Product Analytics in 100 seconds.

As founders and product builders, we're on a mission to produce product market fit.

Along the way, we talk to our users and collect qualitative feedback to improve the product.

There's a quantitative approach too.

Enter product analytics, creating user insight by surfacing usage data.

In product analytics, every interaction is a learning opportunity.

In our app, we capture those interactions, we transfer them into an analytics application, and create data artifacts that inform a new product change.

Four data scopes are important.

The user, who interacts with the product across multiple sessions as tracked by individual events and further described by event properties.

Within this online shop, let's track the click of a product card to understand browsing patterns of our users.

We first include an analytics snippet within our code that lets us use methods to communicate with our analytics tool.

Then we listen to a click event on the product card used to track method, define an event name and add details via properties, like the product category or the product's position on the page.

The event triggers a simple API call to the analytics server, submitting our custom properties along with information about the page it was fired on and device details.

Every event also includes an ID stored as a cookie on the user's device.

This lets the receiving server identify events from the same user and continuously build up an event stream.

When users provide additional information about themselves we can capture that using the identify method, attach user traits and build up a detailed user profile on our analytics platform.

We've been craving to answer why are landing page visitors not buying our products?

Spoiler, analytics won't tell us why.

Instead, a funnel analysis can tell us what is happening across our core conversion events and serve as input to decide what to do about the drop-off.

This is why providing access to the data and sharing it with our team is crucial.

From there on, we might want to dig deeper into the data, or we might want to pull in qualitative feedback and understand why users behave a certain way.

This way, we are not being entirely data-driven, but data-informed.

After all, data being a deep sibling to qualitative feedback tells us what problems to solve.

And then you basically use intuition to figure out what solutions to those problems might be.

We go from observation to insight to our next product hypothesis.

Product analytics can also help us to segment our users, for example, into acquisition channels, or we build so-called user cohorts along behavioral traits.

Here, we are assigning all users who set a goal within our fitness app to the goal setters cohort.

Let's use this cohort in a retention analysis.

We could ask what percentage of new subscribers keeps completing at least one workout per week?

And what's the difference between those that set goals and those that don't?

Such a retention curve falling to zero would not indicate product market fit.

This one much more showing less immediate drop-off and higher persistent usage in subsequent weeks.

The ability to segment users also helps us running A-B tests.

Before users receive our app content, we randomly assign a product variant and then direct users to this variant showing either A or B.

We also set a cookie to track what variant they're seeing.

In our analytics tool, we can then group users by this experiment property and understand how the introduced change impacts user behavior.

What gets measured gets managed, so focus on the quality of your metrics.

A North Star metric provides this clarity, but make sure to break it down.

Like the learning app Duolingo, optimizing for daily active users, but breaking that down into user activity states, managed by teams, optimizing a more movable metric that is informed by lower level product events.

With the Object Action Naming Convention, we keep our events set up consistent.

With an accessible tracking plan, we keep it transparent.

Our analytics needs are growing?

Consider using a customer data platform to connect multiple data sources to multiple destinations such as a data warehouse for joining analytics with business data, a marketing platform for delivering personalized experiences, or a webhook triggering custom workflows based on incoming event data.

Product analytics can be evaluative helping us to understand how did this particular product change impact user behavior.

It can also be generative enabling us to identify events that drive desired outcomes, helping us to prioritize new strategic opportunities.

Take Bourbon, a location app designed to make plans, check in at locations and share photos.

By looking at behavioral data, they uncovered that a small group of users wasn't actually using most of their features, but heavily sharing photos.

Bourbon decided to focus solely on photo sharing and launch with a new logo and name.

Ultimately, whether we use our measurements to inform strategy or evaluate features, the measure of who we are is what we do with what we have.

So as product builders on this beautiful iterative journey of chasing our vision, it is imperative to understand which ideas to cut short and which ones to take further.

Power your next iterations with product analytics.

And while doing so, don't forget, stay product-led.