Category: Data

Implementing Product Analytics

  Maryam Shah

linkedIn
Twitter
Instagram

  Maryam Shah

linkedIn
Implementing Product Analytics

Understanding digital products and how customers use them is key to creating stellar products. Lucky for us, product analytics tools allow us to capture and visualize this information rapidly and translate it into effective decision-making.

If you are new to product analytics, you may want to check our blog on Introduction to Product Analytics first.

Implementing product analytics for your business can seem like a daunting task, which, if not done properly, can become another resource-draining project for your business with minimal actionable insights. So, let’s walk through the recommended steps to implement product analytics effectively:

1. Become aware of your business context

When starting off, it is important to acknowledge that each business is unique and has a different set of objectives, goals, and vision. Hence, the implementation of product analytics needs to align with your business’ respective goals.

Become context-aware by understanding your business first. You may want to ask yourself the following questions:
What problem is my product solving?
What is the long-term vision for my business?
Who is the target audience for my product?
What challenges does my industry face?
Who are the industry leaders for my product domain?

Answering these questions will help you align your product analytics implementation with your business needs. Now you will understand what you need to know about your product and the improvements required.

2. Understand your product: draw a customer journey map

The second step is to understand your existing product. The best way to do that is to chart out a customer journey map. This map lists all the actions your user takes or the screens/webpages they view in a sequential manner. Understanding this user flow is important in understanding your analytics so make sure you draw this in significant detail. This will be your reference point when you get to step 4, which is defining your sub-metrics.

User journey for an e-commerce platform.

Fig 2. Example of a basic customer journey map on an e-commerce platform (Note: this is definitely not detailed!)

3. Identify your North Star Metric (NSM)

Identify your product’s end goal. This links to step 1, where we tried to understand why we created our product and what we want to achieve through it.
To translate this idea as a product term, ask yourself what’s the one action that you always want your user to take/increase. This will become your north star metric: the core metric you want to improve (which is key to driving your business).

North star metric

For e-commerce, the north star metric will be “purchases completed”. Everything becomes a secondary objective which is in some way trying to eventually improve the north star. Improving the search feature, the homepage listings or the checkout process are all improvements aimed at making a user complete a purchase. So, a good digital product will be designed to make the user easily and repeatedly complete purchases.

For Spotify, the north star can be “soundtracks played” while for Netflix it can be “videos played”. Be sure to spend some time at this step to find the right north star metric for your business. This is perhaps the most important step in getting your analytics right.

4. Define your sub-metrics

Once you have the north star sorted out, start understanding how you can make users improve the north star metric? We will call these secondary objectives sub-metrics that help us improve the main objective.
A good way to identify sub-metrics is to look at your product as an amalgamation of different modules. You can identify these modules through your customer journey map drawn in step 2. Subsequently, you can list metrics that you want to improve for each module.

Let’s follow the same example of an e-commerce platform. We can identify the following modules in a basic e-commerce product:

  • Login/signup
  • Product search
  • Product catalog
  • Cart checkout
  • Rewards & loyalty program
  • App settings

Each module has its own set of user events, user flows, and link toward the main objective. E.g. cart checkout is detrimental to completing a purchase but general app settings may not be as important. This would depend on your product and the industry norms.

Once you have the modules, you can list the metrics, which help you quantify the efficiency of the module’s performance. It’s key to refer to your detailed customer journey map at this step to understand the intricacies of your product. Good metrics capture product performance and quantitatively reflect any improvements which take place in the product.

Let’s zoom in on the Search Module to understand this better.

How do we quantify user behavior here? You may want to ask the following questions:

  • Percentage of daily users using ‘search’ – This helps us understand how many people continue to use the search feature. Any sharp changes in this trend can indicate breakage or improvement which may require further inquiry.
  • Search to add-to-cart ratio – This helps us quantify the users who actually find the product they were looking for and successfully place it in their cart.
  • Time spent on ‘search’ – This helps us understand if it’s taking users too long to locate products. Benchmarking with the industry’s average search time will further contextualize this information.
  • Time taken for search results to display – A long loading time will be a red flag for your platform whereas quick displays mean users don’t have to wait to see products.
  • The average number of search results displayed – If a lot of customers are unable to see any search results, they may be at risk of leaving the platform. Not being shown any products can be bad for product discovery as well.

This was not an exhaustive list of metrics and can vary across products. The key is to only list relevant metrics and not create a long list for the sake of completeness.

5. Integrate a product analytics tool

At this step, you have a list of sub-metrics for each module that you want to measure and visualize through a tool. So, first, look for the right product analytics tool. You may want to compare pricing plans, analytics capabilities, and use cases to find the right fit for your product. Some common tools used in the industry are Google Analytics, Mixpanel, Pendo, and Heap.

To integrate this tool, you will have to identify the events you want to track along with their event and user properties (discussed here). Refer to your customer journey map and sub-metrics list for this activity.

Continuing our example on the search module, we may want to have the following events:

  • Searched – upon submitting a search request (event properties could be “search word used”, “auto-suggest-selected”).
  • Clicked search bar – upon clicking the search bar.
  • Search results screen – displays search results (event property could be “number of results displayed”).

Again, remember to keep this list manageable and only include events and properties that may give you valuable insights. Be sure to check your pricing plan for any limitations at this step. Your tool provider and development team can handle the rest.

6. Visualize, learn, and improve

Once you have your tool integrated, you can visualize all your sub-metrics and your north star metric to effectively track, benchmark, and improve your metrics. Insights from these visualizations can become very powerful in improving your product.

Enjoyed Reading?

Read more from our blog

Read more from
our blog

Introduction to Product Analytics

  Maryam Shah

linkedIn
Twitter
Instagram

  Maryam Shah

linkedIn
Introduction to Product Analytics

Product analytics answers questions regarding user behavior around digital products. Insights and visualizations on product usage offer opportunities to improve product features and design products best suited to the users’ needs.

 

What is Product Analytics?

The process of tracking the events (actions) users trigger on a digital product across a user journey.

Each event has accompanying event properties (action details) and user properties (user’s details) for enriched analysis of user behavior.

User journey for an e-commerce platform.

Fig 1. User journey for an e-commerce platform with typical actions/events taken by a user

Events

Any actions a user takes on a website/application are called events. Clicking a button or any menu option on the screen is a trackable event. For an e-commerce website/application, an event can be clicking on “add-to-cart” or “product-search”.

Tracking events helps us understand which actions users take frequently, which ones they get stuck on and which ones they rarely use. This can provide great insights into user behaviour around your product.

 

Event Properties

Each event can have more data points of interest that elaborate on the nature of the event.
Following our e-commerce example, the event “add-to-cart” can have the name, quantity, and price of the product added to the cart with it as the event property.

Event properties can provide really detailed information about the actions taken. However, knowing which properties can provide valuable insights is important since tracking and storing unnecessary data can have significant costs.

User Properties

User properties are details about who the user is. These can be demographic details such as their location and gender or behavioral details linked to the product such as the number of purchases made, product categories explored, and reward points earned.

Behavioral user properties are updated based on the events a user triggers. E.g. the user making a purchase has certain reward points and cash-in-wallet before they make a purchase. The purchase event impacts their user properties as the reward points increase and cash-in-wallet decreases.

User properties enable us to divide users into different segments. Segments can interest the marketing team for targeted content or the product team for A/B testing.

Example of types of data collected through events.

Why is Product Analytics Important?

  1. Verify existing product features’ performance – How is the app/website currently performing? Are users easily able to use the product? Do they have any preferred sections/pages?
  2. Identify gaps in the product for improvements – Which features are not being used? Are pages loading slowly? Where do users stop using the product?
  3. Impact testing for newly launched features – How is the newly launched feature being used? Is it performing well?

 

Who is Product Analytics Beneficial for?

 

Role Value
Product Managers  Understand user behavior around the product to optimize and improve the product
Product Operations Measure user behavior for new product feature rollouts
UI/UX Team  Understand users better to design and build products
Marketing/Growth Team Understand user segments and use targeted content for growth
Customers Benefit from the value creation from the product analytics process

Enjoyed Reading?

Read more from our blog

Read more from
our blog

Value Engineering from Data

  Hussnain Ahmed

linkedIn
Twitter
Instagram

  Hussnain Ahmed

linkedIn
Value Engineering from Data

“The world’s most valuable resource is no longer oil, but data.” The Economist for May 06, 2017.

Indeed it is; however, data is only as valuable as the insights it can provide. Analytics help us make better decisions by uncovering hidden patterns and trends in our data. Doing so, on one hand, can impact business efficiency, and on the other hand, improve the quality of human experience.

When it comes to business efficiency, data can help create better products and services, as well as generate new revenue streams. This well-being is then passed on to the end consumer who now has high-value products to improve the quality of their life with.

Quotes

We need collaboration between data engineers, data scientists, and business subject matter experts to generate value from data.

Now the big question is: How can we create value from data continuously?

 

Data is not all about technology

Data is often seen as a purely technological topic. Technology is a critical enabler in collecting, storing, processing, and publishing data but analytics require a fair share of domain expertise. We need collaboration between data engineers, data scientists, and business subject matter experts to generate value from data.

Organizations invest in setting up data teams, platforms, and agile development practices. However, the journey of continuous value generation from data neither starts nor ends at such technical capabilities.

 

Business teams must own data initiatives

The data analytics story begins with business problems, such as a mobile app company being unable to retain users. Such problems are often part of strategic initiatives, i.e. the company’s goal is to reach X number of active users, but they are having trouble doing so. To understand user behavior and correlate it with user retention, we must first define the critical business concepts such as what is an active user, what is a churned-out user, how much time of inactivity is acceptable, etc.

An organization may run multiple initiatives simultaneously, and triaging to prioritize analytical use cases helps efficiently use data teams and technical capabilities. All such activities need ownership from relevant business stakeholders in the organization.

Business teams must own data initiatives

As we embark on developing data, it is critical that we invest time in designing and planning to ensure we identify key user personas and map out their journeys. This will allow us to gain a complete understanding of the various use cases.

Additionally, we need to lay out the data architecture to accommodate the ingress, storage, and processing of data. Finally, we must take the business use cases and formulate them into smaller data science tasks that can be executed by data developers.

The data implementation projects vary in scope and complexity, therefore it is essential to estimate the resources and time required for successful completion. At this stage, data platforms and agile development methodologies can be very helpful in managing simultaneous implementations and reducing time to value.

Quotes

Data is more valuable than oil because we can continuously extract value from it if we apply the proper mechanisms.

Data products

As data analytics becomes commonplace in a company, it is vital to treat the outcomes of projects – such as reports, dashboards, and models – as digital products and services. This way, we can ensure that the value generated by these projects is continuous and high-quality. To do this, we must maintain the data infrastructure, the quality of the data itself, and the quality of generated insights. We must ensure we have people, processes, tools, and governance in place to maintain data products and services.

Data is a renewable resource, unlike oil. It is more valuable than oil because we can continuously extract value from it if we apply the proper mechanisms.

Enjoyed Reading?

Read more from our blog

Read more from
our blog