A new quantitative systematic procedure, called a cost relevance tree, for group objective evaluation of development projects for their eventual cost-reduction effectiveness in a given segment of a business is described. The procedure is applied to the construction, repair, and rearrangement of Bell System underground (conduit) plant. This systematic procedure quantifies recent and applicable costs, projects these costs into the future, and finally provides a format of development project program rankings where future underground plant innovations can be identified and evaluated. This procedure is an extension of Honeywell's PATTERN technique attempting to add some economic dimension without undue complexity to the relative value of development projects. While similar to the Delphi method in the use of expert evaluation, attempts to achieve its controlled, but time-consuming, independent thinking were not made. Resultant data deficiencies plus a lack of benefit-cost analysis and the uncertainties of future projections are compensated by a sensitivity analysis. This analysis also shows the effect of input data uncertainty on project ranking.