top of page

How startups should use product analytics (even without an analyst)

Understanding and analyzing product metrics can be a game-changer on the path toward healthy unit economics.

This article gives a rough picture of what tools early-stage startups can use to collect data, what exactly they should analyze, and how to do it without an in-house analyst.

External analytics tools

Popular external tools for product analytics

External tools are about 70% accurate. They link to your platform and show basic insights on how customers interact with a product.

You can track, for example, the number of users who clicked a button to start training in a fitness app and those who completed it. Then, you can identify the drop-offs and their timing. The data can be visualized with the same tools.

In the beginning, early-stage startups often use external tools like Mixpanel or Amplitude to track metrics related to product use: activation, adoption, and retention. They are easier to set up than internal tools and typically come with APIs for integration. But in the long run, I recommend switching to internal tools — storing data in internal databases gives more accuracy and control.

When it comes to categorizing users or identifying those generating the highest profit, it's also more effective to use internal data tools.

Internal analytics tools

Popular internal tools for product analytics

Internal tools ensure 100% accuracy. They are useful for warehousing and analyzing critical business data, such as registrations and transactions, and evaluating the success rate of core user actions.

They keep information in the cloud or other storage systems based on customers’ IDs, including the number and amount of their payments; it allows for a more accurate analysis of each person’s customer journey. I recommend tracking purchase-related metrics with internal tools.

The data can be visualized in Tableau or Looker Studio. Using internal data tools requires knowledge of SQL or hiring an analyst.

Combining external and internal tools

You can use both types of tools for cross-analytics, such as studying both users’ behavior and their spending habits. This gives a full picture: for instance, you can see how many users, after clicking a particular button, made a purchase within a week/month.

Using a combination of internal and external tools also helps to segment customers based on feature adoption, and see their impact on revenue. If it’s a meditation app, for example, analysis might show that users who listen to meditation sounds generate twice as much revenue compared to those preferring fairy tales.

Product metrics to monitor

At the start, businesses should track monetization metrics:

  • conversion to purchase on a sales screen;

  • repeated purchase;

  • average revenue per user (ARPU);

  • average revenue per paying user (ARPPU);

  • how many customers complete registration;

  • how many users reach the purchase page. 

When a startup’s unit economics is positive, the company should also:

  • track how many customers activate within a product (start using it);

  • track how many reach the AHA moment (fully understanding the product’s value);

  • track how many return (continue using it) and churn (stop using it);

  • maximize revenue from paying users;

  • grow its customer base.

Some companies at this point set their North Star, a key performance metric that correlates with revenue. For instance, for Airbnb it’s the number of bookings; for Facebook it’s the number of active users.

No in-house analyst? Get insights on your own 

Early-stage companies often don’t have product analysts on their teams. Here are three ways to find insights on your own. 


Implement NPS surveys

After the first purchase through the Amazon app, you'll see a window with a question: "How likely are you to recommend our product or service to a friend or colleague, on a scale from 0 to 10?" In the space below, you can explain your choice. This is Amazon’s NPS survey.

NPS surveys are a great way to assess — at no cost — how satisfied customers are with your product. Based on answers, you can categorize users into detractors (score 0-6), passives (score 7-8), and promoters (score 9-10). 

Use this data to calculate Net Promoter Score by subtracting the percentage of detractors from the percentage of promoters.

Tip: Bain & Company, the founder of the NPS system, suggests that a score above 20 is considered good, 50 is excellent, and 80 is stellar.

Analyze customer churn

A startup Х discovers that 60% of its customers leave its platform after the first month of use. This percentage represents the customer churn rate (CCR), indicating the number of users who stop using a product or service. 

Here's how to study customer churn:

  1. Identify customers who stopped using your product (like those who used your platform for six months but at some point stopped). 

  2. Exclude deactivated accounts and one-time users.

  3. Send your ex-users a survey, asking about issues with the product, or analyze their customer journeys. Inquire in the survey if they switched to your competitors’ services and why.

  4. Classify reasons for churn. With this data, you can also calculate the percentage of customers who might renew their subscription if you fix the problems.

  5. Brainstorm ways to improve the product.

Tip: It's better to study customer churn by keeping data in internal tools (MySQL, Snowflake, etc.). They provide more accurate information about users than external tools.

Examine drop-offs

Scenario: Out of all users who downloaded your app, 80% registered but only 30% actually used the product. This implies that the biggest drop occurs between registration and product use.

Here’s how to reduce user drop-offs: 

  1. Categorize customers who didn’t use the platform in data analytics tools.

  2. Find out what they did after the registration instead of completing the target action. 

  3. Build hypotheses on why they dropped and conduct A/B tests (you can also send them an NPS survey). 

Tip: External tools, such as Amplitude or Mixpanel, often enable you to upload the data of dropped-out users and analyze their actions step by step. 

But data tells you facts, not reasons. If you can’t understand why users act in a certain way, ask them for feedback.

Liliia Lutsenko is a senior product analyst at Wise and an advisor for the F1V portfolio. Watch her webinar “Numbers that matter: Product analytics for startups” via this link.

Subscribe to our LinkedIn newsletter to get one useful article every week via email and on LinkedIn. Each story can be read in under 5 minutes, enabling you to promptly implement insights into your business.

Photo by Kees Streefkerk on Unsplash

bottom of page