Home > Articles > Software Development & Management

  • Print
  • + Share This
This chapter is from the book

Summary and Conclusions

"Those who foretell the future lie, even if they tell the truth."—Arabian Proverb

eSell Inc. installed a metrics program based on functional metrics to insert predictability—and therefore confidence—into its sales and implementation processes. From that perspective, the program was an unqualified success. After two years of effort to create and refine the tools and techniques required implementing the program, the following observations can be made:

  • Customers appreciate the dialog that has to take place to produce an estimate for the project. It gives them a sense of comfort to be explicitly asked to walk through the project's requirements from their customer's perspective and to participate in the process of visualizing how the requirements will be implemented.

  • Because customers appreciate the process, the sales organization has embraced it as an integral part of their sales process.

  • Because both the sales organization and their prospective customers have embraced the program, the implementation teams do not have to be seen as the "bad guys" when recommending the use of a rational metrics-based process for project estimation and control.

  • Consequently, overall implementations have proceeded smoothly and have produced a highly referenceable customer base that can be leveraged for sales to new customers.

As the above referenced Arabian proverb suggests, the authors believe that the value of the estimation process is not in the accuracy of its predications but in the creation of attainable targets in which the customer and their customer are intimately engaged with the project teams. These attainable targets produce the self-fulfilling prophecy of project completion on time and within budget.

Some unexpected lessons were also learned and helped evolve the process. First, the bootstrapping of the measurement process in an organization unaccustomed to such mental models is slow-going and requires constant attention. For the process to be truly adopted and embraced, each group in the organization had to come to its own conclusions about the value of the extra effort, and each had to be "evangelized" separately and on an ongoing basis. The good news is that such adoption, thus hard-won, is strong and somewhat self-sustaining. Second, in practice it was a challenge to keep a balance between overly detailed and overly general ability-to functional descriptions during the estimation process. This required several consistency and normalization checks on the project managers who were generating the first draft of these lists. Third, at times some of the various functionalities described in separate ability-to statements were too related and intertwined to be structured in a completely self-sufficient and independent manner. Thus there were times when customers could not have individual line-item veto. Finally, for the same reason of threads of dependence between ability-to statements, it was a constant challenge not to double-count elementary processes and thus over-inflate the estimates. Avoiding this took attention to detail, practice, a dedicated team, and documented, consistent rules.

  • + Share This
  • 🔖 Save To Your Account