Using Real Usage Patterns to Guide Product Strategy

In the early-stage startup, usage patterns provide quantifiable metrics to guide the product strategy. 

• What exactly is the ‘Aha!’ moment in your product? 
• How much usage correlates with retention?
• Prioritizing which product features will bring the most output? 

The most well-known companies relied in their product development more on usage patterns than anything else. For example, Facebook. Here is an excerpt from the Acquired podcast: 

“… 10 friends in 14 days. The other cited 7 friends in 10 days of how do you create the magic moment? … You set the threshold and then you’re like, okay, if we can deliver this delightful experience where now people have a rich newsfeed and the people they care the most about in the world to interact with, they’re going to retain.” 

Facebook’s growth team examined what number of friends early on is a threshold to creating the magic moment that makes up an engaged user. Based on this discovery, the product development prioritized “People you may know…” and improved the news feed experience.

Then there is also the Bible app from YouVersion. The market of Bible apps is extremely competitive, yet this one managed to dominate the niche. The source of product strategy was nothing but usage data and user psychology. Key usage insights were:

• The importance of daily engagement led them to develop daily notifications, starting with reminders about their daily reading. If the user doesn’t complete it, a notification offers to try a less challenging one. 
• The power of daily Bible verse variability for the user’s goal: there were quite a few people who waited till midnight, so that they could see the message the app sends.

In this blog post, we’ll discuss what to look for in usage patterns to guide your startup’s product strategy.

Tool-centric vs Behavior-Centric Data Analytics of Usage Patterns

Industry behemoths like Facebook and the Bible app have analytics built into the product and use their own proprietary insight mining methods. This is expensive and takes a long while to implement. For startups, especially early-stage ones, there are out-of-the-box solutions such as Mixpanel and Amplitude for tracking real usage patterns. 

Surely, the market of product analytics is way richer with tools like Google Analytics, Salesforce, Zendesk, and others. Those are still valuable in a way, and teams often use them. However, they constitute a tool-centric approach, rather than a behavior-centric one. Within a tool-centric approach, the teams focused on each having their own tool:

• The marketing department keeps tabs on acquisition and traffic via the page views, sessions, and traffic sources from Google Analytics.
• The product team measures feature adoption with clicks and features used from Mixpanel or Amplitude.
• The sales team focuses on revenue and accounts with Salesforce.
• And the customer success team tracks tickets/issues from Zendesk to ensure customer retention and satisfaction.

Each tool operates with its own data and captures its own distinct part of the user journey. This approach gives siloed insight pertaining to the team. While these tools are used in this way still deliver great benefits, they miss the following:

• Which onboarding experience ensures higher customer lifetime value (LTV)?
• Interaction with which feature, or even better, sequence of interactions, caused an upgrade?
• Which early user behavior can predict churn vs retention within a month?
• What is the usage difference between successful and unsuccessful users?
• Which interaction breaks the user journey?

Compare & Contrast: Screens with Highest Churn vs Insights from Usage Patterns

Below is a screenshot from a product analytics tool that tracks the usual: monthly active users, session count, screens with the highest churn, and so on. The traditional analysis will go something like this: the top three churning screens are checkout, homepage, and product details. So, the team will decide to redesign those screens to reduce churn and consequently increase conversion. While everything is logical so far, tracking usage patterns can be much more valuable.

Example 1: Where do the users navigate from?

For instance, let’s take churn on the checkout screen. Deeper behavior analytics might reveal that:

• Users who add more than 2 items, use filters, and visit several product details pages have a churn of 35%;
• Users who add 1 item and come directly from an ad churn at 70%.

So, the problem is not the checkout screen in itself. The problem is where the user comes from and the expectations set there. In this case, a product decision of redesigning the checkout page is a wasteful activity. The real impact will come from aligning the ad campaign with product details on the web app. 

Example 2: What sequence of actions leads to churn?

Another example is the sequence of actions. Let’s take the same checkout churn. Analyzing real usage patterns also includes analyzing sequences of user actions.

For instance, let’s consider this sequence: user navigates to homepage -> category -> product detail page -> add to cart -> checkout -> final costs with delivery & ETA -> Dropout or Purchase Completed. When a user goes to the product detail page, the delivery cost and & ETA are hidden behind a small link, like “Delivery Info”. 70% of users who do not click it drop out at checkout. The majority of those who view it complete the purchase. The regular ‘screen church’ data will not show this, and instead of redesigning checkout, it is best to make delivery info available without having to click on a link. The value for product development is immense: an unnecessary checkout redesign, which is unlikely to generate any ROI, versus a small tweak that is likely to generate substantial ROI. 

Summary

This does not mean that page views, churn by screens, or rage clicks are unnecessary. They are health metrics: they can help easily detect a problematic area. However, to correctly identify the cause and determine the best product decision with the highest ROI, one needs to dig deeper into user behavior. This is where tracking usage patterns comes in. 

Analytics of Product Usage Patterns

Let’s take, as an example, a marketing campaign creation and scheduling tool. The team defines basic functionality: 

• a short onboarding flow,
• a library for content, 
• a scheduler, 
• an AI assistant for creation
• Social media integrations for previewing and publishing posts.

In the product-led growth model, analytics takes a behavior-centric approach. The tools are set up to track the full user journey, and this is done in-product. The goal is to learn what experience made the users retain, upsell, or drop off. 

Visitor Landing: What events should you track before users sign up?

For the Bible app mentioned earlier, registration is not necessary. They actually have the majority of users never going through the sign-up process. Yet, those users are still increasing their bottom line through engagement and virality. 

The point, though, is that with tools like Mixpanel, you can still track usage patterns in the form of events even if the user of your app is not registered or logged in. Initially, Mixpanel treats all users as anonymous. It assigns each user their unique token and records all the behavior through events specified in your app. So, for instance, you can know whether a visitor opened the app, started onboarding, checked the pricing page, etc. 

Even more crucially, though, the product must send the information over to Mixpanel on where the user comes from. This requires proper product setup for campaigns, such as utm parameters, tracking invite links from other users, campaign IDs, and such. This way, you can refine your marketing by tracking where the most converting users come from, what blocks visitors from converting, and so on.

Starting User Journey: What leads to user activation?

In the screenshot below, you can see the Mixpanel dashboard for a ride-sharing app that tracks the following:

• First App Open,
• Sign-up,
• Log-in,
• Add Payment Type.

This B2C app considers user activation as a set of 4 straightforward events. Though some might argue that it should also include the event “First Ride Completed”. Overall, every app will have a unique event flow that they consider User Activation.

For most SaaS tools, registration is pivotal. Once a user signs up with your app, Mixpanel merges this information with the token it used to track the anonymous user behavior.

The major point of tailoring analytics specifically to your product and your product strategy is setting up in-product events. These should not be behaviors like a button clicked or a modal opened. Instead, event taxonomy should focus on users accomplishing something meaningful. 

These events are part of defining the product activation hypothesis.

For our marketing campaign creation and scheduling tool example, the activation should be a bit more complex. For instance:

• Onboarding started
• Signup Completed
• Onboarding Completed
• Connected Social Account
• A Marketing Project Started
• First Content Created (Aha! moment)
• First Post Scheduled (Value experience, user activation)

Having these events tracked, you’ll know whether each of them is required for activation. You can also assess the quality of activation. For instance, some sequences will include completion of onboarding, others won’t, and you’ll have the opportunity to evaluate them against the retention data. It will show the importance of the onboarding experience and whether it needs more work.

Inside the User Journey: What usage patterns make users stay?

To track retention, your team should set up events relating to core product value. The sequence of events should allow you to measure recurring usage patterns. For our marketing tool example, these can be:

• Post Created,
• Post Edited,
• Post Scheduled,
• Post Published.

With this setup, you can track usage patterns that will tell you mathematically the usage of a regular user, the usage of a power user, and which usage is the threshold for retention, among quite a few other things. In its turn, this will help you:

• Break down users into cohorts based on quantifiable usage patterns, 
• rethink pricing, for example, based on the difference between regular and power users, 
• Define a retention threshold and insert onboarding steps that will nudge new users to complete those, ensuring higher conversion rates. 

Moreover, in this setup, you can consider separate events for features that cost your business a lot. For instance, AI-generated Post, or AI Image Generation. Those features significantly increase the costs of serving users for your business. This will help you define toxic power users who abuse your service. As such, this will help you preserve your margins and quantify usage caps. 

Connecting to Business Outcomes: Usage Patterns for Revenues

Business outcomes are set up as user events as well. For instance:

• Pricing Viewed (interest),
• Subscription Purchased (revenue),
• Upgrade Completed (upselling),
• Subscription Cancelled,
• Payment Failed.

Here, you lock in the full user journey, from knowing where the user came from through usage to purchasing decision or lack thereof. The discoveries can go along the lines: 

“Most paying users come from colleague invites,  Pricing Viewed 3 times, create 7 posts a week, and publish at least 5 posts a week.”

So, then you’d re-focus your product strategy from ad campaigns to promotions for sending invites. This may lead you to develop an internal growth engine, such as exchanging user invites for extra product usage. And, actually, there are a myriad of product decisions for developing and improving your revenue engine. Therefore, insights from Mixpanel/Amplitude help prioritize those in your MVP roadmap analytics in a way that leads to the highest business outcomes, whether it be engagement, retention, or direct revenues. 

Extra: Tracking Event Duration

Every user event is time-stamped. So, you might track durations by specifying each event with creating variations such as [Event X] Started and [Event X] Finished. This will allow for such insight for your product strategy as: “users who created posts within 24 hours activated, users who took more than 1 day to create a post churned”. Learning this, you might push your product strategy toward developing contextual in-app messaging and AI-prompting to help users create posts within the 24-hour window. 

The number of completed events and the duration of working with events can help identify cliff behavior. This means identifying usage that leads to substantially higher retention. Insights often look like: “Users who spend under 20 minutes completing a post retain 5x times more.” This may involve diving deeper into content formats and developing specific prompts to complete longer or more complex posts faster. It might also lead to developing an app or email notifications that remind users to complete their started posts. 

Final Words

Usage patterns are paramount for your product strategy. Not only do they help you track meaningful user actions rather than clicks or views, but they also help you:

• Create high-impact marketing,
• Ensure early user experiences that convert,
• Make data-informed decisions on user cohorts,
• Spot toxic power users and cap usage to preserve profit margins,
• Adjust pricing and devise smarter upselling,
• And, finally, get deep insights into what usage patterns make users LOVE your product and tweak your product roadmap to nudge more users down this path. 

FAQ: Using Real Usage Patterns to Guide Product Strategy