In the debate about joint procurement the attention sits on bundling. Who bundles what, in which category, at what volume. The second question is rarely asked, although it co-determines whether the project is workable.
Joint purchasing and joint information flow are not the same thing.
This article describes an economic mechanism and names the rules within which it has to be assessed. It contains no legal advice and replaces none.
Bundling demand is a deliberate decision. It is taken by the participants, it is the purpose of the cooperation, and its effect is foreseeable.
Information flow, by contrast, arises as a technical by-product. As soon as several legally independent companies use the same platform, demand, volumes, prices and supplier assignments sit in one shared system. Who sees what is a configuration question.
That question often gets asked only once the system is already running. At that point answering it is considerably more laborious than before.
In competition economics, tacit collusion describes an alignment of market behaviour without any agreement. It requires three conditions. The participants meet repeatedly, they can observe each other's behaviour, and they can react to it.
For changing suppliers in indirect spend these conditions regularly do not hold. Between the member companies of a buying group the situation is different. They are permanently part of the same organization, they serve neighbouring or overlapping sales territories, and a shared platform makes the behaviour of the others visible.
More becomes visible than it first appears. Purchasing data allows conclusions about sales. The volume purchased shows the scale of the business. The breadth of the assortment shows its orientation. The purchase price, combined with an observed retail price, allows an estimate of the margin calculation.
None of these figures is the subject of a tender. All of them arise in one.
The point of this section is not that anyone is doing something impermissible. The effect occurs precisely when nobody intends it. That is exactly why it is a question of technical design and not of good will. A platform that limits visibility from the outset ensures the question does not arise.
The following rules form the framework. They are named here so you know what to look for and whom to ask. Naming them does not imply that any of them applies in your case, and the list makes no claim to completeness.
The cartel prohibition under section 1 of the German Act against Restraints of Competition and Article 101 TFEU covers not only agreements but also concerted practices. Section 3 of the same Act contains the provision on cooperation between small and medium-sized enterprises, on which mid-market buying cooperations typically rely. The European Commission's 2023 Horizontal Guidelines are relevant in two chapters, the one on purchasing agreements and the one on information exchange. For buying groups organized as cooperatives, the German Federal Cartel Office's guidelines on the compatibility of cooperative structures with competition law are the obvious source.
On market share thresholds we deliberately state no figure. Values circulate whose origin cannot always be traced, and applying them presupposes a definition of the relevant market. That definition is itself a legal question and in individual cases the harder part of the assessment. A number quoted without it does not help, it misleads.
Which requirements apply in a specific case depends on the legal form of the cooperation, on the markets concerned and on the participants' shares in those markets. Only a lawyer practising in competition law can make that assessment. It should exist before a shared platform is introduced, not afterwards, because its outcome determines the configuration.
Regardless of how the legal assessment turns out in a specific case, an organization has to answer the following questions, because they determine how the system is set up. They are phrased as questions, not as requirements.
Who sees an individual member's demand volume? Only the member itself, the central office as well, or all participants?
Who sees which prices? A distinction is needed between bid prices during a live procedure, the awarded price, and historical prices of completed procedures.
Who sees which member buys from which supplier?
Is data aggregated before display? From what number of members onward? And can the aggregate be traced back to the individual member?
At what point does information become visible? At the same time as the procedure, or with a time lag?
Who sets these visibility rights, who can change them, and is a change recorded?
Can it be shown afterwards which setting applied at a given point in time?
The seventh question is the one most often overlooked. A separation that is in place but cannot be evidenced does not help at the moment somebody asks about it.
Does this affect us at all? The question arises in full sharpness where members compete with each other in their local markets. Where sales territories do not overlap, the situation is different. The assessment still belongs before implementation.
Isn't this solved by access rights in any system? Role-based rights are the precondition, not the solution. What matters is whether the separation can be configured along the seven questions above and evidenced afterwards.
How laborious is the clarification? Manageable, and considerably smaller than the effort of changing a running configuration after the fact. In the tool selection of one of Europe's largest retail cooperations, cusoso was the only one of twenty vendors examined that could deliver the required separation.
Chapter 7 of the study treats both questions separately, names the legal framework in full and contains the question catalogue with commentary.
→ To the white paper
Joint purchasing and joint information flow are not the same thing.
This article describes an economic mechanism and names the rules within which it has to be assessed. It contains no legal advice and replaces none.
Two questions that regularly get conflated
Bundling demand is a deliberate decision. It is taken by the participants, it is the purpose of the cooperation, and its effect is foreseeable.
Information flow, by contrast, arises as a technical by-product. As soon as several legally independent companies use the same platform, demand, volumes, prices and supplier assignments sit in one shared system. Who sees what is a configuration question.
That question often gets asked only once the system is already running. At that point answering it is considerably more laborious than before.
Why visibility alone changes behaviour
In competition economics, tacit collusion describes an alignment of market behaviour without any agreement. It requires three conditions. The participants meet repeatedly, they can observe each other's behaviour, and they can react to it.
For changing suppliers in indirect spend these conditions regularly do not hold. Between the member companies of a buying group the situation is different. They are permanently part of the same organization, they serve neighbouring or overlapping sales territories, and a shared platform makes the behaviour of the others visible.
More becomes visible than it first appears. Purchasing data allows conclusions about sales. The volume purchased shows the scale of the business. The breadth of the assortment shows its orientation. The purchase price, combined with an observed retail price, allows an estimate of the margin calculation.
None of these figures is the subject of a tender. All of them arise in one.
The point of this section is not that anyone is doing something impermissible. The effect occurs precisely when nobody intends it. That is exactly why it is a question of technical design and not of good will. A platform that limits visibility from the outset ensures the question does not arise.
The framework for assessing this
The following rules form the framework. They are named here so you know what to look for and whom to ask. Naming them does not imply that any of them applies in your case, and the list makes no claim to completeness.
The cartel prohibition under section 1 of the German Act against Restraints of Competition and Article 101 TFEU covers not only agreements but also concerted practices. Section 3 of the same Act contains the provision on cooperation between small and medium-sized enterprises, on which mid-market buying cooperations typically rely. The European Commission's 2023 Horizontal Guidelines are relevant in two chapters, the one on purchasing agreements and the one on information exchange. For buying groups organized as cooperatives, the German Federal Cartel Office's guidelines on the compatibility of cooperative structures with competition law are the obvious source.
On market share thresholds we deliberately state no figure. Values circulate whose origin cannot always be traced, and applying them presupposes a definition of the relevant market. That definition is itself a legal question and in individual cases the harder part of the assessment. A number quoted without it does not help, it misleads.
Which requirements apply in a specific case depends on the legal form of the cooperation, on the markets concerned and on the participants' shares in those markets. Only a lawyer practising in competition law can make that assessment. It should exist before a shared platform is introduced, not afterwards, because its outcome determines the configuration.
Seven questions before implementation
Regardless of how the legal assessment turns out in a specific case, an organization has to answer the following questions, because they determine how the system is set up. They are phrased as questions, not as requirements.
Who sees an individual member's demand volume? Only the member itself, the central office as well, or all participants?
Who sees which prices? A distinction is needed between bid prices during a live procedure, the awarded price, and historical prices of completed procedures.
Who sees which member buys from which supplier?
Is data aggregated before display? From what number of members onward? And can the aggregate be traced back to the individual member?
At what point does information become visible? At the same time as the procedure, or with a time lag?
Who sets these visibility rights, who can change them, and is a change recorded?
Can it be shown afterwards which setting applied at a given point in time?
The seventh question is the one most often overlooked. A separation that is in place but cannot be evidenced does not help at the moment somebody asks about it.
Frequently asked questions
Does this affect us at all? The question arises in full sharpness where members compete with each other in their local markets. Where sales territories do not overlap, the situation is different. The assessment still belongs before implementation.
Isn't this solved by access rights in any system? Role-based rights are the precondition, not the solution. What matters is whether the separation can be configured along the seven questions above and evidenced afterwards.
How laborious is the clarification? Manageable, and considerably smaller than the effort of changing a running configuration after the fact. In the tool selection of one of Europe's largest retail cooperations, cusoso was the only one of twenty vendors examined that could deliver the required separation.
The next step
Chapter 7 of the study treats both questions separately, names the legal framework in full and contains the question catalogue with commentary.
→ To the white paper