The design of woven fabrics has remained largely unchanged for generations - and so have the problems. But installing design software to automate the process is a big decision.

Traditionally, a designer's written definition is used to produce a woven sample. The designer examines the sample, identifies improvements and revises the specification. A new sample is then woven, and the cycle is repeated until the designer is happy with the results.

All this is time consuming, error prone and expensive. Weaving of samples must await the availability of a loom and may require the interruption of a production run. Transcription errors when specifications are revised can mean an entire cycle is wasted.

ScotWeave is a software system that addresses these problems, enabling a designer to create a new design and immediately see its appearance on screen. A quality printer can capture the final appearance of the woven fabric. Designers and fabric buyers can work directly with these print-outs, and the need for woven samples is virtually eliminated.

The Constructive Evaluation method was used to provide a predictive evaluation of the impact of ScotWeave at a typical customer employing several fabric designers. This was to support a decision on whether to buy ScotWeave.

For this evaluation, the quality-by-design chain has four levels: the customer has strategic objectives, that could be addressed by an intervention (in this case a project to install ScotWeave), that will change the characteristics of the fabric design process, by using the ScotWeave system.

At these four levels the important stakeholders include, respectively, the customer's financial director, operations manager, chief fabric designer, and head of purchasing.

At the strategic objectives level the key characteristics include lower costs and improved service to fabric buyers.

At the intervention level they include the elimination of woven sample costs and acceptability to fabric designers.

At the fabric design process level the characteristics include the use of paper samples, the acceptability of these to designers and fabric buyers, and the retention of familiar design concepts.

At the ScotWeave level the characteristics include the realism, accuracy and colour calibration of the paper samples, support for the entire design process, and support for rapid experimentation.

Part of the quality-by-design spreadsheet for the intervention and fabric design process levels is shown in Figure 1. Figure 2 shows the quality profile for the intervention level.

ScotWeave is well established and runs under Windows, so the risks are low. Even so, there is a possible risk at the design process level, for example, that the paper samples may not be acceptable to fabric buyers, even though they are realistic and accurate. An appropriate avoidance measure would be to show some illustrative paper samples and their woven equivalents to key buyers before committing to buying ScotWeave.

The results of the evaluation are largely given by the quality profile for the strategic objectives level, which shows projected annual savings in operating costs of £73,000 (worst case); a projected total cost of introducing the system of £41,500 (worst case); and a dramatic reduction in the average time taken to respond to buyer requests for changes to a sample: from more than 500 elapsed hours to 24.

The Constructive Evaluation report packages the quality-by-design spreadsheets, the quality profile, and the risk register. How might the report be used to support the decision on whether to buy ScotWeave?

The key decision maker might be the customer's managing director. The strategic objectives profile showing that ScotWeave pays for itself in seven months is pretty convincing, and the managing director may take the decision in the confidence that this profile has emerged from a systematic evaluation.

However, a more cynical managing director might suspect that the figures are too good to be true, and might require supporting evidence for the promised cost savings.

Here, the quality-by-design spreadsheets can be used to trace the sources of the savings through the various levels. These first show how the savings result from virtually eliminating woven samples, and then show the features of ScotWeave that makes such elimination possible. The accompanying quality profiles provide the actual figures, both for the current manual design process and for the new computer-based process, used in calculating the savings.

The managing director can thus trace through the entire chain of argument and check that it is sound and convincing. Any residual worries about hidden risks may be eased by consulting the risk register.

Suppose that another stakeholder, the operations manager, is happy with the case for a system but needs to be convinced on the choice of ScotWeave in particular.

The approach here is, again, to trace through the quality-by-design spreadsheets, but this time starting at the ScotWeave level. Tracing in this direction shows how various characteristics of ScotWeave are central to its acceptability to fabric designers and buyers, and ultimately to addressing the strategic objectives. A rival product that lacks some of the key characteristics would not be a suitable alternative.

In practice, the evaluation very clearly shows that the case for ScotWeave is compelling. The more that decision makers study and discuss the evaluation report, the more convinced they become.