It argued three benefits: lower purchased costs, lower procurement and supplier-management costs, and reduced noncompliance with greater stakeholder satisfaction. Its foundational proposition was that companies eventually buy too many things from too many different suppliers, and its analytical direction was consolidation-first, with provision for approved exceptions.
On 6 February 2014 I wrote that the paper was surprising for all the wrong reasons.
Not because supplier consolidation is wrong. Consolidation makes sense when strategically applied — complex products requiring specialized ongoing support, where the field of capable suppliers is naturally limited anyway. My objection was to the category it had chosen.
Why indirect spend was the wrong ground
Over a fourteen-year monitoring period, eleven years of it after the characteristics were identified, two recurring commodity behaviors emerged.
A Historic Flat Line commodity shows relatively stable pricing over an extended period and a narrow floor-to-ceiling price range. Direct materials and specialty products commonly — though not universally — behave this way. They account for roughly 80% of what an organization spends, but only 10% of the time its procurement function spends buying.
A Dynamic Flux commodity shows continuing price movement, often with a wide floor-to-ceiling range. Indirect materials, and MRO in particular, frequently — again, not universally — behave this way. They are the mirror image: 15 to 20% of the spend, and 90% of the time.
The finding was not that every direct or indirect commodity behaves identically. It was that the purchasing model has to follow the commodity’s observed behavior rather than its accounting classification.
Both were published on this blog in 2007 and first documented in Acres of Diamonds, which received limited direct circulation in 2004 before its public release in 2005 — still available for download.
That is the whole objection in one sentence: Hackett selected indirect spend — a heterogeneous portfolio containing many Dynamic Flux commodities — and drew a consolidation-first direction from aggregate benchmark correlations before distinguishing which commodities benefited from concentration and which required continuing access to a dynamic market. Apply a Historic Flat Line purchasing model indiscriminately to Dynamic Flux commodities, and the apparent volume-discount savings can disappear quickly, sometimes immediately. It is why buyers say they can beat the centrally negotiated price with one phone call to a local supplier. In the cases I cited in 2014, a government department had paid on average more than 23% above the going market price for its MRO products, and a major American retailer experienced a similar outcome with its rationalization initiative. In both, the correction ran the opposite way to the prescription: MRO cost performance improved by widening access to the market, not by narrowing the supply base. In the government case, that strategic expansion was the 1998 RAM work.
The methodological difference, stated plainly
Hackett’s paper demonstrated correlation. Organizations with fewer indirect suppliers also showed lower purchased costs, fewer procurement staff, greater automation.
Correlation within a benchmark sample is real information. It does not establish that reducing supplier numbers produced those results. There are usually other unidentified conditions that could account for some or all of the savings, such as clean data, clear procurement authority, sound category segmentation, and disciplined exception handling. Supplier count may therefore have been an indicator of organizational maturity rather than the independent cause of improved performance.
A benchmark tells you what high-performing organizations have in common. It does not tell you what your organization should do next. If benchmarks alone were sufficient, the outcome record across past generations of technology would look very different than it does.
Hackett benchmarked the visible configuration. I was modeling the behavior underneath it. That difference is the entire distance between the two positions in 2012 — and it is the distance that has been closing ever since.
What changed between 2012 and 2026
In 2017 Hackett reported that nearly 85% of procurement organizations expected digital transformation to fundamentally change service delivery, while only 32% had a formal digital strategy and only 25% had the resources and competencies in place. The gap between intent and capability was now Hackett’s own finding, in its own numbers.
In 2021 the Digital World Class® research quantified what separated the top performers: 25% lower cost, 2.5 times the procurement ROI. Still an enablement story, but a measured one.
In 2025 readiness itself became the prerequisite — strategy, governance, culture, technology, data architecture and AI enablement identified as the conditions for scaling. I wrote about that research in November 2025.
Then, in its 2026 Procurement Key Issues Study published on 17 March, Hackett stated that simply activating vendor-provided AI features is a missed opportunity, and that leading teams first assess where AI can drive measurable value in their specific organizational context before choosing between embedded capabilities, AI-native solutions or custom development. Its senior research director put it this way: the focus is shifting to redesigning processes, roles and decision-making so AI delivers measurable value rather than incremental efficiency.
Context before technology. Process and roles before deployment.
That is Phase 0™ in spirit, arrived at independently, from benchmark data rather than from a 1998 engagement — if not yet in executable form.
But look again at that last sentence, because it is also where the convergence stops. Processes, roles and decision-making are to be redesigned so that AI can deliver measurable value. One reading is generous: the technology alone was never going to be enough, and the organization has to change with it. The other reading is the one I would flag. The organization is being fitted to the technology. The direction of travel still runs from the tool to the operating model.
That is an equation-based approach — adjust the variables until the model produces the expected output. Phase 0™ runs the other way — a people/AI-centric, agent-based approach. Establish how the organization actually operates first, then decide what technology, if any, fits it. It is the difference between redesigning the work so the tool succeeds and selecting the tool the work can absorb.
The ARA™ RAM 2025™ assessment
These scores assess the cited public Hackett materials against four defined architectural dimensions. They do not assess Hackett’s undisclosed client methods, internal capabilities or intent.
Hackett has reached readiness-first. The open question is whether readiness can be measured against an operating reality the organization has not yet discovered.
Readiness-first is not yet operating-reality-first
A readiness framework that relies primarily on documented processes, declared roles, reported performance and stated decision rights assesses the organization largely as it understands itself.
The 1998 engagement — and others like Virginia’s eVA initiative, which is still delivering value today — knew all of those things too. Documented process, service levels, performance data — all available, all accurate.
What none of it contained, in the 1998 case, was that technicians were holding orders until late afternoon to protect one metric. Nobody had written it down, because nobody had any reason to think it mattered. A readiness assessment confined to declared processes, formal roles and reported performance could have scored that organization highly and still missed the condition governing the outcome.
So the three questions now sit in sequence:
The first two are being answered in public, by firms with far greater reach than mine. The third is the constant — the question that has run from RAM 1998 to ARA™ RAM 2025™ — and it is the one that decides whether the first two land.
On the 2014 disagreement
I have not softened my view of that paper. Its correlations may hold within its sample; the limitation was turning an aggregate relationship into a general operating prescription for a category that behaves dynamically.
But the firm that published it in 2012 is the same firm telling its clients in 2026 to assess their specific organizational context before selecting an approach. That is not a retreat. That is a research organization following its own evidence, and it is worth saying out loud.
A benchmark tells you what high performers have in common. It does not tell you what your organization should do next.
On Friday I’ll show what the method looks like in practice. The lab is free: Friday, September 25 at 9:30 AM Eastern — https://www.linkedin.com/events/7504567838724997120/
Truth Is Believing. Accuracy Is Knowing. Outcome Is Proof.™
-30-
Share this:
Like this:
Related