App Usage Data Analysis

User behavior is more than what they say - it’s what they do. While surveys and usability tests provide valuable insights, log analysis reveals real interaction patterns, helping UX researchers make informed decisions based on data, not just assumptions. By analyzing interactions - clicks, page views, and session times - teams move beyond assumptions to data-driven decisions. Here are five key log analysis methods every UX researcher should know: 1. Clickstream Analysis - Mapping User Journeys Tracks how users navigate a product, highlighting where they drop off or backtrack. Helps refine navigation and improve user flows. 2. Session Analysis - Seeing UX Through the User’s Eyes Session replays reveal hesitation, rage clicks, and abandoned tasks. Helps pinpoint where and why users struggle. 3. Funnel Analysis - Identifying Drop-Off Points Tracks user progression through key workflows like onboarding or checkout, pinpointing exact steps causing drop-offs. 4. Anomaly Detection - Catching UX Issues Early Flags unexpected changes in user behavior, like sudden drops in engagement or error spikes, signaling potential UX problems. 5. Time-on-Task Analysis - Measuring Efficiency Tracks how long users take to complete actions. Longer times may indicate confusion, while shorter times can suggest disengagement.

People see product usage data as primarily for Product Managers and CSMs. But as our CRO I use it daily across my departments. Here's what I track, what I use it for and some of our stats from 2024: First, after having done this for 15+ years, it still shocks me how few buyers ask for usage statistics of current customers before they buy. There is a big assumption there that they will behave differently from their peers at other companies. We track more, but here are 5 core things that I care about: 1) # of Sessions over various time periods 2) # of Users/Account 3) Automation Features Active 4) Top Used Features 5) User Titles. Yours may vary depending on the product - some products no one logs in, it just works in the background and others are data analysis platforms, etc. We happen to have both so it's important to me that we've got automations working in the background that should be as well as people logging in for data analysis. Our North Star stat for the first 3 (the last two are informational and drive learnings for us): 1) # of Sessions: Very straightforward - I want at least 1/Day for primary account users. This tells me that in general, they need our product for daily work. 90% of our accounts have this, 94% are active weekly. The balance are probably onboarding customers in any given week. No notes. Our team, both product and support, has done an absolutely amazing job on this. We'll want the same in 2025. 2) # of Users/Account: I want at least 2-3 users that do the above. If we have that I feel pretty good, but always happy to have more that login intermittently. Our average account has 6 Users/Week. No notes again. We'll look to do the same in 2025. 3) Automation Features: We have features that automate daily work for ad platform optimizations - we have others but these are the most important to me. I want to see them in use for at least the core ad platforms our customers spend 90% of their budget on - that's LinkedIn & Google for 2024. We just launched this in 2024. We had 70% that met this and my goal is 95% in 2025. And here are uses for each department I manage now or have in the past: 1) Marketing: Find case studies/testimonials, find your top titles for audiences, find/expand on ICP Accounts, find which campaigns to push based on which feature gets the most usage. There is so much here that marketing teams rarely take advantage of. 2) New Business: Same as above and they probably use it the least. Find reference customers, find the top users w/ titles that match your buyer. Incorporate most used features into your pitch. They can refer you new accounts. You will instantly become a better seller if you use this data. 3) Customer Success: Most obviously renewals, with usage being a good gauge for likelihood to renew. But top users are also most likely to upsell! Would love to hear how everyone else is using this data - I'm scratching the surface here!

🌟 Day 226: 📱 Mobile Device Usage and User Behavior Dataset Today, I delved into the Mobile Device Usage and User Behavior Dataset from Kaggle, focusing on data analysis and visualization using Python. This project provided an opportunity to explore user behavior patterns and their relationship with mobile device usage. Here are the key accomplishments: - Data Preparation: I began by cleaning the dataset, addressing missing values, and renaming columns for clarity. This ensured that the data was ready for analysis. - Exploratory Data Analysis (EDA):  - Analyzed various attributes such as app usage time, screen on time, and battery drain to identify trends across different demographics.  - Created visualizations to explore correlations between features, enhancing understanding of user behavior. - Advanced Visualizations:  - Developed heatmaps, scatter plots, and box plots to illustrate relationships among variables.  - Utilized pair plots to visualize interactions between features segmented by user behavior classes. - Classification Modeling: Implemented several classification algorithms (Logistic Regression, Random Forest, SVC) to predict user behavior based on device usage patterns. Evaluated model performance using accuracy scores and confusion matrices. - Regression Analysis: Trained regression models to predict battery drain based on user behavior data. Metrics such as Mean Squared Error and R² Score were calculated to assess model effectiveness. This project significantly enhanced my skills in data analysis and visualization techniques while providing insights into mobile device usage trends. The experience highlighted the importance of exploratory analysis in uncovering meaningful patterns within datasets. Check out the full project on GitHub: https://lnkd.in/eqY_7e9k #Python #DataAnalysis #MobileUsage #UserBehavior #Kaggle #DataScience #Visualization #LearningJourney 🚀 Looking forward to continuing this exciting exploration of data!

Struggles of doing data science in the real world 🤦: What do you do when there’s no A/B test but you still need insights? I recently faced that challenge (again): 👉 The growth team asked me to evaluate the impact of a new mobile app feature on conversions (a week after it launched) In the real world, data is messy, and A/B tests aren’t always an option. As a Data Scientist, you need to learn to be resourceful Here’s how I approached it: 1️⃣ Segmented analysis: I created pre- and post-launch groups based on user signup dates. 2️⃣ Exploratory data analysis (EDA): Visualized conversion trends, layering in cohort and seasonal comparisons. 3️⃣ Statistical testing: Ran an independent t-test to validate observed changes, carefully checking assumptions like normality and variance equality. Result? A clear signal of increased conversions on iOS, while Android showed minimal impact. 💡 Key takeaway: T-tests (or similar methods) can still deliver actionable insights outside traditional A/B testing, but validating assumptions and adding context is critical to making reliable conclusions. I broke down my full workflow and the lessons learned in my latest newsletter article (If you’re curious, check the link in the comments👇) What’s your go-to method for analyzing feature impacts without a perfect experimental setup?

How does "Product Analytics" Work? With a story, let's understand the fundamentals behind popular tools like Mixpanel, Amplitude, Google Analytics, Fullstory, MoEngage etc. Story: The other day I wanted to order Noodles 🍜 for myself and like most of you I headed to our beloved app Swiggy (sorry Zomato) Here is the journey that I followed: 1️⃣ Open the app 2️⃣ Search for Noodles 3️⃣ Click on one of the restaurants 4️⃣ Add to cart 5️⃣ Apply Coupon 6️⃣ Make Payment 7️⃣ (Im)Patiently wait for the order To make my experience better ❤️ & To Get me to spend more money 🤑, the PMs at Swiggy need this data. The right analytical tools will help the Swiggy team: ⤴ Capture these interactions 📦 Store these interactions ✨ Create reports from these interactions 💡 Take the right action The developers embed a code in the app, which sends the important interactions (app open, clicks, screen change) to these tools. These tools process and store the data at their end. These tools give you a friendly interface to get simple and advanced reports like: 👉🏽 Funnel - How people are moving from one step in journey to another. PMs can Discover any blockers and drop-offs. 👉🏽 Trends - How feature usage is growing or decreasing. 👉🏽 Segments - Creating and analysing various user groups based on behaviour, demographics etc. 👉🏽 Cohort Analysis - Are people engaged in the app, retention quality. and many more. The interactions are sent in the form of events. Each event contains a name and properties to contain any additional information about the event. For example: Button Click event might contains properties like text on button, operation system, app version etc. Similarly user information is also sent and store to relate events with particular users. User information is also arranged in the form of properties, for example: A user can have name, location, favorite restaurant etc. While tools like Mixpanel, MoEngage and other provide out of box solutions for product analytics, many capable teams also choose to do this by their own by following stack: 1. Event Capture from app: Plain Javascript, React, Angular etc. 2. Event Storage and Streaming: MYSQL, Mongo, Kafka etc. 3. Visualization: PowerBI, Redash, Metabase etc. Thats a 101 on how product analytics tools work. Want to know more? I am hosting a Free Live Session on "Analytics for Product Managers" this Saturday, Register here: https://lnkd.in/g-9VuNS4 See you! #TechMadeEasy #productmanagement