If somebody asks you to measure the air temperature using only your eyes, you would call that person crazy, right? You can’t detect how fast or slow molecules are moving. It’s a useful metric. We know the temperature - we can react accordingly. -20C (-4F) is different from -20C (14F), and you can even use a different coat or something.
Now, ask an entrepreneur about systems. Probably, most of them will shake their heads. Ask about metrics, and they would say something like, “Oh, yeah. We have our main metric - profit!” Well, at least you have it but is it useful? Ask about OSP (Overall System Performance) and how to measure it, and you will get silence.
Because people often forget that any company is a complex system. It has multiple parts, and some of them might be quite bad (even shitty), but OSP will be incredible.
Take, for example, Amazon. Did you know that they have the shittiest server-side programming ever? It’s wild west there. Literally, everybody uses whatever they want and how they want. One team uses one server language, and the other team uses COBOL (a language from nineteen-sixties). Yet, everything works smoothly for users, and it’s blazing fast.
Their carefully chosen metrics for that include speed and how often the system is down. That’s it. Anything else doesn’t matter. It’s a part of their technological stack.
Now, let’s take a look at what’s in the front. If you take working with customers, they have an unusual metric (actually, insight here) of tickets per order. If there are 10,000 orders and 5 tickets - everybody is happy. If there are 10,000 and 5,000 tickets - something went wrong. Simple as heck.
They are constantly watching such metrics, get feedback from the reality and market, identify weakest “places in the system,” improve and then repeat. Doesn’t matter if it is shitty if it works for a customer (remember, Amazon is the considered to be the most customer-oriented company in the world… at least it was).
They introduce new metrics like… never? You don’t need a lot of metrics - you need only those that show OSP. That’s it. Now, take other businesses - they either have too many metrics (and they don’t know what’s happening) or too little or even none (and they don’t know what’s happening).
As far as I saw, most businesses focus too much on internal processes and interactions and forget (almost completely) that the system is exposed to the market.
Market weights your value, not your internal structure.
There are many crazy stories about how people in different departments focus on “good” metrics, but eventually, it appears to be useless. Like, from my experience, you can spend 10 hours improving your web app and save 3KB of data for each request. Then you realize that your data size is like 1MB, and those 3KB doesn’t help much (but 10 hours of life is spent almost worthless - no, you didn’t learn anything by doing that). Why did it happen? Poor priorities as a result of useless metrics (or absence of them).
It sounds funny, and you can say, “pfh, those guys are stupid.” Just try to check your organization or business and how people spend their time there. You will see that’s only the tip of the iceberg.
Any company should triple-check their metrics, throw unused or overused, focus on customers and value (long-term value) and optimize for that. Everything else is like measuring the air temperature with your eyes.