In practice it is often difficult to isolate the contributions of an intervention to the volume of offences committed. For example, it might be that a new forensic technique increases the number of offenders being identified and charged with an offence. This does not measure directly the number of offences prevented but does enable the analyst to compare the relative effectiveness of the new technique compared with traditional practice. If the costs of the two ‘techniques’ are known then it is possible to compare directly the cost per unit of achieving an arrest using the two methods. The ‘cost effective’ method can then be identified as the one delivering an arrest at lower cost: see here for further analysis of this application of cost effectiveness to the choice of forensic technique.
For a practical guide see the Home Office toolkit. This tool is very much in line with the Dhiri and Brand (1999) Guidance for Crime Reduction Programme Evaluation.

