In business cases for procurement software the saving is regularly applied to the entire tendered volume. Ten million euros of volume, five percent savings, half a million euros of benefit, set against the annual cost of the tool.
This assumes that without the tool nothing would have been tendered at all.
As a rule that is not the case. Part of the volume already runs through structured procedures today, and the savings achieved there would exist without the tool as well.
The additional effect of a tool consists of two parts. The saving on volume that newly enters competition, and an improvement on the volume that would have been tendered anyway.
Only the sum of the two is the effect that may be attributed to the tool.
At a current tender rate of forty percent, the naive calculation overstates that additional effect by about half. Under mid-range assumptions the effect factor is 2.9 percent of bundleable volume. If the saving is applied to the entire volume, it becomes 4.5 percent. The overstatement is 55 percent.
The same test we apply to other parties' savings claims applies here in our own case.
Four positions, of which business cases regularly show only the first.
The licence. The usual model in this market segment has two parts, a base fee for the platform and a fee per connected unit.
Implementation. Master data, category structure, integration with existing systems, training. These costs arise once and are spread over the expected useful life for the annual view.
The central office's ongoing effort. Running procedures, maintaining supplier and article data, supporting members. This item is most often left out, although in small organizations it can exceed the licence.
The members' ongoing effort. Reporting demand, coordination, approvals. Small per unit, not negligible in total.
The calculation can be done without specific vendor prices. Pricing models differ, and they change. What does not change is the structure.
The effect factor states how much additional effect each euro of bundleable volume produces. The required volume follows from annual cost divided by that factor.
For three sets of assumptions the picture is as follows. Conservative, with a current tender rate of 60 percent, a rate with the tool of 90 percent, a 4.0 percent price reduction on newly tendered volume and 0.5 percent additional reduction on already tendered volume, the effect factor is 1.5 percent. The required volume is then around sixty-seven times annual cost.
Under mid-range assumptions, with 40 and 90 percent, 5.0 and 1.0 percent, it is 2.9 percent and therefore around thirty-four times. Under favourable assumptions, with 20 and 95 percent, 5.0 and 1.5 percent, it is 4.1 percent and around twenty-five times.
Across all three sets, the required volume sits at twenty-five to sixty-seven times annual total cost. This order of magnitude is deliberately rough and replaces no calculation. It helps where nothing is being calculated at all so far.
A central office does not calculate for itself. It calculates in front of a governing body that will check the number, and for members who are independent entrepreneurs.
A calculation that overstates the additional effect by half will be noticed by the second year at the latest. At that point it is not the software that is up for discussion, but the central office's credibility in the next investment decision.
The most sensitive input in the whole calculation is the current tender rate. At 60 instead of 40 percent, the effect factor under mid-range assumptions falls from 2.9 to 2.1 percent, and the required volume rises from thirty-four times to forty-eight times. If you can establish only one number properly, establish this one.
How do we determine our current tender rate? Through the categories. Which were awarded in structured form last year, which were renewed, which were not looked at? That is a list, not a data analysis.
Can't we include process costs after all? Anyone who wants to will find the starting values in the study. We do not carry them in the result calculation, because saved working time only changes the annual result if it leads to lower personnel costs, and because freed capacity goes into additional tenders whose effect is already captured as coverage.
What about the benefit from evidence and information separation? Set at zero in these calculations. The reason is the same one for which we do not accept claims without a stated reference basis. An expected value for an event with low probability and very high damage would not be a calculation, it would be a number that produces the desired result.
Chapter 8 of the study contains the full break-even calculation, appendix A.2.3 all three sets of assumptions with the calculation path, and appendix A.1 every assumption with its effect should the value applied not hold.
→ To the white paper
This assumes that without the tool nothing would have been tendered at all.
As a rule that is not the case. Part of the volume already runs through structured procedures today, and the savings achieved there would exist without the tool as well.
The size of the error
The additional effect of a tool consists of two parts. The saving on volume that newly enters competition, and an improvement on the volume that would have been tendered anyway.
Only the sum of the two is the effect that may be attributed to the tool.
At a current tender rate of forty percent, the naive calculation overstates that additional effect by about half. Under mid-range assumptions the effect factor is 2.9 percent of bundleable volume. If the saving is applied to the entire volume, it becomes 4.5 percent. The overstatement is 55 percent.
The same test we apply to other parties' savings claims applies here in our own case.
What belongs on the cost side
Four positions, of which business cases regularly show only the first.
The licence. The usual model in this market segment has two parts, a base fee for the platform and a fee per connected unit.
Implementation. Master data, category structure, integration with existing systems, training. These costs arise once and are spread over the expected useful life for the annual view.
The central office's ongoing effort. Running procedures, maintaining supplier and article data, supporting members. This item is most often left out, although in small organizations it can exceed the licence.
The members' ongoing effort. Reporting demand, coordination, approvals. Small per unit, not negligible in total.
The structure of the break-even calculation
The calculation can be done without specific vendor prices. Pricing models differ, and they change. What does not change is the structure.
The effect factor states how much additional effect each euro of bundleable volume produces. The required volume follows from annual cost divided by that factor.
For three sets of assumptions the picture is as follows. Conservative, with a current tender rate of 60 percent, a rate with the tool of 90 percent, a 4.0 percent price reduction on newly tendered volume and 0.5 percent additional reduction on already tendered volume, the effect factor is 1.5 percent. The required volume is then around sixty-seven times annual cost.
Under mid-range assumptions, with 40 and 90 percent, 5.0 and 1.0 percent, it is 2.9 percent and therefore around thirty-four times. Under favourable assumptions, with 20 and 95 percent, 5.0 and 1.5 percent, it is 4.1 percent and around twenty-five times.
Across all three sets, the required volume sits at twenty-five to sixty-seven times annual total cost. This order of magnitude is deliberately rough and replaces no calculation. It helps where nothing is being calculated at all so far.
Why this counts more for a central office than elsewhere
A central office does not calculate for itself. It calculates in front of a governing body that will check the number, and for members who are independent entrepreneurs.
A calculation that overstates the additional effect by half will be noticed by the second year at the latest. At that point it is not the software that is up for discussion, but the central office's credibility in the next investment decision.
The most sensitive input in the whole calculation is the current tender rate. At 60 instead of 40 percent, the effect factor under mid-range assumptions falls from 2.9 to 2.1 percent, and the required volume rises from thirty-four times to forty-eight times. If you can establish only one number properly, establish this one.
Frequently asked questions
How do we determine our current tender rate? Through the categories. Which were awarded in structured form last year, which were renewed, which were not looked at? That is a list, not a data analysis.
Can't we include process costs after all? Anyone who wants to will find the starting values in the study. We do not carry them in the result calculation, because saved working time only changes the annual result if it leads to lower personnel costs, and because freed capacity goes into additional tenders whose effect is already captured as coverage.
What about the benefit from evidence and information separation? Set at zero in these calculations. The reason is the same one for which we do not accept claims without a stated reference basis. An expected value for an event with low probability and very high damage would not be a calculation, it would be a number that produces the desired result.
The next step
Chapter 8 of the study contains the full break-even calculation, appendix A.2.3 all three sets of assumptions with the calculation path, and appendix A.1 every assumption with its effect should the value applied not hold.
→ To the white paper