Two more I use a lot that aren’t vanity metrics like dashboard usage:
1) Cost reduced
2) Future cost avoidance
They’re not very important for growing companies where revenue growth will be the most important commercial metric, but super relevant to more mature / stable enterprises that are going through a cost-cutting focus
#2 is also relevant to fast-growing orgs where some costs are currently quite low, but can easily spiral out of control without pre-emptive measures (which will feel thankless if not reported on properly)
Great read, Eric. I've been making the same switch in how I approach my work.
I now optimise for decision impact. That means selecting the single biggest business process I can move, delivering highly curated observations on top of it (using LLMs to analyse and synthesise), and making sure there are clear action points or feedback on their side.
I absolutely understand your point when I imagine this taking place in a company whose product is a SaaS.
But when I imagine the data team being in a health care company, there is a legitimate interest in numbers that are not tied to ROI. There are medications for certain diseases that only a doctor can prescribe, the pharmaceutical manufacturer cannot simply sell them, and in Europe health care data is very well protected, that means you are not allowed to know which doctor has which patients. The company needs an immense amount of data of all kinds to somehow figure out, but not all of that data leads to an immediate decision, and you cannot measure the ROI directly.
Or take automotive. A company that manufactures cars is huge. The path to measuring which data has which ROI is extremely long. Think of all the data collected through connected drive for example, about the vehicle status. It is simply there. Nobody at the automotive manufacturer can decide that I now have to go to the workshop just because the vehicle status says I should.
You can certainly measure in other ways whether the data is useful, but it is often not directly financially measurable. There are also things that have a value that is not monetary, like customer satisfaction, being broadly positioned or building up capacities and capabilities.
I do understand your point though: how can a data team take away this pressure that is now building up.
Maybe this pressure can be reduced when the data teams connect more in two directions: business and infrastructure. Those who already had a lot to do with the business side can understand even better what the processes are, and those who were already working a lot on infrastructure can go even further in that direction.
But I'm not sure if it would work in a SaaS to be honest.
Two more I use a lot that aren’t vanity metrics like dashboard usage:
1) Cost reduced
2) Future cost avoidance
They’re not very important for growing companies where revenue growth will be the most important commercial metric, but super relevant to more mature / stable enterprises that are going through a cost-cutting focus
#2 is also relevant to fast-growing orgs where some costs are currently quite low, but can easily spiral out of control without pre-emptive measures (which will feel thankless if not reported on properly)
This line was huge:
"What I mean: What incremental revenue can you trace to a decision that wouldn’t have happened otherwise?"
Great read, Eric. I've been making the same switch in how I approach my work.
I now optimise for decision impact. That means selecting the single biggest business process I can move, delivering highly curated observations on top of it (using LLMs to analyse and synthesise), and making sure there are clear action points or feedback on their side.
I absolutely understand your point when I imagine this taking place in a company whose product is a SaaS.
But when I imagine the data team being in a health care company, there is a legitimate interest in numbers that are not tied to ROI. There are medications for certain diseases that only a doctor can prescribe, the pharmaceutical manufacturer cannot simply sell them, and in Europe health care data is very well protected, that means you are not allowed to know which doctor has which patients. The company needs an immense amount of data of all kinds to somehow figure out, but not all of that data leads to an immediate decision, and you cannot measure the ROI directly.
Or take automotive. A company that manufactures cars is huge. The path to measuring which data has which ROI is extremely long. Think of all the data collected through connected drive for example, about the vehicle status. It is simply there. Nobody at the automotive manufacturer can decide that I now have to go to the workshop just because the vehicle status says I should.
You can certainly measure in other ways whether the data is useful, but it is often not directly financially measurable. There are also things that have a value that is not monetary, like customer satisfaction, being broadly positioned or building up capacities and capabilities.
I do understand your point though: how can a data team take away this pressure that is now building up.
Maybe this pressure can be reduced when the data teams connect more in two directions: business and infrastructure. Those who already had a lot to do with the business side can understand even better what the processes are, and those who were already working a lot on infrastructure can go even further in that direction.
But I'm not sure if it would work in a SaaS to be honest.