History: Parametric Cost Estimating generations

Parametric cost estimating has been created and matured through a number of system generations. What were those generations and how mature are we now?

Generation 1.0 - Mainframe

Commercial parametric models began with first generation models adopting the earliest mainframe computers. Prior to the adoption of computers, using graph paper, a pencil and a slide rule it was possible to derive parametric models, but the distribution of that knowledge for commercial use was difficult unless published in a book.

With the adoption of computers in estimating it was possible to make a modest material gain from the licensing of commercial parametric models.

This support allowed further research and more importantly effortless sharing of parametric solutions across the aerospace and defense community. The life time of these early parametric models spanned from 1975 to 1990.

Early Parametricians needed to be taught text editor and keyboard skills as part of the commercial parametric courses. Cost Engineers of the 1970s were still marvelling at the electronic calculator when the cream of the youngest, most talented Cost Engineers had their productivity slowed to a grinding halt by one finger and a keyboard.

Estimates were prepared and processed overnight on the mainframe, only to find in the morning that a vital input was missing and the whole estimate would need to be run again the next night.

However, intellectual property in the form of the trade secret algorithms contained within commercial models were deemed sufficiently valuable that management and Cost Engineer were prepared to accept mainframe computers as an imperfect solution.

Generation 2.0 - Personal Computers

Second generation parametric models began to arrive with the embracing of personal computers (PC) in industry in 1988 and are still upgraded and used today. As personal computers became powerful enough to deal with the thousands of mathematical calculations required in the parametric models, the software of the first generation models were ported onto these computer platforms. The second generation models also had the advantage of Graphical User Interfaces (GUI) and easier dialogue with the user. This led to the more efficient use of commercial parametric models and the further expansion of their application.

The commercial parametric model became portable and could be taken to subcontractor or customer sites for negotiations.

Whereas the processing of the first generation parametric models was done centrally, using dumb terminals via 1,200 bits per second or 2.4 kbps modems to retrieve the results. The second generation commercial parametric models had to deal with the issue of protecting the trade secret equations within the programming.

The software was compiled to run faster on personal computers, which led to the trade secret equations within the programming becoming scrambled – to deal with the issue of protection – which in term led to the problem of validation and verification.

The term ‘black-box’ was commonly used by staff who found commercial parametric models unfathomable.

In these second generation models the users were ultimately unable to see the mathematics or parametric algorithms; they had to depend on the quality assurance of the commercial parametric models vender.

Generation 2.5 - Knowledge Based

The second generation models developed a Knowledge Base to create a means to safely store known data and therefore contribute justification of the values being used in the parametric analysis.

The simple access to creditable, factual, historical data which is impossible to challenge and results in more confidence in a parametric prediction. This approach started to store the cost data in a central shared database with group access to the cost teams to contribute to when verified and utilise in the prediction.

Generation 3.0 - Client/server computer

Finally, at the turn of the millennium, third generation parametric models arrived on the market. Taking advantage of the latest client/server computer architectures combined with open parametric estimating frameworks. Third generation parametric models enable commercial parametric models vendors to port their models into this framework and implement bespoke organization specific models. Within the open parametric framework these models (from different origins) can use the common cost engineering features of the framework like labour rates, escalation, risk analysis, reporting and so forth.

Cost savings on third generation models are possible because the same user interface is used for all the models whether software, IT or hardware. Training can be modular, with framework training conducted only once followed by model, specific training, providing a more pleasurable user experience.

Furthermore, validation and verification of the models is simpler as the open environment can be scrutinised by any authority to review the implementation of the cost estimating relationships.

Generation X - What is next?

What is the next generation of creating cost, schedule and risk impact estimates? There is still a demand for cost estimating in Defence and Security, but what is the next approach? Can there be an IA source? Is there a Commercial solution or will individuals generate solutions in their organisation?

Take a look at this article regarding Forecasting the Future of cost Forecasting

For more knowledge and appreciation of the parametric history read the 'Systems Cost Engineering' book. This article is based upon Chapter 23 - The History of Parametrics.

https://www.routledge.com/Systems-Cost-Engineering-Program-Affordability-Management-and-Cost-Control/Shermon/p/book/9781138253865