Operationalizing Decision Yield - measuring performance over time
A number of previous examples introduced Decision Yield, discussed how to develop questions to measure it and then how to operationalize it in a performance audit. If your concern with a decision is to ensure that you maintain or improve the way you make a decision then measuring decision yield over time is a great way to do this. To track decision yield like this you will need to assign numeric values to the 5 dimensions to weight them so you can combine them into an overall score. You could track all five dimensions independently but it may be easier to bring a single combined score into your performance management environment and show management its current value, its trends etc in whatever format you use for key performance indicators.. You can then assess each of the 5 dimensions on some regular basis to track the values and convert them to an indicator using these weights. For example: the weekly performance of an insurance underwriting system could be tracked using the Decision Yield of the "underwrite new policy" decision.
You would likely use performance management or business intelligence tools to collect data to answer many of your questions (the more quantitative ones like number of quotes issued, number of policies written etc) and some kind of online questionnaire to capture qualitative information like typical response of prospects to the price quoted. Plugging these answers into a formula will allow you to generate a number that you can then track as a graph. Such a graph might show how a new website impacted Decision Yield by reducing consistency or how a change in the regulations led to much higher rates of manual review and thus higher cost and lower Decision Yield. The graph will give you early warning of problems as you evolve and adapt the decisioning process that you use.
You will need to consider different perspectives from different groups also. Below is an example of the kind of scorecard you can develop – allowing each group of stakeholders to weight each of the five dimensions. You typically measure a dimension in terms of how close to "best practice" you are and so this gives you a score of something from 0 to 1.00 where 1.00 is as good as it gets at present – remember that your score on a dimension should decrease if you do nothing. Applying the weights from your stakeholders give you a single number – something you can track on a dashboard as a Key Performance Indicator or KPI.
A graph showing how the calculated Decision Yield value can be tracked over time to show rises, when the decision yield improves, and declines when something goes wrong. Dashboards and Business Activity Monitoring software can be used to determine when to respond to changes in Decision Yield. In the example we can see that the decision yield has risen steadily largely as a result of improvement in Speed. Precision jumped early and then stabilized while Consistency and Cost have remained fairly constant. Agility
has fluctuated significantly including a large drop in the most recent period, something probably worth investigating. A graph like this can be a valuable tracking tool.
In addition a radar chart could be used to plot current against target and trailing average results for example the graph below shows the same data as the one above, this time considering the average of the last 3 readings with the initial and target numbers. Again you can see that Speed, and to a lesser extent Precision, are the improvements being made at present.