Here's a well-written article that reflects something I've noticed across a number of jobs during my career: KPIs are often the problem, not the solution.
Yes, businesses should measure performance. They should seek to improve the metrics that matter.
But setting KPI targets forces staff to focus solely on reaching a particular metric for a particular target, at the expense of everything else.
@skribe I'd go further and say they really shouldn't exist at all.
Particularly for lower level employees, KPIs can shift blame for structural problems within the organisation (including it's policies and procedures) on to the individual.
Not reaching that sales target? It's not that the market is saturated, the product is crap, the sales channels are wrong, or the marketing is ineffective. It's just you aren't trying hard enough!
Higher up, it incentivises managers to sacrifice everything else in order to get a particular number looking good. A KPI on costs gets managers to focus on getting short-term costs down, no matter what impact that cost cutting has on long-term profitability.
It's just a bad way to run an organisation.
@ajsadauskas it's a good article and touches on the knock on effects of poor service to the wider society, which I hadn't considered.
But a good set of KPIs are supposed to balance this by requiring KPIs to be offset. For example, a profitability measure should be balanced by both a customer service measure and by a staff satisfaction measure.
You tend to get what you measure (and not get what you don't) so without KPIs, how do you optimise your service?
@zorrobandito This video gives a good rundown of the distinction between metrics and targets.
When you turn a metric from a system into a target, its value as a measurement is gone.
#business #economics #economy #capitalism #management #manage
@ajsadauskas @zorrobandito very true, but unfortunately a lot of people do not accept that. They argue, they need something to measure to see whether they are closer to their target. As they learned:
- Set a goal
- apply measures to meet the goal
- check whether you got closer to the goal
- if yes, continue
- else adapt measures
While this works for a steam engine, it does not work for people.
(BTW thanks for the video)
@prefec2 @ajsadauskas I'm not convinced. Your arguments assume that targets are set with no regard for dependent metrics. If that's happening, then that's bad/dumb, but doesn't mean that targets are inherently flawed.
I've run both call centres and ops teams and set efficiency and customer service targets with great success.
@zorrobandito @ajsadauskas the key argument in the above comment is that when you apply metrics to a social system and the measurements are used to evaluate people, people will optimise for the metrics and not for the goal. If you use metrics only to understand a social system and then use that to inform your management decisions then the effect is less so.
1/2
@zorrobandito @ajsadauskas for example in computer science, companies use(d) lines of code, number of tests, number of commits, and ticket completion time, as metrics to measure success, but then people optimised for these metrics. It does not result in better work.
The problem is that all these metrics do not directly measure better work, they measure properties that are also affected by good work, but could also be improved differently.
Now at @
skribe 🇺🇦 
Zorro
Reiner Jung 🇬🇱 🇺🇦 🇪🇺
TofuTheSquirrel