Using Data Analytics To Build Better Products

Posted by

I am fascinated by the application of Data Analytics to solving everyday technology problems.

In particular, I love how data can provide a guide to make better decisions, mitigate risk, and help lead to better outcomes for users.

In this post, I will discuss how different analytical tools can be used to help you decide what to build and why and how data can guide your design and product iterations.

When building products, and in particular MVPs (minimally viable products), I try to follow this framework:

  1. I start with the customer and work backwards. 
  2. I use insights from customers to build tools or products that add value to these constituents.
  3. I aim to simplify. I seek to be externally aware and look for new ideas from everywhere.

One place to get useful ideas is in data.

I have used user, website, and statistical datasets to generate new product ideas.

I want to pass these lessons on to you. In particular, I try to leverage qualitative data, regression analysis, and statistical ranges to guide my design decisions.

Qualitative Data

I first start with qualitative data. This data type is non-numerical in nature and can be gathered via user studies and interviews.

As a builder you should never be done learning and you should always seek to improve your products by asking potential and current users probing questions.

Before building a product or MVP, you should ask potential users the following types of questions:

  1. What do you want/need?
  2. What’s your problem or issue?
  3. How would you benefit from this product?
  4. How would this product change your behaviour?
  5. What do you like most – and dislike most – about the product/vision?
  6. What could be improved about the product/vision?
  7. What is the easiest/hardest part about using the product?

The feedback that these users provide is one form of data. Leverage this data to drive products forward. This is true if the product is pure digital tech, hardware (like computers, servers, LED lighting), consumer health, or athletics

As a builder you must insist on having high standards — and you can’t have high standards without getting user feedback.

The data you collect will help you raise the bar and deliver better quality products.

Regression Analysis 

In data analytics, regression analysis is a set of statistical processes for estimating the relationships between a dependent variable (often called the ‘outcome variable’) and one or more independent variables (often called ‘predictors’, ‘covariates’, or ‘features’).

When you built or designed your last product did you leverage a regression analysis to help you better understand the intent and behaviour of your users? Did you segment users into cohorts and look at trends?

If you are one of those product designers who overlooks regression analysis because the concept is either foreign or there is a lack of understanding of how to do it, now is the time to dive deeper and learn how to perform this analysis.

Because regression analysis is used to estimate the strength and the direction of the relationship between two linearly related variables, the tool can help you answer critical product design questions like this: if my users do X, are they more likely to do Y?

Without this guidance, building the right tools for the right users is a lot harder.

Understanding the Range (of behaviours)

The range is the difference between the lowest and highest values. It is easy to calculate yet critical to building the right products.

In the following data set, what is the range: 7, 2, 3, 4, 1, -5, 10, 9?

To find the range, find the smallest and largest value and calculate their spread. In the example above, the dataset spans from -5 to 10, equating to a range of 15 units.

Why is understanding the range so important for building products?

Firstly, it helps you understand the upper and lower limits, or possibilities for product adoption. If you are going to build a product that has a small range, the total number of people who might benefit from what you are building might be narrower.

Secondly, seeing the range (and the data within the dataset) can help you understand dispersion, which is the extent to which a distribution is stretched or squeezed. If you see that your users (or data) are closely tied to the middle of the data or closer to your upper or lower bound, you can make informed design decisions that benefit these cohorts. I am a big fan of this approach.

Here is one example that can illustrate how to use this type of data to make an informed product decision. I look at some basic attributes of user cohorts for my website (ages of the users, time spent on the site, and locations) to produce content most relevant to the largest number of site visitors.

By using insights about my audience I can produce relevant content and in-product workflows that help real users. For example, I noticed that a large number of my website’s traffic ignored a majority of pages I had taken time to build. In fact, the vast majority of my traffic went (and goes) to only a handful of pages. 

I used this insight to improve the pages that people do visit and to sunset those pages that people bypass.

Without understanding dispersion (and time spent on sites) I could not have made that inference.

If you are new to building, here is a simple sports metaphor. What is the range of scores in a golf outing or basketball game? Use those range of scores to influence what you believe a score might look like (highs, lows, averages) for any given game. That structured logic is now a useful guide to help you understand the future.

Conclusion: Applying Data To Product Design

Designing products is rewarding. Using data to build even better products is an even better use of your time because you can ensure that what you are building adds the most value for the people you are serving. 

Data is not something to fear. Rather, we should all embrace it in its many different flavours. Leveraging qualitative data, regression analysis, and by looking at data dispersion (with a focus on the range) you can start to see who you users are, what they value, and how to best align your product and design decisions with their needs.

Jim Barksdale, the former CEO of Netscape, famously remarked: “If we have data, let’s look at the data. If all we have are opinions, let’s go with mine.” Data can take your well-intentioned and well-informed opinions and unlock new insights for you, your teams, and your products. The beauty is that data is always useful whether or not you are building a mobile app, fintech products, enterprise VPN software, a parenting blog, or affiliate software. These improvements, over time, will add value for your users. And that is the ultimate goal, isn’t it?


With Crowdfire, you can find curated content, schedule your posts, engage with your audience, deep-dive into analytics and create custom reports. Now introducing Social listening. Try it for free.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.