I found an article that I like today, its called 5 steps to transparent metrics, written by Hank Marquis.

The presentation of the article is very plain, but the subject is very relevant: what are you using your KPIs and Metrics for?

All metrics should have an object and a purpose for improvement. Otherwise, what is the point of spending time and money on collecting the metrics?

Metrics are not free, there is often a significant cost associated to collecting, collating, calculating, and translating data for every set of KPI. So if you’re not using a KPI to make an object or a process better, faster, and cheaper - don’t bother collecting it at all as its just creating busy work and wasting money.

To ensure that Metrics are valuable, your KPIs should have at least this 5 parts

1. Object: What are we analyzing?
2. Purpose: What is the purpose for the metrics?
3. Focus: Which attribute of the object are we analyzing?
4. Stakeholder/Viewpoint: Whose perspective are we reflecting?
5. Context Factors: What is the context?

Metrics should have a purpose

Wikipedia on Goal Question Metric Method.

“The first rule of any technology used in a business is that automation applied to an efficient operation will magnify the efficiency. The second is that automation applied to an inefficient operation will magnify the inefficiency.” Bill Gates

Someone asked me today: “…isn’t automation optimization?”. The answer is: NO.

Automation is accomplished by tooling. Optimization is accomplished by humans.

Only a human can look at a set of tasks and put them in a logical and sequential order to accomplish an objective i.e. only a human can say it takes these sequential 20 steps, 8 hours, and 4 people to build a wagon. Only a person can build a process.

Continuing, only humans can do process optimization i.e. looking at the process end-to-end and determine if we reorder step 3 and 4, combine step 6 and 9, and run steps 11 and 12 in parallel, we can eliminate unnecessary effort - then it will only take us 16 steps, 4 hours, and 3 people to do build the wagon. That’s called improvement, reduction and savings.

If a process cannot be done right manually then is reasonable to assume that the process cannot simply be improved by automation. So the answer remains a no, automation does not mean optimization.

Automation is tooling: software, robots, etc. Its objective is to ensure that the steps of process dictated by humans are automatically executed the same way every time and hopefully much faster. If automation is done correctly there is a digital trail of when a task starts and ends. This will then provide the owner of the process a trend of information that can he or she can use to assess and implement even more optimization.

When somebody tells you that the tool is not working start the troubleshooting with this question first: what is this tool being programmed to do? There is a very high probability that the problem is not the tooling but in fact it is the process the tool is automating.

One thing that people often forget is that - processes change. And they change often. If your process changes every 90 days then your automation tooling must be capable of supporting this high rate of change. In this case automation is almost a requirement.

I think it is fair to say that your tooling must be as agile as your processes are. A failure of either is a failure of both. But knowing the difference is the key to success.