Deepika
Deepika
Sep 4, 2018 · 3 min read

A product manager’s thoughts on data

In business school, the importance of gathering good data, analyzing it with appropriate tools and making decisions based on the insights obtained were all repeatedly drilled into us.

With my newly minted MBA degree, and my first real product manager job at thirdACT, I’ve been attempting to put all those lessons into use these last few months.

This article is an attempt to capture some things I’ve learned in the process.

Let’s start with why product managers care about data, and what they hope to do with it. This part is pretty much what we are taught in business school.

Product Managers care about data because

  1. It informs product conception and strategy — You identify demand for your product or a specific new feature in market and customer data, identify gaps which could turn into your product’s competitive advantage over similar products, identify what features work in favor of the product’s growth etc.
  2. It informs iteration and improvements to the product — You understand your customer’s reaction to your product or feature or website, you measure users’ engagement with specific aspects of your website or product, you solicit feedback about current and desired features. You can then use this to predict your customers’ lifetime & churn, and inform areas of improvement in your product.

Sources of Data

All of the above works wonderfully when the data is in clear, organized and easy-to-visualize form. One thing I’ve learned in the past few months is that data in real life is like food in a farmer’s market as opposed to a restaurant — all the ingredients are there, but you have to do work (and sometimes a lot of work!) before you can have a nice meal!

When I was in business school, our data projects involved using data from easily usable, tool-friendly format. All we had to do was plug in data into SPSS, Minitab or Excel, then pick the correct concepts and techniques to apply and voila! We would have some nice insights.

In the real-world, data that we get is messy, disorganized and often needs to be cleaned and processed before it is ready for use. Friends at bigger companies with mature products often have access to a mature data warehouse and an extensive set of tools to work with it. When you are starting out — especially in an industry like ours (real estate) that’s not known for great digital records — there are a lot of spreadsheets and, sometimes, completely unstructured data.

At thirdACT, I built a data pipeline for collecting and organizing property data from a web based data source that didn’t offer comprehensive exports. A mix of selenium based web automation and some python scripting gives me clean, structured data in a nicer tabular form. I then use Excel or Tableau to come up with different insights in easy and visually appealing charts.

Tools & Techniques for data analysis and insights

This isn’t a comprehensive list, but is merely an attempt to describe the set of tools and techniques I’ve gotten a chance to use.

SQL DBs & SQL queries: Great for storing large amounts of structured data, and then generating reports quickly and easily from that data.

Python, pandas and statsmodels: I’ve only just begun scratching the surface of what we can do with Python. Pandas and statsmodels are libraries that give you a fantastic toolkit to work with data in python. Since it’s a general programming language, you can use python for everything from one-off cleaning scripts to fix a batch of data, to complex machine learning on large datasets.

R: R is powerful and general like python, but was written exclusively for working with data and statistics.

SPSS, Minitab, Tableau etc: These are proprietary tools that offer the ability to perform complex statistical and machine learning analysis on your data without the need for the extensive programming required by Python or R. I’ve only used trial or free versions of these programs and on small datasets at school and briefly (with Tableau) at thirdACT. If your company uses the full version, these are potentially a faster way to perform detailed analysis of data.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade