API Analytics: Power Users and Valuable Marketing Channels

I struggled with the title of this article as it covers a good amount of analysis methods for finding vital marketing channels, power users, etc. Here are some I thought of:- Uncovering the Power Users of Your API: A Deep Dive Analysis- Data-Driven Insights for Your API- Finding Your Most Valuable Customers and Marketing ChannelsNonetheless… let’s continue.

Dataset and Tools

Prerequisites for this analysis include summing up usage and revenue

I’m using Python’s Jupyter notebook (can also use SQL) and a fake dataset I generated. Find the csv here. Here’s a brief on each field/column in the dataset:

user_id: A unique identifier for each user of the API.user_age: The age of the user.referred_medium: The channel through which the user learned about the API (e.g., Reddit, Google Ads, Facebook, Sales Reachout).referred_campaign: The specific marketing campaign that referred the user to the API (applies only when referred_medium is 'Google Ads').employment_industry: The industry in which the user is employed.month: The month in which the user used the API.pricing_plan: The user's pricing plan (Basic, Premium, or Enterprise).total_api_calls: The total number of API calls made by the user.hubspot_owner: The HubSpot owner from the sales team responsible for the user account (applies only when referred_medium is 'Sales Reachout').retention_month: The number of months the user has been using the API.price: The cost of each pricing plan chosen by the user.

The dataset looks something like this in a dataframe

Metrics to Calculate

We use median and not average here to avoid factoring in outliers that deviate us from the actual metric. Combining fields, as seen above, outvalues the insights given to us by single areas (e.g., user age group + employment industry > only user group )

API Usage: (Monthly)

• Median API usage by HubSpot owner + pricing plan

• Median API usage by age group + employment industry

• Median API usage by referred medium + campaign

Revenue: (Monthly)

• Median revenue by referred medium + campaign

• Median revenue generated by age group + employment industry

• Median revenue generated by HubSpot owner + pricing plan

The Actual Analysis

User Analytics

Power or value users provide the business with the most revenue. Therefore, we find power users by aggregating the revenue by user’s age and industry of employment.

df_sum = df.groupby(['user_age', 'employment_industry'])['total_api_calls', 'price'].sum().sort_values('price', ascending=False)

Using a heatmap, this aggregation above sorts the data by subscription revenue and API usage. Some deductions to take away from the above are that:

• Users between 18-25 working in:

  • Finance and Retail produce the most significant monthly revenue from all user groups.

  • Education use the API the most monthly.

• 20% of the users from the retail industry make up 2 of the top 5 user groups generating the most subscription revenue

• Compared to other groups, API usage is higher than the revenue produced amongst:

  • 25-30 working in Finance

  • 18-25 working in Education

Our power users are between 18-25 from Retail and Finance and 25-30 from Education. This should do 50% of the job of a google ads manager. Now it’s up to you to create campaigns to target more of these users.

Marketing and Sales Analytics

So what worked and what didn’t? Or maybe who worked and who didn’t?

Revenue by Channel 

We aggregate the metrics by referred_medium and referred_campaign and see 4 types of mediums for referral and 3 campaigns run for PPC (Pay Per Click) Ads through Google.

Sales reach out is an instant winner here, followed by Reddit and Facebook. The median monthly revenue from sales reachout is almost the same as the total revenue from running google ads. This does not indicate that sales reachout works better than google ads, as we need insight into the cost comparison of both. Google Ads also generates users with potentially a higher profit margin due to their lower product usage.

Revenue by Hubspot Owner (Sales Analytics)

We will cover a deep dive for Hubspot analysis in the upcoming substacks. For now, let’s do a petite POV into revenue by the salesperson.

Regarding plan cost: Enterprise > Premium > Basic and revenue generated follow the same order.

Here are some takeaways and insights:

• John :

  • Generated revenue equaling Bob and Jane’s combined through sales of the Enterprise and Premium plans

  • Brought in users that have high product usage, which could mean lower profit margins

• Bob and Alice:

  • generated almost the same amount of revenue equally but as Bob in the lead

  • brought sold premium plans to users that provide a relatively higher profit margin due to their average product usage

So we understood that:

• Users between 18-25 from Retail and Finance and 25-30 from Education produce the most revenue

• Revenue from sales reachout = Revenue from PPC

• John’s sales precipitate the most revenue by a salesperson (Congrats John)

This leaves room for A/B testing in each department (Sales and marketing) and experimenting with the pricing plans for the product.

It also concludes today’s article. Here’s how to keep yourself updated with the next one.