Author
Ansh Agrawal (page 2)
I help startups set up analytics, understand user behaviour, and gain actionable insights to drive product growth
Latest posts
Analytics Debt: The Silent Killer Nobody Talks About Until Series B
The short version: Analytics debt is the gap between the data you have and the data you'd need to answer the questions you're about to be asked. It accumulates in five predictable stages, and the damage peaks at Stage 3 — when the company believes it's data-driven and the dashboards are built on broken events. You can audit yours in an afternoon with three checks and no new tools. Every startup I've worked with has analytics debt. Every single one. Not because they're bad at data. Because they
· 9 min read
D2C Playbook 1 -- The Data Foundation: Fix Your Owned Layer Before You Scale Spend
The short version Here's the thing nobody tells you when you're about to scale spend. Every channel you're about to pour money into gives you data you can't fully trust. Quick commerce hands you almost nothing about who your customer is. Meta and Google grade their own homework and, surprise, they're doing great. The one place you actually own the full picture is the website and app you built. So that's the layer you have to get right first. And most brands have it quietly broken. Not broken i
· 17 min readGamification is a multiplier. But multiply by zero, and you get nothing.
Duolingo has 19 million daily active users. Foursquare, once held up as the gold standard of gamified apps with 50 million users at peak, quietly retired its own badges and points system before splitting into two separate apps. Both used streaks, badges, leaderboards but only one survived its own mechanics. What does the research on gamification actually say? Yu-kai Chou, whose Octalysis gamification framework has been adopted by LEGO, Microsoft, and Porsche, offers a blunt diagnostic he
· 4 min read
My Love-Hate Relationship With PostHog (And Why I Keep Recommending It Anyway)
The short version: PostHog's pure product analytics is genuinely weaker than Mixpanel's or Amplitude's, and complex reports push you into SQL. I still recommend it to most founders, because it bundles session replay, feature flags, experiments and a real data warehouse into one tool at a stage when standing those up separately costs more than the analytics depth you give up. I have a complicated relationship with PostHog. Always have. There are things about it that genuinely frustrate me. Thin
· 4 min readAnyIP: Helped get more visibility into marketing, and improving CAC, spend efficiency
In short: AnyIP didn't own its analytics and had little visibility into marketing performance. We cleaned and structured the raw data and built structured reporting on top. They gained visibility into marketing, improved CAC and spend efficiency. About AnyIP AnyIP is a high-performance proxy and networking infrastructure provider used by developers, marketers, and automation teams to route traffic reliably across the globe. With multiple marketing channels driving acquisition, the team needed
· 1 min readFRAI: Improved paid conversion rate (2x)
In short: FRAI had no clarity on which user segments converted or why others dropped off. We ran deep behavioural analysis, turned it into product recommendations, and designed the A/B tests. Paid conversion rate doubled, and they kept a repeatable experimentation process. About FRAI FRAI is an AI-driven product designed to help users clear interviews using the help of AI-interviewer for their live job interviews. As the product grew, the team needed clarity on user behaviour, conversion driv
· 1 min readTermplus: Setup the entire Data Analytics Infra, & unified user journey
In short: TermPlus had no clear analytics setup and fragmented data across the product. We ran a deep audit, wrote the tracking plan, and worked with their dev team to implement it. They ended up with a full analytics infrastructure and a unified user journey. About TermPlus TermPlus is a digital financial services platform offering streamlined insurance and banking workflows for users across Australia. As the product scaled, the team needed clearer visibility into user behaviour, product per
· 2 min readAnswering Agent: Helped create a call score using AI
In short: Answering Agent had no visibility into the quality of its AI calls. We studied the product and industry benchmarks, defined a full event set, and built an AI-generated call score. They could finally measure call quality directly. About Answering Agent Answering Agent is an AI-powered calling platform that automates customer outreach, follow-ups, and support conversations. As an AI-first product, the quality of each call directly impacts customer satisfaction, product trust, and conv
· 2 min readSol: Helped clean up complex data structure into a simple accurate setup
In short: Sol is a complex product that had ended up with tracking to match. We defined core event categories instead of granular events and moved the complexity into properties. The result was a simple, accurate setup that scales. About Sol Sol is a powerful productivity and collaboration app designed for individuals to dive into meditation, exercises, find new friends and more. With a fast-moving product roadmap and weekly feature releases, maintaining clean and consistent analytics was bec
· 2 min readWellnessCoach: Helped regain trust in data
In short: Wellness Coach was overtracking, with an inconsistent event structure that made the data unusable. We audited the setup and rebuilt a clean, simple tracking structure. The team could trust their numbers again. About Wellness Coach Wellness Coach is a B2B digital wellbeing platform that offers guided meditation, coaching sessions, habit programs, and enterprise wellness solutions. With millions of users engaging across mobile and web, having clean, reliable analytics is essential for
· 1 min readSpeedyloans: Reduced cost (-30%) and increased engagement (+1%)
In short: Speedyloans was sending millions of emails, driving up cost and putting domain reputation at risk, with 100+ user attributes and no prioritisation. We segmented and modelled it properly. Cost down 30%, engagement up 1%. About Speedyloans Speedyloans is an online lending platform that provides fast, accessible loan advances to users across the US. With a high-volume user base and millions of outbound communication events, optimizing engagement workflows and improving conversion targe
· 2 min readzeroone: Improved Week 2 retention (+15%), engagement (+30%), and implementation of new tracking
In short: Few zeroone users completed an artwork upload after onboarding. We moved uploads into onboarding as an optional step and cut unnecessary friction. Week 2 retention rose 15% and engagement 30%, with new tracking shipped alongside. About zeroone zeroone is a blockchain-powered social app where users upload, mint, and collect digital artwork. Designed for both creators and collectors, zeroone blends social engagement with crypto-native mechanics, making early activation and repeat usag
· 4 min read