This might feel like a quant post, but it’s more than macroeconomics. The data field is already feeling the ripples of economic forces. ZIRP (Zero % Interest Rate Policy) has ended, which means that the era’s hiring, pricing strategies, budgets, and leadership philosophies are also at an end. The story begins with macro factors, but it’s really a growth story about how badly businesses need their data teams to take on a new role.
Hi Vin, I am a lead AI PM at health insurance company and working on calculating impact upfront and prioritising different initiatives. However, calculating impact upfront sometimes could be challenging as there is so much uncertainty/complexity in the business process that is going to consume it. It makes it difficult to track ROI as well. Sometimes, it is not possible to collect metadata related to a baseline process. Doing impact calculations takes time and you have to balance it with not losing the project to external vendor or competing internal team or stakeholder interest. I was able to connect the dots (model output to business KPIs) for some of the projects and it makes decision making super easy for everyone. So totally get the value but it could be challenging for some data products.
Clarion call for the current AI age. Powerfully argued and written.
Hi Vin, I am a lead AI PM at health insurance company and working on calculating impact upfront and prioritising different initiatives. However, calculating impact upfront sometimes could be challenging as there is so much uncertainty/complexity in the business process that is going to consume it. It makes it difficult to track ROI as well. Sometimes, it is not possible to collect metadata related to a baseline process. Doing impact calculations takes time and you have to balance it with not losing the project to external vendor or competing internal team or stakeholder interest. I was able to connect the dots (model output to business KPIs) for some of the projects and it makes decision making super easy for everyone. So totally get the value but it could be challenging for some data products.