The objection comes up in almost every conversation about a shared tendering platform. Bundling only pays off once all members join in. As long as some of the member companies stay out, the volume is too small and the effort too large.
The calculation shows the opposite.
What gets tendered is a category, not a total volume. Under a narrow definition, a member company with five million euros in revenue reaches around 55,000 euros per category per year. Twenty member companies together reach around 1.1 million.
The jump that matters lies between one participant and a double-digit number of participants. That is where an order no supplier prices carefully turns into an order several bid for seriously.
Whether forty or sixty member companies join after that changes the result gradually, but not fundamentally. The second participant matters more for the effect than the fiftieth.
Experience shows that projects in buying groups do not fail on economics. They fail on consent.
An approach that requires everyone's agreement needs a resolution that binds everyone. That is a different process from a project that starts with twenty member companies and grows. The first requires a decision of principle, the second a decision to participate.
An approach that starts with twenty fails less often. And it delivers something no resolution can deliver, namely a completed procedure the remaining members can observe before deciding.
Participation can be decided per category. A member that awards its energy supply jointly does not have to make the same decision for workwear.
That sounds obvious but is not in practice. In many conversations the platform is discussed as if it were a decision of principle about the independence of the member company. It is a sequence of individual decisions, each tied to a specific category and a specific procedure.
For the central office this has a practical consequence. Getting started works better with two or three categories where the specification is uncontroversial than with a complete switch. A member that took part in workwear and knows the procedure decides differently about the next category.
One qualification belongs here. There is a size below which a member's participation worsens the group's economics rather than improving them.
The usual pricing model in this market segment has two parts, a base fee for the platform and a fee per connected unit. Approximately, then, the minimum volume per unit equals the annual fee per unit divided by the effect factor.
A fee of 3,000 euros a year per connected unit requires, under mid-range assumptions, a bundleable volume of around 100,000 euros per unit, and under conservative assumptions around 200,000 euros. By the calculation above, that corresponds to a retail company from around one to two million euros in annual revenue.
Two things follow. Very small members do not carry their own fee. And a linear fee per unit works against exactly the mechanism that matters, because it prices every additional participant the same although the twentieth adds less than the second. Anyone who wants to generate participation should check whether the chosen pricing model rewards it or taxes it. Tiered and capped models change the picture considerably.
How many member companies do we need to start? A double-digit number in the respective category is the orientation, not a fixed threshold. What matters is the total per category, and one large member can replace several small ones.
What about the member companies that do not join? They decide again with the next category. In practice that is the most common way participation grows, not through persuasion up front, but through a completed procedure whose result is visible.
Do our members lose independence? In demand bundling, the member decides per category whether to participate and keeps its own demand. The separate question is who sees which data inside the system. That is a configuration question, and it belongs before the rollout, not after.
Section 6.6 of the study addresses the participation question, and chapter 8 covers economics at small participant numbers, including the minimum size per unit and the effect of different pricing models.
→ To the white paper
The calculation shows the opposite.
Where the jump actually lies
What gets tendered is a category, not a total volume. Under a narrow definition, a member company with five million euros in revenue reaches around 55,000 euros per category per year. Twenty member companies together reach around 1.1 million.
The jump that matters lies between one participant and a double-digit number of participants. That is where an order no supplier prices carefully turns into an order several bid for seriously.
Whether forty or sixty member companies join after that changes the result gradually, but not fundamentally. The second participant matters more for the effect than the fiftieth.
Why this is the decisive point in practice
Experience shows that projects in buying groups do not fail on economics. They fail on consent.
An approach that requires everyone's agreement needs a resolution that binds everyone. That is a different process from a project that starts with twenty member companies and grows. The first requires a decision of principle, the second a decision to participate.
An approach that starts with twenty fails less often. And it delivers something no resolution can deliver, namely a completed procedure the remaining members can observe before deciding.
The second relief
Participation can be decided per category. A member that awards its energy supply jointly does not have to make the same decision for workwear.
That sounds obvious but is not in practice. In many conversations the platform is discussed as if it were a decision of principle about the independence of the member company. It is a sequence of individual decisions, each tied to a specific category and a specific procedure.
For the central office this has a practical consequence. Getting started works better with two or three categories where the specification is uncontroversial than with a complete switch. A member that took part in workwear and knows the procedure decides differently about the next category.
Where the lower limit lies
One qualification belongs here. There is a size below which a member's participation worsens the group's economics rather than improving them.
The usual pricing model in this market segment has two parts, a base fee for the platform and a fee per connected unit. Approximately, then, the minimum volume per unit equals the annual fee per unit divided by the effect factor.
A fee of 3,000 euros a year per connected unit requires, under mid-range assumptions, a bundleable volume of around 100,000 euros per unit, and under conservative assumptions around 200,000 euros. By the calculation above, that corresponds to a retail company from around one to two million euros in annual revenue.
Two things follow. Very small members do not carry their own fee. And a linear fee per unit works against exactly the mechanism that matters, because it prices every additional participant the same although the twentieth adds less than the second. Anyone who wants to generate participation should check whether the chosen pricing model rewards it or taxes it. Tiered and capped models change the picture considerably.
Frequently asked questions
How many member companies do we need to start? A double-digit number in the respective category is the orientation, not a fixed threshold. What matters is the total per category, and one large member can replace several small ones.
What about the member companies that do not join? They decide again with the next category. In practice that is the most common way participation grows, not through persuasion up front, but through a completed procedure whose result is visible.
Do our members lose independence? In demand bundling, the member decides per category whether to participate and keeps its own demand. The separate question is who sees which data inside the system. That is a configuration question, and it belongs before the rollout, not after.
The next step
Section 6.6 of the study addresses the participation question, and chapter 8 covers economics at small participant numbers, including the minimum size per unit and the effect of different pricing models.
→ To the white paper