Product Analysis

ToM Joseph

1.) Which 5 of these providers is the most active with Fitogram and why?

Chart showing feature weightages

Engagement Score

To determine most active providers , a metric ‘Engagement Score’ is defined :

Engagement Score= w₁ x f₁ + w₂ x f₂ + w₃ x f₃ + ..... w_n x f_n

Where w is the weightage(-1,1) given to feature f

​

​

​

Most Active Providers

Most Active providers have high engagement score

2.) Group the providers into 4 different categories using the data at hand. Explain how you did this, which data points you used to create each category, and what each category represents.

User Cohorts

• Users are divided into 4 segments based on ‘Engagement Score’
• Light Users has scores in range (0-3000)
• Medium Users (3000-6000)
• Heavy Users (6000-9000)
• Super Users (9000+)
• Each of this segments represent how the users interact with fitogram pro
• Behavior of different segments is explained in the slide-Behavioral Cohorts

3.) Which of these providers do you feel is likely to churn and use another software? If you had more data, what indicators of this might you see over time?

Least Active Providers

• Decrease in engagement leads to Churn.
• Providers with least engagement score are more likely to churn
• If we have more data, we could calculate Daily Engagement Score by omitting the feature - Bookings All Time from earlier formula and find how different user cohorts engage over time . Questions like ’why engagement score is decreasing ?’ , ‘what software features are less used over time(and by which cohort) ? ’ etc. could be answered.It helps to identify where providers face trouble in using the software or if customer service needs a step up

4.) What information are you missing from this table to help you know more about the provider?

• Type of business ( private company,proprietorship,freelance etc. )
• Number of employees
• Revenue
• Location - region
• Contact details, telephone ,email,website
• Lead(focal point for contact) contact name,email and telephone
• Channel through which they found Fitogram (social media,search engine etc.)

​

5.) Of this data, what is the least and the most useful in your opinion and why?

• By dropping one row which has all Null values,we get the above chart
• It can be inferred that most used feature for Providers is Bookings -which in our case can be called as the ‘North Star’ metric
• Least useful feature is ‘occupied seats total’, same information is captured by ‘percent_seats_occupied’

6.) Create one single visualization of this data of your choice. Explain why you chose that visualization type and what it represents.

Behavioral Cohorts

• Adjacent visualization helps to identify behavioral patterns of different User Cohorts. It is possible to answer questions like ‘Why do different user cohorts use fitogram pro and how do they use it‘ ?
• It can be found that Super users mostly use the product for USC Bookings and they make highest number of intercom chats followed by Light Users.This indicates that the chats are mostly related to USC booking or events held�(Pearson’s correlation value shows positive correlation between events_held,Bookings All Time with Intercom chats)
• ‘Medium Users’ has the highest average of ‘Customer Bookings count’.It means that they actively promote use of fitogram pro to their clients.If more clients use the platform, it is more likely that they start using more features and become a paid user
• Heavy users mostly use the platform for pro booking and conducting events
• Verticals are represented by different colors

Correlation Matrix (Pearson’s r)

Above chart shows correlation between features