3 ms·

> I personally do think he got some things wrong ("if you can't measure it, you can't manage it" mantra being one of them ) Could you explain this point more.

> I personally do think he got some things wrong ("if you can't measure it, you can't manage it" mantra being one of them ) Could you explain this point more. I'm genuinely curious.

A4ET8a8uTh0 4y ago
Sure. Some context may be needed. I suppose my views are formed in this way partially, because I became an MBA after having soul crushing jobs, which does tend to change your perspective a little bit ( and give you an insight into what is going on in a head of the front line drone that is forced to implement orders the best way he/she can ). << "if you can't measure it, you can't manage it" 1. It is wrong in a literal sense. You obviously can. Maybe you should not ( which is a valid question to ask ), but you absolutely can manage things and not measure everything. 2. It is wrong when taken metaphorically to a logical conclusion. The mantra seems to encourage ( and did encourage based on some of the things managers tried to measure -- lines of code come to mind ) measuring everything. In accounting, there is a concept similar law's 'de minimis' ( I kinda don't want to use that name, because I worry people wills start questioning that from tax perspective, which is its own animal I don't want to touch, but in practical terms, the name matches the concept exactly - basically 'too small to matter'), where accounting for some items in some processes is borderline pointless ( say a screw ). Every so often, people in accounting classes go 'but what if', 'but we can track it'. And the answer is invariably: 'You can, but is it worth it?' Naturally, you could counter it with: 'Well, you don't know if you don't measure it.', but there has to be a point at which the tracking of atoms get too crazy. So that is my basic thought on the matter. I can expand a little more, but I think that covers it. Note, I am not saying it is not worth it to measure stuff. I am saying it is important to be discriminating over what you pick to measure, because, people being people, will optimize for it.
triceratops 4y ago
https://en.wikipedia.org/wiki/Goodhart%27s_law https://en.wikipedia.org/wiki/Goodhart%27s_law
extragood 4y ago
From my perspective: there are some things you fundamentally can't measure. One of the teams I manage is technical support (B2B SaaS). Support's core functionality in an organization is to prevent churn. If a customer reaches out to support, they have some issue that is blocking them from using your product effectively. If you are unable to unblock them, then you risk them switching to a competitor. The problem is that not every support ticket/interaction will result in churn so we can't really directly measure how effective we are at our core function. So instead we measure the number of tickets raised, customer satisfaction, adherence to SLAs, etc, which is as good of an approximation as we can get with the resources we have available to us. In my case, those numbers all look reasonably good, so I'm treated as an effective manager, even though we don't (and maybe can't) really know how much churn we've prevented. I wish we could though - it would make my arguments to executives on behalf of the support team that much easier to say, hey we retained $XXXXXX in revenue over this period.

> I personally do think he got some things wrong ("if you can't measure it, you can't manage it" mantra being one of them ) Could you explain this point more. I'm genuinely curious.

A4ET8a8uTh0 4y ago

Sure. Some context may be needed. I suppose my views are formed in this way partially, because I became an MBA after having soul crushing jobs, which does tend to change your perspective a little bit ( and give you an insight into what is going on in a head of the front line drone that is forced to implement orders the best way he/she can ). << "if you can't measure it, you can't manage it" 1. It is wrong in a literal sense. You obviously can. Maybe you should not ( which is a valid question to ask ), but you absolutely can manage things and not measure everything. 2. It is wrong when taken metaphorically to a logical conclusion. The mantra seems to encourage ( and did encourage based on some of the things managers tried to measure -- lines of code come to mind ) measuring everything. In accounting, there is a concept similar law's 'de minimis' ( I kinda don't want to use that name, because I worry people wills start questioning that from tax perspective, which is its own animal I don't want to touch, but in practical terms, the name matches the concept exactly - basically 'too small to matter'), where accounting for some items in some processes is borderline pointless ( say a screw ). Every so often, people in accounting classes go 'but what if', 'but we can track it'. And the answer is invariably: 'You can, but is it worth it?' Naturally, you could counter it with: 'Well, you don't know if you don't measure it.', but there has to be a point at which the tracking of atoms get too crazy. So that is my basic thought on the matter. I can expand a little more, but I think that covers it. Note, I am not saying it is not worth it to measure stuff. I am saying it is important to be discriminating over what you pick to measure, because, people being people, will optimize for it.

triceratops 4y ago

https://en.wikipedia.org/wiki/Goodhart%27s_law https://en.wikipedia.org/wiki/Goodhart%27s_law

extragood 4y ago

From my perspective: there are some things you fundamentally can't measure. One of the teams I manage is technical support (B2B SaaS). Support's core functionality in an organization is to prevent churn. If a customer reaches out to support, they have some issue that is blocking them from using your product effectively. If you are unable to unblock them, then you risk them switching to a competitor. The problem is that not every support ticket/interaction will result in churn so we can't really directly measure how effective we are at our core function. So instead we measure the number of tickets raised, customer satisfaction, adherence to SLAs, etc, which is as good of an approximation as we can get with the resources we have available to us. In my case, those numbers all look reasonably good, so I'm treated as an effective manager, even though we don't (and maybe can't) really know how much churn we've prevented. I wish we could though - it would make my arguments to executives on behalf of the support team that much easier to say, hey we retained $XXXXXX in revenue over this period.