During the early and mid 1980s, the United States Postal Service (USPS) was faced with the decision whether to extend postal automation, and if so, by what means. The Office of Technology Assessment (OTA) was asked to investigate the advisability of proceeding with Phase 2 of the postal automation strategy on both technical and economic grounds. The USPS's strategy was judged to be technically feasible, although a technology other than the USPS's choice was deemed a worthy alternative.
The decision ultimately centered on which technology would perform the best in terms of economic savings. The USPS's choice was for single-line optical character readers, while the proposed alternative was for a multi-line reader which could convert a full address that used only the standard five-digit Zip code to Zip + 4. Both types of reader then place a bar code on the letter, and the bar code is read by an automatic sorter. The advantage of Zip + 4 is that the automatic sorter can then sort the mail to the level of a carrier route, rather than a post office or postal zone. This question was further complicated by uncertainty about the use of ZIP + 4 by consumers. Historically, the USPS had overestimated consumer use of its innovations.
To address this complex and uncertain situation, Decision Science Consortium was contracted to perform a decision analysis of postal automation alternatives. A complex decision tree with six decision branches was developed. Each decision branch except the one for canceling the automation altogether was subjected to a probabilistic analysis of three factors: rate of Zip + 4 usage, savings percentage factor, and usage savings factor. Next, for each path in the decision tree, a detailed cash-flow analysis was developed in order to compare the outcomes of the various alternatives.
The results of this analysis indicated that the NPVs of the five alternatives that would continue the postal automation ranged between $900 million and $1.5 billion. On an expected value basis, all of these options were preferable to canceling the automation, and the option to convert from single-line to multi-line optical character readers was the optimal decision.
Finally, sensitivity analyses were performed to consider the uncertainty in the evaluations. Based on these analyses, the following conclusions were reached:
- Any continuation of postal automation was better than canceling.
- Converting to multi-line optical character readers was preferred.
- Uncertainty about the cost of the multi-line readers contributed very little to the uncertainty of its NPV.
- Uncertainty about Zip + 4 usage contributed the most to variations in NPV.
Thus, the USPS's main arguments against the use of multi-line readers, that their price and performance were uncertain, were found to be insignificant when compared with other factors, particularly the uncertainty of rate of Zip + 4 usage. This analysis formed the basis of the OTA's report and recommendations to Congress, and the decision was made to convert to multi-line readers. The savings to the USPS (and US taxpayers) was estimated to be $1.5 billion, some $200 million more than had the USPS's first choice been employed.