Product Analytics: A Comprehensive Guide to Using Data for Better Product Decisions

Product Analytics: A Comprehensive Guide to Using Data for Better Product Decisions

I remember the first time that a client told me how much analytics had helped their business.

They were able to increase their sign up rate for their product by 22% while reducing their marketing costs. It wasn’t magic or fancy tactics. They simply used their analytics data to make informed decisions.

They didn’t have to guess or take huge bets. They knew exactly what was working and what they needed to do more of.

This is power of analytics and more specifically, product analytics.

Product Analytics Defined

It seems like the definition of product analytics is always changing, but I like definition from the folks over at Amplitude:

“Product analytics show you who your users are, what they want, and how to keep them”

This article will be an essential and comprehensive crash course into the world of product analytics. It will explain why you should care, what tools you can use, and what reports will be most helpful.

If you have some familiarity with with this topic, then you can skip around to the sections that are relevant to you. Think about it like your very own “choose your own adventure” game.

Who Benefits the Most from Product Analytics?

Every company has “users” they would like to retain and get more like them.

However, when we talk about product analytics these days, we are really talking about companies who offer products such as web apps or mobile apps. Ecommerce companies who get the majority of their revenue from digital properties would also fit into this group.

To understand why these types of companies would benefit the most, let’s look at the types of questions that you can answer once you set up the right analytics tools.

Questions that Product Analytics Can Help You Answer

We are looking for answers to business questions around two challenges: how to acquire users at a lower cost and how to retain more users. Questions such as:

To answer these questions, you need to be able to track a user from the moment they first learn about you to the moment they become a customer.

Product Analytics gives you a robust view of your user (Image Source)

If a user is able to do all of this through digital channels, then product analytics can really help you.

Don’t Invest Too Early in These Tools

A few months ago, I was talking to a company that wanted to get the best analytics tools for their product (a web app for doctors). They were almost done with their MVP and they were getting ready to start their marketing.

They gave me a list of 100+ questions that they wanted to answer. The questions were good but their main challenge was that they had zero users. It didn’t matter how good the tools they implemented were, they would simply end up with empty charts and reports.

To be data-driven, you need data.

Instead, they needed to focus on getting their first 100 users before they invested in some of the analytics tools that I’m going to mention later on in this article.

This is a common mistake that I see new companies make all the time. They want to use the same tools that establishes companies use even though they don’t have anywhere near the amount of traffic or users.

Your company needs to have certain “minimum benchmarks” before you consider investing in advanced analytics tools. Data is only useful if you have enough of it. You can’t take one data point and extract valid conclusions from it.

I recommend companies consider investing in product analytics tools when they reach the following numbers:

  • 100+ B2B users (companies) or 2000+ B2C users (consumers)
  • Actively experimenting with different marketing channels
  • Spending $1000+ per month on user acquisition

If you don’t hit these numbers, you can focus on qualitative data e.g. interviews, surveys, etc to get the answers that you’re looking for.

How to Get Started by Getting the Right Foundation in Place

Now that we know who can benefit the most from product analytics, let’s talk about the how to get started.

This isn’t about tools because tools can and will change. Instead, you need to think about certain foundational pieces that will make your life easier in the long term.

1. Learning to Connect Business Goals to Data

Analytics is meant to help you grow your business. This may mean more revenue, more users, more referrals, or anything else that matters to you.

I constantly see companies that have a lot of data but no way of using that data to actually grow their companies.

To avoid this, spend some time trying to understand what parts of your business could be improved with more data. This is similar to the questions that I listed above but you want to focus on areas where you’re performing below industry averages.

Perhaps your onboarding funnel has a poor conversion rate or your user retention is too low. This is where data can help you.

The best analytics projects are the ones that start with a crystal clear objective (e.g. we want to figure out how convert more of our free users to paid users).

Have crystal clear goals and objectives

2. Create a Tracking Plan Before Writing Any Code

Most of the tools that you will be interested in fall into a category called “event driven tools”. Think Mixpanel, Amplitude, Kissmetrics and Intercom. These analytics tools rely on events to collect data.

An event is simply an action that a user takes (e.g. signing up for your product, uploading a picture or playing a song).

Playing a song would be an “event’ for Spotify

You can also send properties alongside any event. If a user uploaded a picture, you might also want to know what kind of picture it was (jpg, png, gif) or what size the picture was (500px by 500px). Properties is where most of the magic occurs and they are critical to getting the most out of your data.

A tracking plan is an Excel or Google Spreadsheet that contains all the events and properties that you would like to track. Doing this planning in a spreadsheet will help you avoid critical errors.

Almost every client that reached out for analytics help didn’t bother to create a tracking plan. They simply figure out what data they wanted to see and started writing the code necessary to track those events and properties.

3 months later, they realized that they are missing important events or properties and they can’t actually get valuable insights. They now have to go back and fix (i.e. redo) their initial implementation.

Choosing the Right Analytics Tools

Everyone loves tools. Good thing because there’s an ever increasing number of analytics tools in the market. Choosing among the hundreds of options can be a full time job for any company.

I can’t give you an overview of every single tool in existence. Instead, I’ll talk about the big players, some of which you already know.

Something to keep in mind is that you won’t be able to find a single tool that does everything. This magic unicorn doesn’t exist yet. Instead, you will end up with a “stack” of 2-3 tools that answer different questions.

1. Segment.com

Segment.com lives on top of all of your analytics tools and it’s main purpose is to help you simplify the implementation process of multiple tools.

For example, let’s imagine that you want to use Mixpanel and Intercom at the same time. Both of these tools rely on events to get data and they will need roughly the same events and properties. Instead of implementing the same data twice, you can send your data to Segment.com and they will send it to Mixpanel or Intercom.

You can track your data once and send it to hundreds of tools. Using Segment.com will also make it easier to move away from tools e.g. moving from Mixpanel to Amplitude without having to rewrite your tracking code.

2. Google Analytics

Google Analytics is the gold standard for analyzing marketing traffic. Nearly every company will keep this tool to understand everything that happens up to the sign up step. This means looking at marketing channels and what users did before signing up.

After a user signs up, we can then move to tools like Mixpanel or Amplitude (listed below).

3. Mixpanel

Mixpanel is great at helping you understand what your users are doing post sign up. Did they complete the onboarding process? Are they coming back and using our product? How can we engage our users to take a certain action?

They focus primarily on mobile apps and media companies.

4. Amplitude

Amplitude is similar to Mixpanel but their product has a more modern feel. Like Mixpanel, they focus on what happens after a user signs up and what users are doing inside your product.

They are a great option for ecommerce companies and companies who have cross platform products (i.e. a user can use your product on mobile, web, and tablet at the same time).

5. Kissmetrics

Kissmetrics is one of the original competitors to Mixpanel and and they also focus on the same questions or challenges.

They have recently added notifications which let you use emails to engage them or bring them back to your product.

6. Heap Analytics

Heap Analytics is similar to Mixpanel and Amplitude but they focus on companies who don’t want to undertake big implementations. Heap “tracks everything your users do automatically” without any extra code, making it easier to implement than Mixpanel, Amplitude or Kissmetrics.

If you don’t have development resources and your products are mostly on the web, then this can be viable option.

A typical stack of tools would look something like this:

  • Segment.com
  • Google Analytics
  • Mixpanel OR Amplitude OR Kissmetrics OR Heap Analytics

A Look at the Most Useful Reports like Funnel and Cohort Analysis

Let’s take a look at the most useful reports that you will use on a regular basis. Most of the tools that I listed above will able to create these reports or an equivalent.

1. Retroactive Funnels

The concept of a funnel is common in the digital marketing world. We can use the same idea to understand the different steps or journeys that we want our users to take.

This funnel report from Mixpanel looks at the different actions needed to complete an onboarding funnel. They start by signing up for the product, completing their profile and taking a critical action which would marked them as “Onboarded”.

We can instantly see the drop off for each step and where we should spend our energy. In this case, getting more people to go from step 2 to 3 can be a big win.

2. Cohort Analysis

Cohort Analysis lets us see the retention for any group of users (i.e. a cohort). The report looks like this:

On the left hand side, we can see how many new users sign up for our product. We can then see what percentage of those users came back and used our product after 1 day, 2 days and so on.

You may find that you’re really good at getting new users but you’re not that good at getting them to come back and use your product.

3. Segmentation to Find Insights

You can also segment your data to find valuable insights. This is where properties come into play.

In this report, we segment our “signup completed” event by the “authentication_type” property so we can see how our users like to sign up. We can then see that 98% of our users prefer to use email instead of Facebook or Twitter.

We might consider removing the social options altogether and seeing if that improves the signup rate.

4. Event-Based Notifications

We can also use our events and properties to send targeted messages to our users. Instead of simply sending them 5 emails after they sign up, we could tailor our communication to only send messages that relevant to them.

Imagine that you’re trying to guide users through an onboarding sequence. Some users might complete only 1 step while others might complete multiple steps. We can set up our messages to only be sent if users have taken or not take a certain action (i.e. event).

This is exactly what tools like Intercom.com let you do. You can define criteria like “users who signed up less than 30 days ago” AND who did a certain action.

Long Term Problems to Watch Out For

Product analytics data is powerful, but only if you’re able to use it over the long term. There are several problems that all companies will run into and that can kill your analytics projects if you’re not vigilant.

Here’s the most common problems that you should be aware of.

1. Naming Conventions

Naming convention refers to what you choose to name your events and properties. Your naming convention should be logical and easy to understand. For example, if you have an event that tracks when a user purchases something then you should call that event “purchase.”

Image Source

I’ve seen companies that complicate their naming conventions and end up using weird formulas that aren’t easily understood. Your naming convention should easy enough that it could be understood by a new hire within a few days of joining the company.

2. Lack of Training or Education

Training or education is an important part of any analytics project. If your product manager doesn’t know how to use Amplitude, how do you expect them to get the insights that they need?

You can do one session where you give your entire team an analytics crash course but you will also need to follow up with everyone individually as they start to generate their own reports.

Something that works well is to set up a Slack channel where anyone can post questions around the analytics tracking. These questions might be around what events to use or even how to create a specific report in an analytics tool.

3. Accuracy of Data

Finally, your company will need to actively work to maintain the accuracy of your data. If your team can’t trust the data, they won’t use it. Expect to see tracking errors come up after a new product release or major product changes.

Conclusion

I love analytics data, and I think it can help companies skyrocket their growth.

But I also know that product analytics isn’t for every company. You should only consider investing in this area until you meet the minimum benchmarks I listed above.

You should also assume that this is a six month project even if it only takes you 1 month to implement your analytics tools. It will take your team some time to get familiar with your data and to actually use it.

All that said, knowing how your users interact with your product, being able to discover aha moments and your best customers’ behaviors, and integrating with marketing data and business intelligence makes the case for product analytics quite compelling.

Related Posts

Leave a Reply

Your email address will not be published. Required fields are marked *

Current article:

Product Analytics: A Comprehensive Guide to Using Data for Better Product Decisions