Shu-Yi Zhou

"I think I have ADD." The moment a founder says this to me, I know exactly what their team is dealing with. Roadmaps that change weekly. Ten contradicting Slack threads. Features getting prioritized, then deprioritized after one advisor call. The team stopped knowing which direction is real. Here's what I see constantly: A founder spends weeks prioritizing features with the team. Then someone mentions a new AI tool at a conference or a feature. Suddenly everything pivots to AI - not because customers need it, but because it's a buzzword. Meanwhile, paying customers are still waiting for the features they actually requested. The wild part? This is also what makes founders successful. Research shows entrepreneurship and ADHD share common traits: impulsivity, novelty-seeking, pattern recognition, comfort with risk. These traits are exactly why startups exist. The willingness to pivot and chase ideas drives innovation. As one founder put it: "This brain is scrappy, solution-oriented, and wildly resourceful." That's founder energy. It's a superpower. Until it's not. Without discipline, that same energy derails teams. Imagine making real progress, then walking into a meeting where the goalposts moved again. You can't feel successful. You can't build momentum. Seed-to-Series B companies don't have capacity to chase every idea. When focus disappears, progress disappears. The team feels it—hard. The solution isn't changing who you are. It's adding backbone. Someone who creates structure. Who keeps new ideas contained until they're fully thought through instead of letting them leak into the team and create chaos. Who asks the critical question: Does this new idea move us closer to our 90-day goal, or does it distract us? Most founders with "ADD energy" aren't broken. They're visionary! But when every idea gets immediate airtime, the company spins its wheels. You don't need to change. You need someone who channels your vision into execution without killing the innovation. That's the difference between founder brain and founder chaos. Do you have the operational backbone to match your vision?

10

If you've ever worked at a startup in SF, you've worked 996 (or something close). What's different now is that founders have to explicitly state their expectations instead of leaving them implicit. This shift suggests two possibilities: 1. Founders aren't as effective at motivating teams to work hard. This is probably true for some founders, but there are enough seemingly strong founders publicly touting 996 that it likely doesn't explain the whole trend. 2. Tech workers have become more transactional. In the 2000s, you could pay people pennies and still get quality work because they genuinely loved building software. Today, the talent pool is larger, but people aren't here for the love of the game. Many are here for number go up. These smart, mercenary workers know how to game the system. They optimize for money and status while minimizing effort. Which is exactly why founders now have to be explicit about expectations from day one.

59

15 Comments

Something's shifting in software pricing, and it's good news. Stripe's AI Scaling Playbook just dropped data on companies hitting $1M ARR in under a year. They're all doing the same thing: Hybrid pricing models: subscriptions & usage-based billing. Not instead of SaaS pricing. On top of it. For 20 years, SaaS was simple: $99/user/month, whether they logged in once or 1,000 times. Predictable? Yes. Fair? Not really. Light users subsidized power users. Everyone overpaid for simplicity. AI products are changing that, because the economics force it. Every generation costs compute. Every query burns tokens. Real marginal costs that SaaS never had. Companies are adding usage-based components. But smart ones aren't abandoning subscriptions. They're offering both: → Base subscription (predictability you're used to) → Usage billing (fairness tied to consumption) → You choose what fits Real examples: Browserbase: bundled browser hours & pay-as-you-go overage. Free tier lets you learn usage first. Result? 17% more signups. Hex: compute-based billing, 500M events processed. Customers set thresholds while they learn. ElevenLabs: hundreds of thousands of subscribers, one engineer manages all billing. Entire marketplace runs on usage-based pricing. What they're doing right: ✓ Clear metrics (hours, minutes, compute, not abstract "credits") ✓ Real-time usage visibility (dashboards, not surprise bills) ✓ Free tiers for risk-free learning ✓ Multiple options (subscription, hybrid, pure usage) AWS proved this works 15 years ago. Same questions then: "How do I budget? What if it spikes?" Now? Every company forecasts cloud costs, sets alerts, optimizes. Learned behavior. AI pricing is the same curve, just reaching end-users instead of IT. Higher bar for clarity. Better education needed. But same destination: You pay for what you use, not what the vendor hopes you'll use. Why this matters: → Start small, scale as you grow → Light usage = light bills (stop subsidizing power users) → Free tiers let you test first → Bills reflect value you're getting In 5 years, this WILL BE standard. Usage-based pricing becomes default for AI tools. Flat subscriptions become premium for people who prefer predictability. You already pay this way for electricity, Uber, cloud storage, mobile data. AI is extending that model to creative and productivity tools. If you're evaluating AI tools: prioritize real-time usage visibility, free tiers, and pricing flexibility. If you're building pricing: meet customers where they are (subscriptions), show the benefits of usage-based, make the transition effortless. SaaS pricing isn't dying. It's evolving into something that works better for everyone. Grab the playbook: https://lnkd.in/e-kUfEwm Where are you on this? → Already optimizing AI spend like cloud costs → Hybrid models (predictable base & usage) → Flat subscriptions, watching the shift Comment below, I'm tracking how this plays out.

4

1 Comment

PMs: Stop treating your demo like an afterthought. It's costing you 65% of potential conversions.   I've seen countless product teams build, ship, announce - then act surprised when adoption is slow.   Wrike did the opposite. They embedded Arcade into their onboarding flow and saw a 65% increase in paid conversions for new users who engaged with them. They understood that demos aren't just for launch. They're part of the entire product lifecycle.   If you're a PM or product leader, here's what I recommend:   1. Before you build: Share Figma prototypes in Arcade to test concepts with customers. Get feedback on the experience before writing a single line of code.   2. During development: Create interactive demos that let beta users explore and provide input. You're not explaining the feature - you're letting them experience it.   3. After launch: Embed demos in changelogs and release emails. Track engagement to understand what's resonating and what's confusing.   That's how you turn your demo into a key part of the product development process.

13

3 Comments

After years leading product teams at Headspace and building growth engines at Hims & Hers (and a few startups of my own), I’m working with Intro to support founders and operators through 1:1 advisory sessions. I’ve spent most of my career in that messy middle between “we have an idea” and “this is a real business.” What I’ve learned the hard way — across venture-backed startups, Techstars, and inside growth-stage companies — is that the biggest unlocks usually come from honest, practical conversations with someone who’s been in the arena. Here’s how I can help: ✔️ Tips on 0 to 1 products  ✔️ SEO, organic, & product-led growth  ✔️ How to validate ideas  ✔️ Guidance on MVP launches  ✔️ Best practices for growth  ✔️ Advice on retention strategy  ✔️ How to scale teams  ✔️ Guidance on product leadership  I care deeply about helping founders and operators build products people actually love, and avoid some of the expensive mistakes I’ve made along the way. If you’re navigating growth, ambiguity, or your next big product bet, I’d love to be useful. Honored to be on Intro alongside experts like Alexis Ohanian (Founder of Reddit), Alli Webb (Founder of DryBar), & Sarah Leary (Co-founder of Nextdoor). The link to book is in the first comment.

256

37 Comments

The goal isn’t to raise the most money… it’s to keep the most when it matters. Most founders optimize for headlines, not outcomes. Capital is fuel, not validation. Use it like oxygen, not gasoline. Just a quick thought for you fellow founders.

18

The most important startup metric isn’t revenue. It's whether people use your product. We used to think it was revenue. We chased revenue hard. We scaled fast. We booked millions in just a few months. But then we started to look deeper. A pattern emerged. The customers who signed up without a demo, who clicked an ad and converted instantly, rarely signed in again. It was tempting to dismiss it. Maybe they trusted the "AI agent" to just run. Maybe they didn’t need to log in often. But when we looked at the data in retrospect, the story was clear. Logins were a leading indicator. The users who came back daily, who explored new features, who gave feedback, and who engaged, were the ones who renewed. They were the ones who referred others. They were the ones who stuck. Since then, we’ve made one shift: Make customers love the product. At all costs. It's our primary KPI. When they do, everything compounds: 1/ NDR climbs 2/ True CAC drops as they refer other customers 3/ Upsell and cross-sell become far easier 4/ Reviews get posted and case studies become easy Bad revenue slows you down. Good usage builds real businesses. If you measure one thing this quarter, measure this: How many of your customers actually love using your product? Then do whatever it takes to make that number go up.

109

39 Comments

One type of search I’ve become comfortable walking away from is the ‘silver-bullet hire’. These are the red flags that almost always come up: 1. Vague messaging about what problem your product actually solves 2. An unclear ICP, often too broad with no defined personas 3. Lack of data on sales metrics, from top to bottom of funnel 4. No clear sales process (or citing sporadic use of a methodology as a process) 5. External locus of control e.g. “that hire didn’t work out because they didn’t work hard enough” or “they couldn't do outbound” 6. Unstructured onboarding, or the same generic onboarding used for every role 7. Ambiguity around how success is measured 8. Low literacy of the job function, hiring someone to “figure it out” 9. Prickliness when challenged on any of the above If these sound familiar, you’re probably not ready to hire yet.

26

1 Comment

"Yes, SHE can." First time I asked my product manager about a feature detail and got a "she" back. Pretty wild moment. It's timely. Teams that make ads are being asked to 3x their volume every year (sometimes more) without hiring (not possible without some AI magic). So while AI means the content treadmill keeps going faster, the good news is that I don't think just grinding harder has to be the answer to keeping up. Agents can be that extra pair of "hands" to do the competitor research, customer research, iteration analysis and asset generation so that we can all run faster without running out of breath. Excited and terrified to see it all play out.

7

1 Comment

YC's advice of "Focus on one thing" has killed more startups than bad products. I know because it almost killed ours. 3 months in, a close mentor sat us down and said: "You're doing too many things. Pick one and go all in." So we listened. We picked one approach and spent three months perfecting it. By the time we shipped, nobody cared. The market had moved. The irony? In trying to follow YC's advice, we forgot to follow their most famous one: talk to your customers. We were so heads down building the "one thing" that we stopped listening to the people we were building it for. Meanwhile, one of our engineers had been tinkering with something on the side. Turns out he'd been talking to users on his own and building out deep IDE integrations. His side project is now the fastest growing feature at Willow Voice. I realized that I had interpreted the advice as "focus on one solution" when the advice actually meant "focus on one problem." Dropbox focused on one problem but tested dozens of approaches. Airbnb focused on one problem but pivoted three times. They didn't focus on one solution. They focused on one problem. And they never stopped talking to customers while doing it. That's what we do now. Every engineer owns a bet end-to-end. Then every Sunday, we dogfood with Willow Voice and build out the feature in a few hours for fun. This way we can run a ton of experiments. Last month two survived! One of them doubled our activation rate. One metric. Multiple bets. Kill ideas fast and never stop talking to users 👉 — Written with Willow

54

4 Comments

If you solve a problem at the incentive layer, everything flows downstream. Before Light Labs, I worked in growth & GTM, helping Samsara scale revenue 100x. Today, Samsara is a $23B public company. To support this growth, we built an incredibly mature attribution model to clearly understand where to double down on marketing investments and where to cut spending. When I first entered the CPG space, I was genuinely surprised to find no attribution model for certifications. Brands spend tens or hundreds of thousands of dollars each year on certifications without clarity on the return on investment. Certifications are critical—consumers recognize and trust them—but we saw an opportunity to make testing measurable. By tying clear ROI to testing, brands can reframe a quality expense as a marketing investment. We created the Product Insight Panel to help brands transparently share real lab results with their customers. The impact has been incredible: brands see an average revenue lift of 12%, ranging between 1% and 22%. This ultimately creates a virtuous cycle that benefits the entire food system: brands test → transparently share results → grow revenue → become category leaders → invest more confidently in testing. If you're curious about the lift we've seen, the tests we've run, and how transparency can meaningfully impact your business, feel free to DM me.

46

Stripe is quietly costing you 1 point of net margin and sitting on your cash longer than it should. I've integrated payment gateways since 2003. In 23 years across Stripe, PayPal, and a stack of traditional processors, here's the math founders never run. Tax #1: ~1 point of net margin. Stripe Standard - what every founder signs up for on day one: ↳ 2.9% + 30¢ for domestic cards ↳ +1.5% for international cards ↳ +1% if currency conversion is required ↳ +0.5% for manually entered cards ↳ 0.8% for ACH (capped at $5) On $1M annual card volume, that's roughly $32K to Stripe. On $10M, ~$320K. On $50M, ~$1.6M. What a true Interchange-Plus processor charges instead: ↳ 0% markup on interchange and assessments ↳ Flat $0.10 per transaction (all major brands) ↳ +0.1% on Amex ↳ $0.35 flat per ACH/eCheck The blended rate lands closer to 2%. Lower if your mix is debit-heavy or you take ACH. Same checkout flow. Same simple PCI compliance process. The integration is almost identical to Stripe. Tax #2: You don't actually have a merchant account. When you use Stripe, you're a sub-merchant under Stripe's master merchant account. A real merchant account is yours - underwritten in your business name, with a direct relationship to the acquiring bank. What that buys you: ↳ T+1 next-day funding (vs Stripe's T+2 default in the US) ↳ No algorithmic fund holds - Stripe can freeze your money without warning if their risk model flags you ↳ A real support rep who knows your business ↳ Negotiated terms as you scale For a side project doing $20K/month, the Stripe convenience is worth the trade. For a company processing $1M+/yr, the math flips hard - on margin, on cash flow, and on operational risk. This is one of the cheapest P&L wins I've ever found in 23 years of processing cards. It shortens your cash conversion cycle on top of the margin recovered. Most founders never benchmark their processor after the first month. The few who do quietly add a point of margin, nobody on the cap table thought to look for. If you're processing more than $1M/yr on Stripe, run the blended-rate math on your last 90 days of transactions. Compare against an IC+ quote. Get a real merchant account underwritten while you're at it. Want the processor I use, plus the exact specs to ask for to lock in these rates? Drop "IC+" in the comments and I'll send the details over DM. Or DM me directly if you prefer - my profile is open. What's the most expensive default sitting on your P&L right now because nobody's questioned it since you set it up?

8

8 Comments

The single most valuable thing I learned at YC has nothing to do with fundraising, product-market fit, or growth hacking. It's a skill that 90% of founders are terrible at. When you walk into a YC partner meeting, they ask one thing: what does your company do? And then they rip your answer apart. "Simpler." "I still don't get it." "Pretend I'm your mom. Try again." I remember pitching my YC startup with some convoluted explanation about workflow automation and AI-powered insights. I thought it sounded smart. A YC partner looked at me and said, "I have no idea what you just said." Their entire program is obsessed with one principle: if you can't explain what you do in one sentence, you don't understand it well enough yourself. That lesson stuck with me. And when I started Mesa with my co-founder Oliver, I applied it from day one. Every product description, every sales call, every investor update — I ask myself: can I make this simpler? The answer is almost always yes. It took us months to compress our value prop into one line any engineer gets in five seconds: GitHub was built for humans writing code. AI agents write most of the code now. Mesa is the platform built for that world. That sentence is worth more than any pitch deck we've ever built. If you can't explain your company to a stranger in under 10 seconds, that's not a marketing problem. That's a clarity-of-thinking problem. Fix the thinking. The words follow.

22

2 Comments

Roadmaps are a commitment to deliver impact, not features. They’re a contract with the business: product pods will deliver measurable impact today that drives business outcomes tomorrow. At Curinos, my org's quarterly roadmaps have three sections: 1️⃣ Impact commitments → what we promise every quarter. Leading metrics like engagement, adoption, and breadth of usage that ladder into revenue, EBITDA, and strategic focus. 2️⃣ Considered initiatives → the features most likely to move those metrics this quarter. Smaller bets we can build, launch, measure, and iterate quickly. 3️⃣ Research & tech initiatives → work that won’t show up in this quarter’s metrics but sets the foundation: tech debt, security fixes, UX research, proofs-of-concept. Real work that has to be accounted for, or teams get asked, “What are you spending your time on?” The discipline is simple: - Deliver measurable impact every quarter. - Show how today’s work builds tomorrow’s success. - Make the invisible work visible. 👉 How does your org ensure roadmaps drive business impact, not just feature delivery?

4

2 Comments

Engineering as a career is dying. I'm only partially joking At Rounds, we've been getting way more demand for GTM and sales talent recently. AI made building so cheap that the ones who can sell have leverage SWE postings are down 35%. GTM engineering postings grew 205% in a year. If you're not technical or don't code, start learning GTM engineering. Clay, n8n, Smartlead, basic Python. It's the highest leverage skills you can pick up right now. (s/o Clodo if you want it done for you) Being ridiculously good looking helps in this field too, evidence below.

60

17 Comments

Hot take I can’t shake: software is about to explode, not get replaced. As volume explodes, the next failure point bottlenecks us: GTM. teams can ship like hell now, but we still fumble around with "what to build". GTM gets harder, not easier, as output increases. The answer right now seems to be "optionality", but that literally just means "throw shit at the wall and see what sticks" 😆 I’m asking this to gather collective input from other B2B SaaS leaders. Which vendors are actually useful for upstream GTM signal that shapes product decisions? (Problem selection, PMF, win-loss, competitive intel, narrative integrity) Drop names. I want to see what’s top of mind across the community, because it seems like a wave is coming.

8

4 Comments

This one's for my PMs out there: sharing Steve Huffman's (Reddit CEO) take on product market fit from a recent podcast. "What is Product Market Fit?" "If you have to ask..." Steve then expands with his view: "Did you do everything wrong and it's still growing? That's product market fit." It's about: 1. The product. "The product has to be good." 2. Organic growth. "You can't growth-hack your way into product market fit. [...] If you take your eyes off, it [should not] stop growing." He adds that you can't just buy users, you can't just do little stunts to get users. 3. Actions over customer love. "People loving the product is not product market fit. [...] It can feel good, it can look good, your customers can love it -- you may still not have product market fit." Words of wisdom as we head into the week.

6

If you're in product ops you'll immediately recognize this dilemma. A massive copy-paste challenge to maintain a commercialization/marketing view of the world...and a product view of the world. Made 300% harder because someone wants a "simple calendar" of launches. The reality is: 1. Initiatives have a M:1 relationship with Launches 2. Different launch tiers have different launch milestones 3. Calendars must include milestones, launch, and initiatives 4. Initiatives can have multiple releases 5. ...but not all releases are launch related 6. Oh, there are massive internal "launches" that marketing doesn't care about And this is why spreadsheets prevail. God-awful spreadsheets with: 1. Dozens of fields (like 6 dates for milestones) 2. Random manual fields to list multiple initiatives 3. Terrible VLOOKUP date madness 4. Product themes that are distinct from anything product uses A classic example of how there is a "right" model, but the UX complexity kills everyone, so they opt for something crappy. (PS: I love these problems. Hit me up if this resonates)

43

8 Comments

Hot take: The most successful SaaS companies of 2026 won't be the ones with the best UI or capabilities. They'll be the ones with the best "Agentic Commerce." Like a lot of people, I’ve been living in Claude Code lately. The velocity is just insane. Between Claude and Vercel, I can go from an idea to a deployed app in minutes. But the wall I keep hitting? Integrations. An LLM is a genius in a vacuum, but it only becomes actually useful the second it can touch the real world. Want a voice agent? You need a Deepgram/Daily API key. Building a compliance app? You need Go1 or EasyLlama. The "AHA!" moment for almost everything I want to build right now is 100% dependent on how fast I can plug in these external components. And that's where the friction is. It's a procurement barrier. Call me greedy, but I want the ability to purchase and provision APIs within the CLI. If I’m in my terminal and the agent realizes it needs a specific capability, I should be able to buy and access that key right there within the tool. I shouldn’t have to break flow, go to the browser, "Talk to Sales," or mess with a billing UI just to get a string of text. The easier a company makes it to access their APIs, and the simpler they make their commercial models, the more adoption they’re going to see. Ease of Integration is the new UX. If an LLM can’t parse your docs, authenticate, and "buy" your service in seconds, your product effectively won't exist to the next generation of builders.

16

1 Comment