Analytics & performance
Where the product loses users and performance
Before the team decides what to fix first, it should know where the product is actually losing users. We analyse the metrics of digital products, mobile apps and AI features, pinpoint bottlenecks and recommend specific changes.
Written by Radosław Kołacki Updated
Trusted by
When is product analytics worth it?
The product is seeing drops or downtime. Churn is rising, NPS is falling, or the team does not know where users leave.
An optimisation plan the Product Owner takes to the management board is easier to defend with numbers than with a hunch: first we establish what is actually happening, then what to fix first.
Who is product analytics for?
- Churn is rising and the team wants to understand why customers leave and what drives retention
- Before optimising, the Product Owner or Head of Product wants to know what to fix first
- The performance of a product, mobile app or AI feature needs measuring
- In medtech and health products, performance and retention translate into cost and risk
How it works
How does a product analysis work step by step?
01
We measure
We connect to the data the product already collects, for example in Google Analytics and Hotjar. We check that tracking is complete and fill the gaps before drawing any conclusions.
02
We analyse
We look at retention and churn, conversion on key journeys and the points where users abandon a process. We watch session recordings and click maps to understand why it happens.
03
We pinpoint bottlenecks
We measure Core Web Vitals, Google's indicators of loading speed and page stability, in PageSpeed Insights. We identify the funnel stages where most users leave and the screens where the product slows down.
04
We recommend
We rate each change for impact and order of implementation, and split the whole into quick fixes and long-term actions. We go through the list with the team.
After the analysis it is clear where the product loses users and performance, and which change to start with. The result is conclusions, not raw numbers.
How much does product analytics cost?
from PLN 2,500 net
A one-off analysis costs from PLN 2,500 net. We run analytics in two modes. A one-off analysis has its scope and price agreed before the start. Ongoing monitoring is billed monthly. We prepare the estimate after a conversation about the data the product already collects.
The conversation about the data the product already collects and the initial estimate for the analysis are non-binding. We come back with an answer within a few hours.
„We wholeheartedly recommend Squares. We have worked together for years and they are one of those partners you can rely on 100%. Among many different projects, they built our company website with an advanced integration with an external database. They do not stop at the code.”
Inprogress - sukces w zarządzaniu
Google review, translated from Polish · Full review
Work
How it works in practice
Inprogress: a data-based website audit
For an IT training company we combined Google Analytics and Hotjar data, session recordings and performance tests into one prioritised report.
Case study
Daily Effect: measuring conversations with the AI assistant
We measure answer effectiveness, topics and sentiment of conversations. Over more than 6 months the assistant resolved 72.7% of issues without a human agent.
Case study
pompy.app: a metrics dashboard for the team
The dashboard shows the current number of quote requests, the most viewed models and comparison statistics on a site with a database of more than 600 heat pumps.
Case study
Why Squares
We combine analytics with product work: we say what the numbers mean and what to do about them. The same team can then design and implement the changes.
Questions
Frequently asked questions
What is churn rate and how is it measured?
Churn rate is the share of customers or users who stopped using a product in a given period. It is calculated by dividing the number of customers lost by the number of customers at the start of the period. The figure alone does not say why they leave, so the analysis combines churn and retention with conversion on key paths, session recordings and click maps.
How to find out where a product loses users?
The place where a product loses users shows in the funnel: at the stages where most people drop out. We analyse data from Google Analytics and Hotjar and conversion on key paths, while session recordings and click maps show the cause. The result is the stages and screens with the biggest losses, and changes listed in order of implementation.
What is NPS and what does a falling score mean?
NPS (Net Promoter Score) is a loyalty measure: the share of customers who would recommend the product minus the share of detractors. A falling NPS signals a problem but does not locate it. The analysis sets it against churn, retention and user behaviour on key paths to establish what is really happening and what to fix first.
What is the result of a product analysis?
The result is a report with conclusions and a list of changes in the order they are worth implementing. Each issue has a description and an impact rating, and the whole list is split into quick fixes and long-term work. We go through it with the team. The same Squares team can then design and build the changes.
How much does a product data analysis cost?
A one-off analysis is priced up front: the scope and price are agreed before the start, after a conversation about the data the product already collects. Ongoing monitoring is billed monthly. We give the initial estimate after a conversation, so the figure is known before any decision to work together.
What are Core Web Vitals and how are they measured?
Core Web Vitals are Google's metrics for loading speed and page stability. We measure them in PageSpeed Insights and point to the screens where the product slows down. The results go into the same list of changes as the funnel findings, with an impact rating and an order of implementation.
How is Analytics & performance different from UX monitoring & optimisation?
Analytics & performance answers where a product loses users and performance, and ends with a report and recommendations. UX monitoring & optimisation is continuous work: every month we observe user behaviour, design and implement small changes, then check their effect in the data. The analysis is a diagnosis; monitoring is a monthly cycle of improvements.
What data and access are needed for the analysis?
The analysis needs access to the data the product already collects, above all in Google Analytics and Hotjar. We first check whether measurement is complete and fill the gaps before drawing conclusions. Squares is an official Hotjar partner, and performance is measured in PageSpeed Insights against Core Web Vitals.
Does the analysis cover mobile apps and AI features?
Yes. We analyse websites and web apps, mobile apps and AI-based features. For AI assistants we measure, among other things, answer effectiveness, topics and sentiment of conversations. In the Daily Effect app this measurement showed that the assistant resolved 72.7% of cases without a human agent over more than 6 months.
Can analytics be an ongoing service?
Yes. Analytics runs in two modes: a one-off analysis with scope and price agreed before the start, or ongoing monitoring on a monthly subscription. In the ongoing mode we review the data regularly and update the list of changes, so the optimisation plan rests on current numbers.