4 ms·

> Targets are often useful because they're proxies for the unmeasurable Not just the unmeasurable, there's a lot of KPIs used because they are easier than figu

by gav 8y ago

> Targets are often useful because they're proxies for the unmeasurable Not just the unmeasurable, there's a lot of KPIs used because they are easier than figuring out metrics that are meaningful. Here's some examples I've seen: Number of Visitors: a large manufacturer with no direct channel considered the number of visitors to their website as the most important thing and this was reported to the highest levels. Without some way of determining the quality of the visitors, or if their experience was valuable, the actual amount is useless. Engagement: a bank was tracking how many visits and how minutes customers spent on their website every month, they considered more time a good thing. Personally I want to spend the minimum possible time dealing with my bank, preferably zero interaction, making processes more streamlined rather than having more content is key. Bad metrics allow you to report up the chain that you are moving the needle in some way and justify budgets.

Jtsummers 8y ago
I believe this was meant as a reply to my comment, and I agree with you entirely. Often the easiest metrics are the least useful. I worked in an office that wanted "production hours" (the business was largely labor-oriented, where that was actually a semi-useful metric, but it's not a useful metric in software development). It was gamed, as long as people charged directly to a project they were "productive". The costs of training and all were hidden as much as possible so projects and teams could seem productive. This meant that customers were being billed hundreds of dollars an hour for training that may not have even related to their project (bordering on fraud here, certainly abuse of some kind). But it also meant that the training budget (organizational) was slowly being cut, because no one was using it. Selecting your KPIs carefully may not prevent failure, but choosing them poorly can almost guarantee failure or mediocrity.

by gav 8y ago

Jtsummers 8y ago

I believe this was meant as a reply to my comment, and I agree with you entirely. Often the easiest metrics are the least useful. I worked in an office that wanted "production hours" (the business was largely labor-oriented, where that was actually a semi-useful metric, but it's not a useful metric in software development). It was gamed, as long as people charged directly to a project they were "productive". The costs of training and all were hidden as much as possible so projects and teams could seem productive. This meant that customers were being billed hundreds of dollars an hour for training that may not have even related to their project (bordering on fraud here, certainly abuse of some kind). But it also meant that the training budget (organizational) was slowly being cut, because no one was using it. Selecting your KPIs carefully may not prevent failure, but choosing them poorly can almost guarantee failure or mediocrity.