Shiju Mathew

The Google APM (Associate Product Manager) mafia has raised over $800 million in total funding. In 2002, Marissa Mayer created Google’s Associate Product Manager program to train future product leaders. With an acceptance rate of less than 1%, this program is highly prestigious and it gave early career PMs the responsibility to work on important projects where their contributions made a big difference. The nature of the role - high responsibility and high impact shaped them into exactly the kind of people who go on to build companies: 1. Early ownership on core products: They led product work on important products such as Search and Maps. This meant being responsible for prioritization, collaborating with engineering, and making shipping decisions on products used by hundreds of millions of users. They developed strong product judgment under real conditions. 2. Experience in cross functional management: APMs had to work with engineers, designers, legal, marketing, and execs and align all the different teams towards one goal. 3. They had to learn and adapt quickly: Over the two year program, APMs worked on two completely different teams building different products each year. This context-switching forced them to learn and adapt faster and build frameworks they could reuse anywhere. 4. Direct mentorship from experienced leaders: The program was small and tightly run. Early APMs were mentored by Marissa Mayer and other senior product leaders. These folks in their early 20s got feedback on product thinking, communication and leadership directly from people who had scaled Google. Here are the top 10 most funded former Google APMs turned entrepreneurs: 1. Asana: Co-founded by Justin Rosenstein Funding Raised: $453M 2. Sierra: Founded by Bret Taylor and Clay Bavor Funding Raised: $285M 3. Pepper: Co-founded by Chetan Narain Funding Raised: $46M 4. Limitless: Co-founded by Dan Siroker Funding Raised: $33M 5. Ambrook: Co-founded by Dan Schlosser Funding Raised: $29M 6. Mem: Founded by Kevin Moody Funding Raised: $29M 7. PapayaMobile Inc.: Co-founded by Si Shen papaya Funding Raised: $22M 8. Warmly,: Co-founded by maximus greenwald Funding Raised: $17M 9. Kapwing: Co-founded by Julia Enthoven Funding Raised: $12M 10. Create: Founded by Dhruv Amin Funding Raised: $8.5M Data from Crustdata (YC F24)'s people and company APIs. Did we miss out anyone?

322

14 Comments

Recently had a chance to sit with Joy Banerjee, VP Product Design at Zomato and it was one of the most informative power packed conversations I’ve ever had about product, design, human behaviour and everything in between! What started as a 60 mins conversation on a Sunday, extended far beyond 5 hours talking about how the product/design landscape is changing today, learning from his 2 decade plus of experience building products and shaping behaviours for 100s of millions of users. He had a refreshing take on how design as a field has been such native to India [think Indian architecture] and hence it shouldn't be looked with a myopic view of only optimising funnels. Its a much larger responsibility and learning from nature should help us design behaviours that sustainable and human-centric. Some of the key insights we discuss on how to build products : ✅ The 3C Framework — Choice, Control & Clarity Indian consumers appreciate product flows that give them enough information to feel confident in making high-value purchases. At Cleartrip, reducing friction actually reduced bookings because users wanted to double-check details. Joy shared - give users choice, control, and clarity to help them make decisions. That’s what creates delight for high-value actions. ✅ Honesty is Still the Best Policy! Cancelled orders were leading to massive food wastage at Zomato. Joy’s team introduced the now-famous “Food Rescue” feature - openly telling users they were rescuing cancelled meals. Earlier there was massive apprehension how users would take it however the result: huge reduction in food waste and users loved the chance to be part of something noble and meaningful. ✅ Design Shapes Human Behaviour — So Shape It Right Features like the ‘Like’ button or ‘Order Now’ have permanently shaped how people behave online. Joy reminded us that every design decision has the power to shape behaviour forever - and we must be intentional about the why behind every choice. ✅ Live Your Users Understanding how Indian users engage means living their behaviours - not just surveying them. Joy highlighted shifts like voice-first usage (voice notes for delivery partners) and more visual, prominent cues as patterns that emerged only by closely observing users. Great design comes from living your users and then designing for them! Joy also touched about Shape coding and how this interesting field led to so many physical product becoming successful - by mimicking how humans were already behaving in the world. Thanks to Rati Murti and Team for organising this. Link to both the episodes in comments. #productdesign #product #design #psychology

140

19 Comments

This is it. This is the conversation every leadership team needs to be having right now. "The Orchestration Graph" by WRITER product leader Matan-Paul Shetrit linked in comments is a must-read. The primary constraint on business is no longer execution. It's supervision. For a century, we built companies to overcome the high cost of getting things done. We built hierarchies, departments, and complex processes — all to manage labor-intensive execution. That era is over. With AI agents, execution is becoming abundant, on-demand, and programmatic. The new bottleneck is our ability to direct, govern, and orchestrate this immense new capacity. The firm is evolving from a factory into an "operating system." Your ORG CHART is no longer the map. The real map is the Orchestration Graph: the dynamic, software-defined network of humans, models, and agents that actually does the work. This isn't just a new tool or a productivity hack. It's a fundamental rewiring of the enterprise. It demands we rethink everything: Structure: How do we manage systems, not just people? Strategy: What work do we insource to our agentic "OS" versus outsource to models-as-a-service? Metrics: Are we still measuring human activity, or are we measuring system throughput and intelligence? This is the WRITER call to arms: The companies that win won't just adopt AI; they will restructure themselves around it. They will build their own Orchestration Graph, with governance and institutional memory at the core. They will treat AI not as a feature, but as the new foundation. At WRITER, this is the future we are building every single day — giving companies the platform to create their own secure, governed, and intelligent orchestration layer. The time to act is now. Read the article. Start the conversation with your leaders. And begin rewiring your firm.

923

91 Comments

🚀 Benchmarks vs Evaluations vs Metrics in LLMs – What’s the Difference? If you’re building or working with AI products, chances are you’ve heard these terms being used interchangeably. But here’s the clarity we all need: 🔹 Evaluations → Think of them as real-world report cards. They test how well your LLM performs in practice – not just in labs. Example: Does your Uber AI Assistant answer “How do I change pickup location?” correctly and quickly? 🔹 Benchmarks → Standardized exam papers used to compare models. Example: Testing GPT-4 vs your fine-tuned assistant on 100 Uber FAQs. 🔹 Metrics → The scores on those tests. Accuracy, F1, BLEU, Latency, CSAT… numbers that tell you how good (or bad) your model really is. 👉 Put simply: Benchmarks = What exam you take Evaluations = How you test in real-world Metrics = The score you get Difference in Benchmark, Evals … 💡 Lesson for PMs & AI Builders: Don’t just chase higher benchmark numbers. Focus on evaluations that reflect user trust, satisfaction, and business impact. That’s where real product wins come from. 📌 I’ll also be sharing detailed guides for each of these as we go along — stay tuned! #AIPM #AI #LLM #Evals #Benchmarks #Metrics #ProductManagement

82

3 Comments

8 questions asked in Walmart PM interview [1] Design a guest checkout feature for Walmart online. [2] Estimate the number of Uber rides. [3] Identify one metric for Meta (Facebook) Marketplace that the whole team should monitor daily. [4] What is the market size of toilet paper in the US? [5] You are a PM for an e-commerce company. What are the metrics you will track as PM? [6] How would you tackle the challenge of fake reviews on platforms like Amazon and Flipkart? [7] How can Google help improve the grocery shopping experience of users? [8] If you are opening a new Walmart branch, how would you decide how many cash registers are needed for the store? Which one did you like the most? Follow Vishal Bagla for more insights on product management. #ProductManagers #ProductManagement #Jobs #InterviewPrep #Interviews

170

16 Comments

A great PRD doesn’t just describe the product. A great PRD makes the team happy. Because here’s the truth: every teammate reads a PRD with a different lens 👇 🔹 Sales → They’re looking for use cases. How will this feature solve customer problems? 🔹 QA → They need use cases plus exceptions. What’s in scope, what’s out, where could things break? 🔹 Developers → They want technical details. Data mapping, flows, edge cases that guide implementation. 🔹 Project Manager → They scan for status signals. Change logs, priorities, and dependencies. 🔹 Design → They need quick access to existing design links. Mockups, flows, and style guides. One PRD. Many perspectives. The magic is writing it so everyone finds what they need without frustration. Pro tips: > Start with a template and validate it with your team → so everyone knows exactly where to find what they need. > Keep it alive → update continuously, don’t let it become a stale doc. > Your PRD should be clear not only for the team you have today, but also for future teammates who join tomorrow. Bonus: Don’t forget accessibility. A great PRD doesn’t just cover features, flows, and use cases, it also ensures the product can be used by everyone, including people with disabilities. Build it in from the start, not as an afterthought. Because a PRD isn’t just documentation, It’s a trust-building tool. And trust = faster delivery + fewer surprises. 👉 What’s the one thing you always check first in a PRD? 🙋‍♀️ I’m Caroline, a PM who believes clarity fuels collaboration. I share frameworks and breakdowns to help teams align, simplify, and deliver with less chaos. #ProductManagement #ClarityOverChaos #TeamAlignment #Agile #PMGrowth

46

33 Comments

Last week, I had an eye-opening conversation with the Product Head of a fast-growing AI-powered design platform from Europe. They'd just launched in India, and honestly, I was impressed: ⚫ Intuitive product with world-class UX ⚫ Strong funnel metrics ⚫ Solid user engagement But here's what was alarming: their adoption numbers were terrible. "We can't figure it out," they told me. "People sign up, but so many just drop off at the final step. Maybe Indian customers aren't ready for products like ours yet?" Then we looked at their payment data. Success rate: Just 62% The culprit? They were using the same payment setup that worked perfectly in Europe: card-only payments. Here's the problem: In India, card penetration is under 12%. Most people pay seamlessly via UPI. This team had mastered the hard parts—they built a product people genuinely loved. But 4 out of 10 customers were hitting a wall at checkout. The fix was simple: We rebuilt their payment flow to prioritize local payment methods. The results? ⚫ Success rates jumped within weeks ⚫ Renewals followed suit ⚫ India transformed from a "tough market" to one of their fastest-growing regions The takeaway: You can have the best product, the smartest GTM strategy, and brilliant marketing - but in India, if you don't solve for payments, you're scaling with a leaky bucket. This is exactly why we built Razorpay the way we did - to help global companies succeed in India from day one. P.S. If you're interested to learn more, check out https://lnkd.in/g_tvdYUK Also, you can go through the case study of how we helped a global giant like Hostinger arrive and thrive in India: https://lnkd.in/gfWuRERe Paulius Lapėnas Harshit Shrivastava Ritwik Kadu Tejas Gowda Akanksha Bawa Akshith Rao Rishi Kumar

81

AI has a product problem. Not a model problem. Models are making capability leaps every few weeks but AI-native product innovation hasn’t kept up. Most products are forcing AI into existing UX patterns rather than rethinking an AI-native experience from first principles. Parallels to early mobile (2007-2010) - for years, mobile meant shrinking your website into your phone until Uber reimagined transportation. To be clear, there are lots of amazing AI-first products. See NotebookLM, Lovable, Stitch, Flow. But the pace of innovation on models has been faster, for a lot of reasons which I will explore in future posts.

552

48 Comments

Breakthroughs in product and the way teams work rarely come from pushing harder. They come from seeing the problem differently. A new way of representing the situation can turn confusion into clarity. It’s not just about raw knowledge; it’s about how that knowledge is structured, surfaced, and made usable. Computer science has long known this: the right data structure turns a brute-force problem into an elegant one. Same information, different organization, exponentially different results. Think of it like a city with no street signs or maps. Every building, every service, every shortcut is there, but unless you know how to navigate it, you’re lost. You waste time. You make wrong turns. You miss what’s right in front of you. I’ve seen this often: same customer data, wildly different takeaways. One team sees a funnel to optimize; another sees a journey to understand. A backlog that looks like noise to one team becomes a signal for product strategy when visualized as a question that unlocks new customer insights. Good representations are what turn potential into momentum. They’re the difference between theory and action. The question isn’t whether you have the right information. It’s whether you’re representing it in a way that makes the next step obvious.

63

10 Comments