How have you empowered your PMs to be their own analyst? Any tips for getting PMs up to speed on running their own queries?

“Data depth” is a core competency for any PM and unlocks “data informed’ strategy.  However, how much data analysis a PM will need to do will depend on the maturity of the organization & available resourcing. At the core, every PM must know where the data is logged and in what format.

1. In organizations, where there is an organization wide data platform that democratizes data access & there are analysts on the teams, every PM should still be trained on using this platform and run lightweight queries. ‘Knowing your data’ helps the PM ask the right questions to the analyst and prevent being stuck in ‘analysis paralysis’. Data empowered PMs are also able to leverage their data analysts for more strategic thinking rather than mundane daily reports. 

2. In organizations where there are no dedicated analysts, PMs can try self serving with data but for complex scenarios, if the PM is being pulled away from defining the “what” (data is only 1 aspect of identifying the aspect), then I would encourage the PM to share the need to find someone with that skillset in the organization or raise the staffing constraint to their leadership. This doesn’t have to be a dedicated analyst but at the least a shared pool of analysts. 

All of the above assumes that the required amount of data is available in a  centralized data platform with an ability to query the data. If this doesn’t exist, I would say start with building such a data platform first. 

I would strongly recommend the below approach:

• Proactive:

  • in their product requirements document, PMs need to clearly lay out the success metrics and hypotheses to test, and work with engineering to implement product tagging that will allow them to compute these metrics and in/validate the hypotheses

• Reactive:

  • first, start by leveraging central data teams: ask them clear questions and have them get back to you with reasonable SLAs

  • If there is no central data team or if the SLAs are too long, invest in product analytics software paid with your PM leadership's credit card to answer simple questions. I would recommend Pendo because PMs can do tagging on their own and then answer easy questions about who and what. The experience for conversion funnels and retention analysis is being improved and not there yet compared to an Amplitude or a Mixpanel. So for these complex queries, get the data teams onboard ASAP in the product development process. With Pendo, make sure to keep in the loop a frontend engineering leader so they ensure that having Pendo tagging does not interfere with their synthetic tests.