Product Tip: Mixpanel
Product Analytics For Product People
There is high value of integrating analytics with every new feature in the product cycle to get the key data points. It’s one thing collecting data and it’s another ballgame analysing the data.
Mixpanel: The analytics as a service platform
Mixpanel's core mission is to help the world learn from it's data.
They have built the only product analytics platform that lets everyone in your organization deeply understand each user journey. Get instant insights and fast iterations throughout your product development process.In this case, we are collecting data relating to how customers use your product. Once collected, we can enrich with data from other sources or extract using the Data Export API. This offers a chance to pull data from mixpanel for more analysis. We can load this to a data warehouse too,for even more data magic.It's data everywhere.
Product Metrics:
Entails the core product metrics that we need to track, and their direct influence on the business and user goals.Use the most engaging metrics to help define better features to create an optimized product instead of just focusing on verbal feedback and survey forms.
Use case:
- How many people are clicking the signup page or any other call to action button on the site and where is their previous visit on the site?
- Where are people dropping of during a particular user flow?
- What referring sites are giving us the best ROI?
- Where do customers with the longest lifetime value come from?
- Get insights on user actions and follow up accordingly
- Send notifications at different flow in the user journey and according to the actions the user has taken.
Common terms in Mixpanel:
Events:
An action that the user might take on your product or app. Eg signing up for a product or downloading something
Properties:
Extra pieces of information that we find inside events. Eg An event will contain properties such as city, country, date, browser, location, referer. You can also set your custom property and events and send them alongside events
Profiles:
This is the identity users are given once you identify them properly, from this then you can see any action a particular user took on your product. Used in the people analytics.
Funnels:
This details the user flow journey and events
Segmentation:
Data for an event, segmented and filtered by properties over a time period
Retention:
Get data about how often people are coming back, also can be the cohort analysis
Installation:
- Can be used to track Android, iOS or web apps by adding mixpanel dependencies and initializing the mixpanel libraries like js library for web apps , iOS/Android for mobile apps
- Also client libraries which supports different programming languages. These are official libraries that can be used to easily consume data through the Mixpanel API. It’s a RESTful API and offers different resources
Tips on tracking:
- Use easy or relatable event names so that it’s easier to make sense of the data
- Choose a single place to track in the project, a case of track backend and frontend differently
- Use distinct IDs to track users, preferably the IDs in the database
- Carefully choose what to track and how that directly influences the business goals to avoid tracking vanity metrics. Clearly differentiate actionable and vanity metrics for your particular business.
Using the AARRR startup metric
These KPIs are best tracked alongside Customer Acquisition Cost (CAC) and Customer Lifetime Value (CLV).
- This is a simple way to track actionable metrics that will help grow the product, carefully divided into 5 steps:
Acquisition:
- Entails visitors per month eg 1000
- How long they stay on the site
- How many pages they visit
- What they abandon
- The key channels they originate from: Email, blog, SEO, Twitter, Facebook, Domain, Marketing channels. See which has the highest number, lowest cost and best performing
Activation:
- How many of these visitors are converted in % eg 70% and how many users does this translate to in a month? (700)
- These could include those who sign up for emails, newsletters, blogs and those who actually create an account.
- This entails homepage signups, how long they stay on the site, and the clicks.
- From this you can monitor and do A/B testing on different features in the site.
Retention:
- Of the 700, 20% are coming back after their first visit. (140)
- These could be those who open emails, notifications , are reading the blog and are repeat visitors.
Revenue:
- Out of the (140),only 10% are paying, this will translate to only 14 paying customers
Referral:
- Out of the 700 who are converted, 10% refer their friends to our services.
The AARRR applied using the mixpanel tools:
- A: Users come to site from various channels. Create a signup event. Clear segmentation to see their properties too.
- A: Use funnel.From traffic source to how it transfers to a sign up and how it can be improved
- R: Retention tool according to metric. Signed up and came back.
- R: Through retention tool or funnel ,can view referrals.
- R: Revenue tool and the actual revenue generated.
Types of metrics that are common in product analytics:
- Quantitative: This deals mostly with the usability testing, get as many people as possible monitored
- Qualitative: This deals with traffic analysis and user engagement
- Comparative: From this you can conduct A/B testing on different features
- Competitive: Monitoring and tracking competitor
The components to analyse are diverse. The most common ones are:
- Audience segment which is the people analytics
- Channel source which targets to measure what channels have the best ROI
- Campaigns and themes: Can track which marketing theme converted more users
- Landing page: This tracks if the message on the landing page, the calls to action buttons are communicating well.
Other analytics tools
Defining key metrics is a great exercise to get you thinking more about your product. How people will use your product, how you want them to use it, and what you want both your business, and your customers, to achieve from your product. More of a clear product strategy.
Reference:
We will dig deeper into the actual implementation in the next writeup.
