This Fall we were invited to present to the Silicon Valley Engineering Leadership Group, hosted at Varian in Palo Alto, California on the topics of product development metrics. Ron LIchty, one of the leaders of the group was curious about metrics because has rarely seen them do any good. In fact, he said that metrics are more harm that good.
In this video at the opening of the talk, I describe the all too common situation where people use a bathroom scale as a dieting strategy! You don’t measure the result to get the result. Yo need to measure the input metrics.
In my experience at Apple and other technology organizations, we found that predictive metrics, those metrics that change rapidly and are indicative of the final results were instrumental in creating lasting improvement.
Predictive product development metrics have the following properties:
- They are easy to measure & objective
- Change rapidly (days/weeks, not months/quarters)
- Indicative of the final outcome
- Tracked versys a target curve
Using metrics with these characteristics can help organizations generate lasting change. The steps to create predictive product development metrics are often represented as a tree. Below is an example of one branch of the tree that might have a total of three metrics. This is to create one metric, for example:
- Identify the overall goal or result
- Determine the root cause that challenges goal achievement
- Develop a plan to reverse this root cause
- Create a predictive product development metric that measures progress of plan
- Estimate the organizational & technical complexity of the planned solution
- Estimate the time to get a 50% improvement (1 month if only 2-3 functions, and no technology)
- Create a target curve by plotting out a 50% improvement per month for 4-6 months
- Track the actual progress versus the target curve
You can get more information by going to our Insights section.