There are obvious advantages to receiving a fixed price quote over a time and materials estimate. It is possible to elicit a fixed price quotation from a supplier in response to an invitation to tender (ITT). It just isn’t easy.
If a supplier is prepared to submit a fixed price quotation, it should mean you can have a high degree of confidence in the business specification, implying the supplier understands exactly what you want. Where the requirments are not well understood the supplier will instead offer you a time and materials estimate, which is no guarantee whatsoever of time, cost or quality; a kind of blank cheque.
In a recent procurement exercise where the PrinceLite method was used to draft the specification included in the tender document, one supplier offered a fixed price quotation, and two others offered a time and materials estimate with a fixed price option that was priced as a 20% uplift on the time and materials price.
One way of interpreting this calculation is to equate it with a 20% uncertaintly (risk). This could be read as “we think we can do the job for x, but we are sure we can do it for x + 20%”. This implies there was a 20% ambiguity in the business specification. Rather than building in an uncertainty factor, the preferred response would be to seek clarification and then commit to fixed price. If the uplift was in the order of 100% or more, this would indicate there was something seriously wrong with the business specification, and perhaps no quesitoning was likely to result in sufficient clarification. This may indicate that the difference between the estimate and the quotation is a sign of the quality of the requirements. It is highly suspect that any project should go forward where no fixed price quotations are forthcoming due to the impossibility of assessing the risk. In the current example however, those companies who took the variable approach were the ones who had either posed no questions, or asked only superficial ones in the submission stage.
The winning bidder quoted £85K, the second bidder quoted £145K, and the third bidder £195K. Set in this context, the cost of producing the upfront specification for tender is easily justified. The decision to award in this case was simple, because the unequivacally fixed price quotation was also the least expensive by a wide margin, and it is no surprise this supplier asked all the difficult questions in the submissions stage.
How it is possible to produce an unambiguous business requirements specification to include in a tender document in a short period? The answer is in the application of UML for project management which is defined as part of the PrinceLite method. This includes:
• End to end activity modelling
• Actor identification
• Primary use case modelling (graphical only)
• Elaboration use case modelling (graphical and ‘mid level’)
• Business object modelling (grammatical and lexical modelling)
• State modeling (featuring the ‘transacted’ class to derive business rules)
A PrinceLite informed business requirements specification for tenders is typically 50 pages long. It is orientated around models that are produced at a level of accuracy sufficient to the task of allowing accurate quotation by an experienced software supplier.
The customer must be willing to represent the requirements upfront where they believe it is possible, can be done in a timely manner, and will result in tangible benefits. Some companies may not have the need to bring the modelling expertise inhouse and so the drafting of the specification for the tender can be contracted out. Conversely, the skills required can be taught to the inhouse project managers and business analysts. The training typically takes one week. The production of the business requirements specificaton for procurement typically takes 2 weeks to produce and a futher week of corrections with the business and ‘power users’ to ensure its correctness. Suppliers typically require 2 weeks to respond to the invitation to tender. Therefore, with a degree of experience the cycle from ‘go ahead’ to receipt of tenders can be reduced to 5 weeks. These figures are, by necessity, approximate but they are nonetheless typical. They are based on transactionally-oriented database applications for business and depend to a certain respect on the functional size of the ultimate system.
It must be remembered that the rationale for going through the process of learning how to produce these specificatons for tender is to allow for the organisation to move away from time and materials estimates to fixed price quotations. Ultimately, the purpose is to control software development products more closely, to drive down the cost of acquisition, and to shorten the delivery lead time.
Time and materials estimates favour the supplier. They are an open invitation for costs to escalate. They provide no recourse when the project deadline slips. They are open to accusations from the supplier that the reason for delays is down to the ‘customer changing the requirements’. This is a charge to which the customer is highly vunerable. Ultimately, is a time and materials estimate worth the paper it is written on? It is unclear how the cost is calculated and it cannot be relied upon. Therefore the decision to award based on such an estimate is a fraud. The lowest estimate will win, but it may well escalate far in excess of even the most expensive of the losing competitors estimates. This is a built-in contradiction within the process, that no right minded procurement professional should ever consider. Unambiguous upfront requirements representation for procurement is possible. It just takes a bit more work.