How to Excel as a Data Product Manager

This is a short summary of the talk on how to excel as a data PM with Michael Barros, head of product at Surgo Ventures, an early stage company using AI to solve health and social problems with precision.

Here is the link to the recorded event.

What is a data product manager? Data product managers specialize in data to create products. As online businesses produce tons of data in their day to day operations, working with data can be a differentiator.

According to Michael, there are four data maturity levels every product manager should be aware of. I will discuss each data maturity below. To better understand data product management, Michael used a Health Tech B2B company as an example.

1. Raw data
These are the data that are gathered from users and are completely transactional.

Each time X event occurs Y log is recorded. For patients this means provider visits, lab results, or claim submissions.

2. Information
The next stage after raw data is information. These data are aggregated and organized. This is the time where PM and analysts create definitions of the data they gathered and drive consensus. It is important that everyone in the team has a shared understanding of the data and keep in mind that definitions do evolve and change over time. This is the stage where you want to invite subject matter experts, so they can help you with the definitions you will need to make sense of data.

What does it mean to define data? Defining data means labeling your data according to well known facts and based on the aggregation of your raw data.

For example, let’s say you want to define diabetes. What is diabetes? What does it mean to have diabetes? How can you tell if a person has diabetes? That means, given a set of data, you will be able to tell if a person has diabetes.

Moreover, you may also be interested in knowing what’s the difference between high risk vs low risk diabetes?

3. Knowledge
This is the understanding of the context around definitions. In this stage, you wonder what key factors lead to a person’s diabetes? What are the leading indicators of success? What are the leading indicators of failure? Given a set of data, what are the habits of a person who may have diabetes? What is this person's weight? What is this person’s food consumption like?

4. Wisdom
This is the stage where you move away from what has happened into driving behavior you want to happen. It depends on knowing what info is relevant. You can create a hypothesis, for example, a person is likely to develop diabetes, this is what you can do to prevent it.

Data Quality Exists at each level of maturity

To be able to read, analyze and make hypotheses and conclusions based on the data you have, it is important that your data is in good quality. Here are the criteria for good data quality.

Accuracy - Here you want to know what percentage of data points have questionable accuracy versus those that are certain. For example, of all the 35 patients studied, how many people are falsely diagnosed with diabetes?
Completeness - You want to make sure whether all data is present within a dataset. Check the null values of each record. How much percentage of all the records are missing some data?
Timeliness - The time in which you get your data may be important. Are you using historical records that are within acceptable range that are relevant to work you are doing?
Validity - Make sure all the data such as date are in the right format.

Skills to brush up on

Here are the skills product managers need to have to perform their duties well.

SQL (Structured Query Language) - useful tool to ask questions on data like an analyst
DBT (modular SQL orchestration) - a tool that enables data analysts and engineers to transform data in their warehouses simply by writing select statements
BI tools (Power BI, Tableau, and Looker) - type of softwares designed to retrieve, analyze, transform and report data for business intelligence
Schema and ERD - Schema is the blueprint of how your data is organized in the database whereas the Entity Relationship Diagram shows the relationship of entities such as people and objects from one another in the system

9 great books on how to use Data to build better Products

7 Product Analytics Books To Make Better Data-Driven Decisions