Product Analytics: The Fundamental Guide for 2020

How to get started and evolve your analytics stack over time. All this & more in this one-stop guide to product analytics.
Debdut Mukherjee
Debdut MukherjeeJune 1, 202012 min read

Time and again, folks have used product analytics tools to better understand their customers. Businesses use analytics to determine product health, improve the customer experience, test product-market fit, and ensure that they are making the right investments with limited time and resources. The bottom line is that product analytics enables you to accelerate the growth of your business.

In the end-user era, where you acquire customers by leveraging product-led growth, knowing what specific actions folks take in your product is instrumental in guiding your strategy across product, sales, marketing, and customer success.

Let's check out some of the reasons why companies should dive head-first into product analytics.

Why should companies care about product analytics?

Traditionally, implementing product analytics hasn't been easy. It takes time and involves aligning multiple stakeholders to create a working product analytics engine that you can use to drive your business forward.

However, the effort is worth the reward as some of the fastest-growing companies use product analytics to grow their business. So, let's look at the top reasons why companies implement product analytics.

  • Like I mentioned earlier, you need data to help validate your decisions. Product analytics ensures that you're putting your time and effort into the right things. By connecting what customers are doing, you can design better customer experiences and ultimately increase your core growth metrics.
  • Product analytics enables you to analyze your company's growth. You can perform trend analysis to see what happened in the past, how the business is performing today, and draw predictions into the future. This is hugely beneficial when you're trying to forecast a course for your business.
  • If you're trying to measure product-market fit, you need to bring product analytics into the fray. By measuring retention analysis, you can check whether the same cohort of users return week after week and are deriving value from your product.
  • With product analytics, you can understand how customers are using your product and what features led to retention? This can give you insight into the Aha moments that bring your customers back to your product.

So now that you know why product analytics are important, let's see what questions you can answer.

What questions can product analytics help you answer?

With product analytics, you can answer a wide array of questions about your product's health and growth, among other things. Here are some questions that you ought to know.

What questions can product analytics help you answer?

1. Product analytics helps you with the 'what': With quantitative analysis product analytics can help you answer questions like:

  • What channels have the highest engaged users?
  • What does our retention curves look like?
  • What percent of users are discovering our core features?
  • Where are users getting stuck in onboarding?

With a tool like Amplitude or Mixpanel, you can easily answer the above questions and much more. But be warned, don't forget about qualitative research. By combining qualitative and quantitative analysis, you get a better understanding of your customer and why they are engaging, or not, with your product. Qualitative research helps answer the 'why' questions for you.

2. Validating your hypothesis: Good product teams often use the scientific method to validate their work. They make an observation, ask questions, form a hypothesis, make predictions, test those predictions, and iterate. Here, product analytics comes into play to help inform and validate your hypothesis.

For example, if you're a product manager at Atlassian working on Jira, you might form the following hypothesis using product analytics:

"We believe that educating customers about core concepts during onboarding will result in an increase in retention, because we've observed customer confusion and a X% dropoff after project creation."

To validate this hypothesis at Atlassian, you'd typically run a traditional A/B style experiment where you'd compare the control to the new experience. However, if you're working at a startup, you'll likely not be able to split test to get statistical significance quickly, so it's best to establish a baseline for your onboarding funnel and monitor how things change over time.

It doesn't matter the size of the company, product analytics can help increase your confidence by providing you some signal to validate your hypotheses.

3. Inform roadmap decisions: Product analytics can go a long way in informing your product roadmap. But it's best to start by outlining the growth model for your business. What are the inputs that drive growth? This includes identifying your acquisition channels, user retention rate, etc. which helps you calculate your growth rate as an output of the model. Once you've done this, you can start to see how making tweaks to the input metrics impact the output that you're trying to achieve.

At this point, you can get more sophisticated with your growth model and start to include key product behaviors or growth loops. To do this, you'll need to capture the most valuable actions to generate insight into what customers are doing in your product. With your model in place, you can start making predictions on how the impact of different roadmap items will move your core metrics and which ones you should prioritize. Check out RICE if you're looking for a framework for prioritization.

4. Measure feature adoption: With product analytics, you're able to track how folks are engaging and adopting new features. One of the worst things that can happen is you ship a feature that nobody uses. This is not only demoralizing to the team but also increases the chance your product is entering a death cycle.

Product Death Cycle

Before writing a line of code determine your success criteria for adoption as part of the definition of done. Make sure you're measuring this after your release before moving on to the next item on your roadmap. You don't always get it right on the first go and might need to iterate, or unship the feature and avoid adding product bloat. (improve it, replace it or kill it).

Now that you understand how product analytics can help, let's dive in to see how different teams use it regularly.

Which teams use product analytics and how?

Having reliable customer data has never been so crucial to business success. Product analytics is used by more and more teams nowadays. Here are ways that some of these teams use product analytics.

Which teams use product analytics and how?

🔨 Product Analytics + Product Team

It's a safe bet to say that product teams are one of the primary users. With product analytics, PMs can get answers to questions like:

  • What's my active usage?
  • What percentage of users are engaging with a particular feature?
  • Where are users getting lost during onboarding?
  • Which features contribute to retention?
  • Which cohort of users are the most engaged?

Product analytics can also help a product team personalize their onboarding experience by triggering in-product messages or sending emails based on the actions, or lack thereof, that users have taken. This helps you guide customers on a journey to discovering your Aha moments, improving their overall experience.

You can use tools like Chameleon, Appcues, or Pendo to get started easily.

🙌 Product Analytics + Customer Success Team

Customer success teams can leverage product analytics to measure customer health. By creating a customer health score based on your product's key actions, you can answer questions like:

  • Which accounts are likely to churn?
  • Who are the power users for a particular account?
  • Who would be an ideal referenceable customer?

By measuring product usage, NPS, and other factors, you can determine the overall health score of a customer. This enables you to identify at-risk accounts quickly and proactively intervene to get them back on track. You can use a tool like Sherlock to automate this for you.

💼 Product Analytics + Marketing Team

Traditionally, marketing teams have been measured on marketing qualified leads or MQLs. However, with the fast adoption of product-led growth, modern teams are focusing on qualifying leads based on the actions they take in the product. By focusing on product qualified leads, or PQLs, you're better able to prioritize your acquisition channels and segment customers. You can start to answer questions like:

  • What channels have the lowest CAC and highest activation rate?
  • What channels are driving the most engaged customers?

You can start to spot trends that can inform future marketing campaigns by connecting acquisition to activation. For example, if you notice that a high percentage of folks from a specific channel don't match your product's typical usage patterns, consider tailoring your message to reinforce their behavior, decreasing your customer acquisition cost.

Additionally, you can start to use first-party behavioral data to power targeted campaigns. For example, in Amplitude Engage you can create a segment based on your existing customer base and sync in to Facebook to build a look a like audiences lowering your acquisition cost and boosting your ROI.

💰 Product Analytics + Sales Team

Product analytics can quickly help the sales team determine what value an account is getting at a glance. For example, your sales team might want to know the number of seats deployed and what percent of those are active each month. By distilling your product down to a core value metric (e.g. number of queries each month for Amplitude), you're able to make this data actionable to the broader team. Lastly, it's best to make this data accessible in the tools your organization already uses, like your CRM, so that they're able to act and put it to use quickly.

Depending on your pricing model your sales team likely has opportunities to upsell or cross-sell your existing customer base. These are ways to provide additional value to your customers and generate additional revenue for your business. Most importantly, with product analytics, you already have insight into what customers would benefit from these offerings by looking at your existing customer base.

Now that we have talked about how different teams make use of product analytics let's see how your data stack might evolve depending on your company's stage.

How should your product analytics stack evolve?

While product analytics are important, your investment should be commensurate to the stage of your company. Your analytics stack should evolve overtime as your company grows.

How should your product analytics stack evolve?

👶 Early Stage (0-20 employees)

At this stage, you have a ton on your plate and need some signal that you're headed in the right direction. Simply install:

The goal is to iterate quickly and improve your core product metrics. Stay disciplined and don't overthink it.

💪 Mid-Stage (20-100 employees)

At this stage, you've found product-market fit, congratulations 🎉, and you've hired a few folks who rely on this data to do their job.

  • You'll look at using product analytics data for more use cases like experimentation, marketing automation, personalization, etc. so it's important to incorporate analytics into the teams definition of done and treat it as a feature of your product.
  • You'll likely want to start establishing a better analytics discipline to help fuel future growth; this includes hiring an analytics lead to take ownership of this moving forward.
  • While adopting a self-service tool for product analytics has been great, you'll want to be able to combine product usage data with other data sources, so look at adopting Stitch or Fivetran to sync this data into your data warehouse.
  • With this data now in your warehouse you'll want adopt a BI tool that can be used across your organization for reporting. Check out Looker, Tableau, or Mode Analytics as they're all good solution.
  • Start incorporating analytics into the teams definition of done and treat it as a feature of your product. Make sure it's tested and teams are encouraged to

📈 Growth Stage (100-∞ employees)

This is a critical stage. Your team has grown significantly along with your data stack. At this stage, a few things you should keep in mind to ensure a smooth sailing experience.

  • You may want to shift your analytics event tracking to Snowplow, an open-source data pipeline or build your own. This requires dedicated investment but gives your ownership of the data and control over long-term costs.
  • Set up a process for data quality. Since data is probably coming from a dozen sources, you need to ensure that the data is accurate and trustworthy.
  • Document everything. As the team grows it's important to document your tools, processes, and create a data catalog. Gitlab has done a great job of this with their Data Handbook.

While this might seem like a lot, all journeys start with a single step. If you think we missed something, drop us an email at [email protected].