Product Strategy Decoded Podcast

Mike Goitein

Every week, I pull back the curtain and make strategy easier using a proven framework. You'll get insights of what works today and why, with powerful takeaways you can apply to your own business. michaelgoitein.substack.com

Episodes

  1. The Art and Science of Pivots – Livestream with VC Chris Tottman

    11/19/2025

    The Art and Science of Pivots – Livestream with VC Chris Tottman

    A Conversation with Chris Tottman on the Strategic Pivots That Built Category-Defining Companies Chris Tottman is a partner at Notion.vc and has spent over 30 years building and investing in technology companies. He started his career trading options on the London Stock Exchange floor before founding multiple companies, including MessageLabs, which became Europe’s first major SaaS business before its acquisition by Symantec. Today, he invests in early-stage technology businesses while maintaining what he calls “a constant itch to create, collaborate, and work with people to help them build the best companies they can.” Key Takeaways On Discovery vs. Proving Yourself Right Great founders don’t over-index on proving their original idea. They stay in constant discovery mode, asking what problems keep customers awake at night and what users are actually doing with the product. On Finding Problems Worth Solving Look for pain that makes someone bang the desk. Pain that makes them think, “If I don’t solve this, I might get fired.” That’s where innovation and the wave of adoption comes from. On Market Definition “Uncomfortably narrow” is the goal. The smaller the market definition, the easier it is to target. Wide definitions like “enterprise, SME, global” become impossible to market effectively. On Official Strategy vs. Revealed Strategy Your official strategy is what you tell investors and put on your roadmap. Your revealed strategy is what your usage data shows customers actually value. The gap between those two is where billion-dollar pivots happen. On the Discovery Process Whiteboard out the whole value chain of a product, looking at where time goes and where money goes. Run B2B and B2C research. Have multiple minds working on it, asking “What if this is true? Is that also true?” That’s how you find the biggest, most pressing problems worth solving. The Difference Between Good Founders and Great Founders Michael Goitein: Chris, we started talking about pivots and that notion that when a company starts, they might have a vision, an idea. But what they ultimately become successful for can be miles from what they started out as. You have these incredible stories that I wanted you to share. Chris Tottman: I think the good thing about the stories is that they show two different ways of trying to figure out what the company can potentially become and how to bridge from where you are today to a bigger version of the world. What we find at Notion is that the difference between good founders and great founders is that great founders don’t try to prove themselves right to the end of the world. They have an idea and take it to market, but they don’t spend their time trying to prove that original idea is the killer idea. Even if customers are buying it. What they do instead is stay constantly in discovery mode. They keep asking questions about what problems keep customers awake at night. What are users actually doing? What are they really getting value from? From ISP to Europe’s First Massive SaaS Business Chris Tottman: In the early ‘90s, I was working with a couple of my closest friends who were building the IP networks. They had an ISP with a proposition for SMEs, connecting companies to the internet. To give you context on how early this was: the first two-meg leased line we sold cost £200,000 a year in 1995. Two meg. Companies were struggling just to connect a local area network within their building to the outside world. We built technology so somebody could just plug in a box that had a mail server, some security, and web space all bundled together. Super easy. It was flying off the shelves. We had a tremendously successful company. But here’s the key lesson. We didn’t just sell to customers. We spent significant time in exploration mode, really asking deeper questions about what was keeping them awake at night. What’s stopping them from being successful? How do they lose against competitors? Three or four things started to come up. One was more sophisticated technology around websites. Another was blacklisting certain sites. And third on the list, almost as an afterthought, customers said: “You keep delivering us viruses.” Two things were happening. I could see the engineers, just through their body language, really itching to get to that third item. And it wasn’t the first thing on the list. It wasn’t the obvious thing. One of the founders said, “Is this a little bit like clean water? Back before critical infrastructure, how did you get clean water? You boiled it yourself. But now that’s all done centrally. The internet is the same. We have all this processing power. Since we’re responsible for delivering mail to our customers, should we be responsible for cleaning it?” We decided to back that idea. We created an alpha over a few months, then switched this technology on so we could route our customers’ traffic through this security tower we’d built. One of the engineers had put a bell on it so that every time we caught something nasty, the bell went off. When we switched it on, a shriek emitted out of the terminal. We didn’t know what it was. It was horrendous. We switched the terminal off. Then the poor engineer realized it was his fault. Every millisecond, we were catching something bad. We knew immediately that we’d hit this internet oil. This was a real gusher of a proof point. That created a course of events where we had to internalize: what do we do about this? Because every email user in the world has this problem. Is it B2B? Is it B2C? We created an entirely new company called MessageLabs. It became the first major SaaS business in Europe. We reached $150 million in annual revenue within about eight years and sold to Symantec. But it came from that process. When you’re with customers, don’t just sell your technology. Spend significant time in exploration mode, really understanding what’s going on in their minds. If you’re doing a really great job for customers, you probably have two or three businesses you could build on the back of that insider understanding. From Social Postcards to Direct Booking Economics Chris Tottman: This was around 2013, the early advent of social channels. The founders came to me with an idea in the hotel space. Wouldn’t it be great if hotel brands could use check-in and check-out moments to get user-generated content? They would provide guests with a digital postcard of their stay that guests could post on social media. My first meeting with this founder was in a beautiful London square. I like walking meetings. He was really compelling, but I didn’t like the proposition. I was thinking of the loop: I’m checking in, I’m checking out, I do this thing, I put it on social media, some random person sees it in my network. Are they likely to want to go to that destination? It felt too weak. Too far away from a transaction. But the founder was compelling. And he mentioned something intriguing: online booking agents like Booking.com and Kayak were starting to take significant percentages of booking revenue. As I walked back, I kept thinking about it. Put yourself in the position of a young entrepreneur in the hotel space. You finance the building. You spend millions fitting it out. You assemble an incredible workforce. You open up, and you haven’t even taken a dollar in revenue yet. The CAC payback must be three to four years. And then 15 to 20% of all revenue goes to a digital booking engine? There was something there. So we invited them into the office for whiteboard sessions. We mapped out the whole value chain of the industry. The journey someone goes through when booking a holiday. We looked at where time goes and where money goes. They came with analysis showing that if you’re a small hotel with no purchasing power, booking engines were taking over 30 to 40% of booking revenue. Even big chains were giving up 15 to 20%. And the numbers were getting worse every year. Then they did B2B and B2C customer research. What they found was fascinating. 65% of consumers had the hotel website open because it had better photographs, more detail. They felt more kinship with the hotel itself. But those same consumers also thought the online travel agencies (“OTA”) booking engine was cheaper. The analysis showed that wasn’t necessarily true. And hotels could dynamically price against what was being offered online. So they conceived the product: a pop-up on the hotel website showing three OTA prices alongside the hotel price, which was never more than the OTAs. A “never priced better” approach. I gave them £250,000 to build that product. We launched three months later. It went wildfire. The pent-up demand from hotels wanting to drive consumers to book direct was enormous. In the first two months, we sold to 250 to 350 hotels. Within a couple of years, we had about 1,000 at the mid-market, plus two of the eight biggest hotel chains. The technology itself wasn’t thick enough initially. There wasn’t a big enough moat. So we knew we had to drive the price down to create more demand, more buyers, and then create more sophistication in the product. But it was so interesting to see that process of mapping out the whole industry, having four or five different minds working on it, doing the analysis purely in pursuit of looking for a big enough problem, and trying to conceive a product that might solve the biggest, most pressing problem for those buyers. That’s probably a $20 million revenue business today, serving most of the biggest hotels on the planet. The Mindset That Matters Michael Goitein: Just to toss a wrench into the conversation, we got a question about product-market fit. Once you’re starting to see traction, how do you drive toward it? Chris Tottman: That’s the silver bullet question. We all argue about what product-market fit even is. It’s not something particularly stable. The hyp

    52 min
  2. Video: How AI Strategy is Different

    11/14/2025

    Video: How AI Strategy is Different

    Thanks to everyone in the Substack community who joined the livestream! Click here to see this post as a Gamma presentation. Join me for my next live video in the app on November 19, 2025, at 12:00pm Eastern time with VC and startup thought leader Chris Tottman of the brilliant “Founder’s Corner” newsletter. Here’s the full talk as a refined transcript: Over the past six months, we’ve analyzed how leading companies are approaching AI within their product strategies. Rather than adding features or chasing every new model release, these four companies made fundamentally different choices about how to embed AI into their core capabilities. Their strategies reveal something crucial: AI is never the strategy itself. It becomes the capability that enables your strategic choices. Four Different Paths to Competitive Advantage Otter: Compounding Value Through Proprietary Infrastructure Otter approached AI differently from the start by building their entire stack from scratch rather than relying on publicly available models. They identified a specific constraint: their users operate in professional meetings across remote-first environments, often attending multiple meetings per week. Their strategic choice centered on making transcript and insight sharing the default. A single user attending multiple meetings might extract moderate value, but after 100 meetings, something shifts. Users can search across all conversations to find patterns, ask questions across meetings, and develop deeper organizational knowledge. This radiated value touches roughly 7.5 people per conversation. The system becomes more valuable the more it’s used, creating a compounding competitive advantage that scales with adoption. Granola: Privacy as a Strategic Differentiator While Otter competed in the transcription space, Granola observed a critical gap: existing solutions created friction in sensitive conversations. Otter requires a visible bot in the meeting. Zoom demands explicit recording announcements. These create barriers to using the product where it matters most, in C-suite and venture capital conversations where confidentiality is paramount. Granola made an inverse strategic choice. They built AI processing directly on the device and deleted the recording immediately after, making the AI completely invisible to other meeting participants. This design eliminated the consent requirement and allowed them to achieve SOC 2 Type II compliance, opening access to markets where any leak could tank stock prices. Where Otter asked how to create compounding value through sharing, Granola asked how to unlock value by preserving privacy. The strategic difference determined everything else. Gamma: Building Flexibility Into Infrastructure Gamma’s origin story offers a different lesson. The company started with a conventional idea: bringing PowerPoint online. Running out of runway forced a strategic bet. The team decided to stake their future on AI, but with a critical distinction: they built AI as an orchestration layer sitting above other models rather than committing to any single model. In three months, they mastered something most companies don’t attempt: the ability to call different models for different components of a single task. When creating a presentation, their system analyzes which models deliver the strongest outlines, which produce the best graphics, and which format content most effectively. When a model shows weakness, they can swap in an alternative without disrupting the user experience. As new models emerge, they’re not vulnerable to disruption. They simply integrate the new capability. Gamma targeted creators, knowledge workers, and educators who needed to sell ideas. Instead of replicating PowerPoint’s interface, they reconceived how people present information using cards created on the fly and continuously improved through model comparison. Their breakthrough wasn’t an individual feature. It was the infrastructure that allows them to choose the best tool for each specific task. Shopify: AI as Infrastructure for Acceleration Shopify took yet another approach. Rather than building AI as a visible feature or proprietary capability, they embedded AI deep into their infrastructure layer. The distinction matters: they don’t use AI to automate or reproduce manual tasks. They rethink linear processes into non-linear, continuous workflows from the ground up. Store setup, which previously required days, compresses to hours. SKU imports accelerate. The value compounds through merchant velocity, not through visible feature count. Shopify targets growth-stage merchants who represent a structurally different opportunity than Amazon Marketplace sellers. These merchants never face displacement by Shopify itself, because Shopify operates in the enablement space. The value flows through a partner ecosystem that provides capabilities Shopify could never build alone. AI disappears into the background, functioning as a speed multiplier that serves a specific market with specific constraints. The Underlying Pattern: Strategic Constraints Drive Design Each company confronted a different set of constraints that forced their innovation: Granola couldn’t use a visible bot, so they created invisible on-device processing. This constraint became their differentiator. Otter focused on English-first transcription at 95% accuracy rather than pursuing 100% accuracy or supporting 50 languages. Accepting this limit allowed them to ship and iterate. Gamma rejected building a proprietary model entirely. This constraint forced them toward orchestration, which became their core capability. Shopify approached AI not as a feature factory but as an acceleration layer for existing workflows. The pattern suggests something important: when you work within constraints rather than against them, your strategic choices become clearer. Your capabilities become more differentiated. Your competitive advantage becomes harder to replicate. What Changed in AI Strategy The companies we analyzed didn’t ask traditional questions. They didn’t ask how to add new AI features, how to build a better stack than competitors, or what tasks to automate. Instead, they asked different questions entirely: Where can we serve customers that currently go underserved? What structural advantages do we have in customer access, data, or behavior that competitors cannot replicate? How does AI enable us to reach those customers with capabilities that wouldn’t be possible otherwise? AI enabled their strategic choices rather than replacing them. The strategy remained about serving specific customers in specific ways. AI simply made those choices executable at a scale or speed or quality that wasn’t previously possible. Planning for Disruption in 2026 As organizations enter planning cycles for next year, the velocity of change raises an obvious question: How do you plan when the landscape shifts constantly? The answer may resemble Gamma’s orchestration layer. Rather than trying to predict which specific innovations matter, build a layer above the disruption. Start with your customers and what they actually value. Then assess which new tools, capabilities, and models serve those customers. Not every advancement in 2026 will apply to your business. Some developments will reach your market; others won’t. But if you’ve made clear decisions about where you play and what your customers need, you can adopt different components as they become relevant. This requires stepping back from the firehose of AI announcements and information. The companies that win won’t be those keeping pace with every release. They’ll be those clear about their strategic choices, able to integrate useful tools when they emerge, and ruthless about ignoring everything else. Upcoming Analysis Our next breakdowns focus on companies like Vanta, which demonstrate AI’s value in handling complexity that humans can’t maintain independently. Vanta enabled Granola to achieve SOC 2 Type II compliance in under 12 weeks, a process that traditionally requires months. Beyond point-in-time compliance, they’ve built continuous compliance monitoring, a paradigm shift that fundamentally changes how organizations approach vendor risk and regulatory requirements. Be on the lookout for upcoming breakdowns of leading SaaS AI startups like Gong, Rippling, Ramp, Vercel, Monzo, and Brex. These detailed analyses will shift to a shorter format, moving from 4,000-plus-word breakdowns to approximately 800-1,200 words, making the insights more accessible for reading in roughly 5-7 minutes. Future editions will incorporate more visuals and infographics to make information denser and more digestible. Get full access to Product Strategy Decoded at michaelgoitein.substack.com/subscribe

    29 min

About

Every week, I pull back the curtain and make strategy easier using a proven framework. You'll get insights of what works today and why, with powerful takeaways you can apply to your own business. michaelgoitein.substack.com