YC SUS: Kat Mañalac and Eric Migicovsky discuss Week 2 SUS Lectures
These notes were generated by AI and may contain inaccuracies.
Eric, facilitator at Startup School, joined by Kat Mañalac, YC partner, answering questions on Week 2 lectures: How to Talk to Users and How to Launch Again and Again and Again. Questions from Startup School forums, YouTube comments, Twitter, smoke signals.
Flexo enables building financial models for growing companies in minutes, not days. One-liner is good, but add who it's for: startup founders or small businesses?
Balancing Launches vs. Success
Question: Many channels to launch on—quick launch or spend days/weeks maximizing success per channel? Recommendation: Don't spend weeks perfecting channels. Spend few hours on messaging, couple hours to days on feedback. Pre-order/Kickstarter/press take longer. Stop emphasizing one perfect launch.
Everyone in the world doesn't know who you are right now. Every time you launch, 0.01% of the world finds out. Every single launch is brand new to so many people. Maximize surface area. Hard to get people to care early—want more chances, feedback. Like MVP: if perfect by launch, waited too long.
Katara helps find and eliminate bias in AI training datasets.
Landing Page CTAs and Waitlists
How long is too long for CTAs like "sign up to be notified of alpha launch" or "get on mailing list"? Longer waitlist = harder to reengage. Get MVP out ASAP. CTA as short as possible: users don't care about alpha/beta. Say "sign up for launch notification" or "newsletter." Get feedback on copy: post on Startup School forum for eyeballs. Ask what they think. Obsess less over perfect copy—users don't know what you're talking about yet. Jump in on forum posts you don't understand for honest feedback.
Storylow helps brands hire photographers for digital marketing needs. Have user interviews next week + paid acquisition via incentives.
Mixing User Interviews and Incentives
Pitching coupons/incentives during feedback interviews? Better not to mix. Separate interviews: (1) non-users to learn day-to-day problems, existing solutions; (2) users focused on improving product. Hard for users to context-switch back pre-product world. For non-users: wide-open questions—what products used, paid for similar, how they failed. Ask for feedback without incentives—if real problem, talking about it is incentive. Constantly paying for surveys? Maybe not burning problem, or product not 10x better than alternatives. Good data point for ideas.
Platform simplifying finding effective mental wellness practices/solutions. Issues: unclear if app/website/service; missing who for (consumers? connecting to therapists?) and exact problem.
Influencers/Bloggers Cheap/Free
Effective strategies? Tough. Research: solve problem they/their readers care about. Numbers game: cold outreach to 50 bloggers, maybe 1 responds—send hundreds. Aim lower: Facebook groups, subreddits, Twitter hashtags, Product Hunt, Happy News. Example: Company for cerebral palsy walking device found Facebook group of parents/families, got dozens/hundreds signups. Follow up like sales: don't expect first-email response. Joy (wedding site builder) drip-campaigned 50 wedding sites/blogs after 1/50 responded.
Affordable pickup for shopping returns directly from porch. Launching in Pittsburgh.
Geo-Targeted Local Launch
Lug (Uber for movers): first week, stood outside Ikea—one founder pitched people in line, other drove truck up. Manual ground work early. Leverage personal networks: past coworkers, friends, classmates—universal Amazon return problem. Door hangers on houses. Building elevator flyer worked for one startup. Avoid stickers on Amazon boxes (looks like theft).
Self-driving molecular design services as GPS for drug discovery.
Launching Biotech/Pharma Tech
Press/media: Fierce Biotech, Stat interested in early-stage. Find reporters covering similar companies. List/prioritize, offer exclusives (e.g., launch). If no, next. Launch with reason: conferences, papers, posters. Pitch timed around events. Universities may help publicize student/researcher work via specialized groups.
Online platform to trade purchasing data (e.g., Amazon history) directly with companies—monetizing, transparent, consumer control.
Mom Test: Testing Future Features
Talk specifics, not hypotheticals. Gathered user insights on desired functionalities for MVP evolution—test with others? Build smallest MVP solving their problem, put in hands for feedback. Easier than hypotheticals. Gather problem insights first, build/test solution. People may wish for problems they don't have.
Teaches kids AI using video games.
Saving Launch for Big Event
Does it make sense to save hard launch/full v1.0 for Demo Day? No—launch quickly for traction/feedback/iteration by Demo Day. Don't overcomplicate: push live, tell friends—that's public launch v1. Every launch reaches multitude who never heard of you.
VR training for first responders: users run calls, command virtual crews, in-game radio, avoid virtual deaths. By firefighter for firefighters. One-liner: VR training for first responders.
Solo Founders, Full-Time Jobs, Failed Launches
YC examples of solo founders building big projects full-time? Many built MVP full-time, then went full-time to scale. Working full-time tests if you care enough—convince yourself first. Talked to firefighters/captains? If regular-use MVP not used regularly, missing features/problem fit. Failed launches raise MVP bar.
White label payments SDK enabling enterprise partners to issue digital rewards for end-user value on platforms.
Pricing Experiments with Intermediaries
Paid per transaction; partners upsell to end-users—no A/B testing pricing. Enterprise pricing flexible—not public, different per customer. Most common: founders charge too little. Alternative is their engineers building/maintaining (hundreds/thousands dollars/engineer). Don't fear asking more—they'll push back if burning problem.
Helps users separate inorganic trash by scanning empty package barcodes. (Unclear: companies/businesses/homes?)
Users vs. Paying Customers
User ≠ payer (less consumer, more B2B: buyer manager, user other). Do variety of interviews. Start with same person as user/payer. Figure money flow: who decides budget, already paying for solutions?
Are they willing to pay or do they already have the budget to pay for a solution to this problem? If they don't pay or don't have a budget, they might not be the best first users for your product. We're selling our services to municipal governments. Is there any hack that accelerates their decision-making to pilot test your product? Governments are slow processes, extremely tough. Find out when the budgeting process starts, understand if they actually have budget line items to solve this. It's a non-starter if they don't—that would be six to twelve months of getting back on their radar.
Contact Pointers is a service for automatically updating other people's address books when you change your contact information and letting you control what info they get. We've heard this idea before, funded it before. It's a problem everyone has but not necessarily one people are willing to pay for. There are a lot of dead bodies in the past—recommend looking at them and doing a deep dive.
Hoping to do a friends/family launch soon. What feature targets for an online launch? Strong network effect, but features planned to make it useful sharing with people without the service. How vital is the quality of these features to an online launch? Worried early adopters find it not useful due to small network, harder to get them back later.
Legitimate worry, but not really because there are a lot of people in the world. First few users try it, don't like it, leave—not the end of the world. Product designed for a lot of people. Contrast with niche like sailing coaches (only ~1,000 people—losing 100 is bad). Friends and family will use it as many times as you tell them. Can launch again with improved product; they've launched the same people over and over, and it gets better each time. Problem: sounds like it needs a network effect. Try to build so it's useful even by yourself.
Sonny from Starett Stream Industries. Tell our bank we make silicon case antennas for responsive sensor frameworks, but we actually make AI sex toys that double the length of a woman's orgasm. Question for Kat about relaunch narratives.
Launched this sex toy five years ago as CMO, went viral, then dispute led company to pivot to medical devices, asked to return to refocus on original idea. If you were a potential investor looking at a rebirth company, what indicators prove the team can make new spotlight moments?
Understand what convinced you and the team to pivot back. What have you learned from five years in the industry? You should have major insights others don't. Have you raised a ton of money and burned through it, or is this first raise? Team has expertise in the space—that's exciting. Good luck, it'll be a fun ride.
Rafi, founder of Aventis Media, publisher of The Bridge magazine—digital magazine making news, finance, and career info relevant, engaging, useful to young adults 13-18. Subscription-based: users are kids, customers are parents who pay.
Customer discovery: kids not actively solving the problem, parents are. Asking questions like in the video, but missing an angle.
Users and customers are kind of both the parent. Topics sound long-term focused (post-high school). For 13-18 year olds, focus product down to specific niche like IB students or Model UN students. Then users and customers are the same person—easier. Price accordingly (allowance vs. Harvard admissions value).
Alexander at Insta Cents: one-click micro-donations to support digital content. Initial hunch: talk to content creators who would love new revenue. Most not engaging—prototype to show how it works?
Many solutions to this problem. Look at past attempts and their problems. Talk only to content creators who have experimented with other solutions (e.g., Patreon)—if not, they probably don't care. Focus on how yours is better. If technical and fast to build MVP, do that. Or hack/white-label existing solution to experiment quickly.
Bruno from Glee Board: matches ambitious CEOs with curated executives/founders into ad hoc advisory boards for better decisions.
When exploring user niches, prioritize? Talk to several at once then zoom in, or one group after another?
Prefer sequential: whiteboard all potential user groups, guess most valuable based on Venn diagram (acute problem, frequent problem, willing to pay). Start with top guess. If not good, move to next.
Yoda from SSS Spot: connects drivers in Latin America with public/private parking lot owners. Launched prototype 6 months ago, 15% month-over-month growth—impressive, indicates product-market fit (hopefully revenue growth).
Testing parking bookings for sports games/concerts. Interview new discoverers or active users?
Segment interviews: one group never used product, one using it—questions differ slightly. For new feature like this (new audience), talk to both active and new users. Park-for-work users differ from sports/concert goers.
John Dougherty: fastest way to identify beachhead with paying customers across industries/sectors?
If building to solve your own problem, users look like you. Otherwise, generate hypotheses list. Look at most active users or easiest sales—who you're solving big problem for, willing to pay. Within paying users, double-click on demographics (e.g., two similar users → find 100 more).
Assist Force: most reliable freelance platform for tech companies, replaces outsourcing agencies. No-name European company—how approach Silicon Valley tech companies?
Product Hunt, Hacker News first—hit people at target companies. Differentiate vs. Upwork, TopTal, etc.
Sandro Franco: downside of referencing existing product? E.g., interactive TikTok/Instagram for language learning.
Use "X for Y" if fastest to paint picture: X must be household name, clear why for Y. Best one-liners generate questions, spark interest.
Don't use if X is in crisis (e.g., WeWork for X—avoided in YC apps). Or if obscure (e.g., Buffer for Snapchat—not enough know Buffer).
Kelly Romina: is simple single-page landing page (like Magic) sufficient MVP to raise money, or need extensive features?
Magic raised on that. Depends: if solves problem, yes. Even "Hi, Eric here, call me to solve your problem" works. Some YC founders' whole product is their brain solving problems—tens of thousands on waitlist.
SmartWatch company, main feature real-time hydration monitoring. Sensors up, but other parts not. Organic growth by revenue: selling one piece at a time takes forever.
Common mistake: think organic revenue must sell 1,000 units. No—make one, sell to one, they love it, scale. Hardware founders resist, but that's the way. You've made sensor but not watch—attach to existing SmartWatch (Bluetooth strap hack).
Naga Sri from Tori: connects global shoppers to international travelers. Launch delayed 4 months, huge user drop—how get back?
Re-engage: call every one if few, emails. CEO personally reach out (Gmail, not drip)—no big company CEO does that. Often enough.
Use surveys for PMF instead of building then pivoting? Fine—not building MVP okay to check if problem exists. If yes, build MVP.
Metro Push: push notifications app for public service providers (electricity, water) who have bad rep for informing users. Avoid perceived failure before serving users?
Run yourself first: e.g., Twitter account on city utilities status. Build trust, then pitch providers: "We've built trusted service—give us data."
Jemelle from Nukes Vision: Google Analytics for retail stores. Hard to find first users targeting big like Walmart. Too small for them?
Tough—"whale hunting" (one contract = millions). Don't start there. Go after smaller customers with problem, fast buying decisions/demo. Get paying customer for feedback fast. Unlikely Walmart first without connection.
Michael from Biz (shows businesses around you, helps receive payments): okay to stop, rebuild, relaunch (not exactly pivoting)?
If discovered business won't work (e.g., after customer talks: no one has problem, not painful), consider options. Insanity: doing same over expecting different. Okay to pivot/change, or say not right thing. Maybe learned better problem from interviews.
Gifted: platform to manage wish lists from multiple sites, shop gifts for friends. Traction from high schoolers sharing things they love, but low spend/revenue.
Answered your question: if using but not spending, probably not right demographic. Change product for high schoolers + revenue, or switch demographic.
Cristina from ListAlya: in user interviews, figuring out features they want when scope is understanding behavior, not what we can build.
Incredibly important. Treat interviews as 15 minutes where user talks, you listen to their life: where they live, work, problem frequency. Get full context. If they pitch solutions, refocus.
Ask them why they're asking for that solution, try to understand what's driving the problem behind the ask. Guide the conversation towards problems they have and solutions they think would be good, then build it and give it back to them.
Elliot from Unify asks: How to successfully launch a big update for a mobile app to convert more free users to paying users?
Go through the entire launch process again: launch on Product Hunt, Unify 2.0, Hacker News. Always recommend companies do a blog post on their own site about the update, what it means for users, and why they built it this way. Treat it like your first launch and hit all the same channels.
Test the update with existing users first—treat early launches as a beta with 100 or 1,000 free users. If they're not upgrading to paid, figure out the words that cause upgrades and use that as a marketing experiment for the public launch. Treat marketing like engineering: an iterative improvement process with experiments and hypotheses to tune and tweak.
Best marketing messages for launches often come from users—write down actual sentences users say and repeat them to other users to see if it clicks. Users who have the problem and get excited may share words to excite others.
Wellin asks: We empower yoga teachers to do events at local homes. We've grown quickly, forcing a waitlist for new instructors. Should we hire more to expand in our current city, expand to a new city with a new MVP, or talk to investors to build an MVP into a platform?
Congratulations on the backlog of new instructors—this happens often with local startups.
Three options:
- Grow in existing city
- Expand to a new city
- Get investor dollars to improve the platform
Answer is grow in existing city—you already have a backlog there. Easier to build an MVP in one location than multiple. Perfect it at one location, then copy-paste to another. Hard to improve if two+ locations are live simultaneously.
Avoid investor money if possible—get money from actual customers to fund the business. Makes fundraising easier later. Trick to fundraising is to not need investor money; investors love that.
Serious platform for preoperative tele consultations to prepare patients for surgery from home. How to launch fast in healthcare vertical with legal/ethical discussions for telemedicine in your region?
Launching fast and breaking things is good—except not in healthcare/medicine with legal or ethical implications.
Think of your product as an evolution/extension of existing products. If telemedicine or tele-consultation apps exist in your market, approach them to add a module for pre-op tele consultations. In the US, use HIPAA-compliant platforms—copy their tech stack into your app if worried about HIPAA.
Focus on one region to avoid balancing ethical/legal ramifications across regions.
Audio-to-text platform transcribes speaker sessions at conferences in real-time for accessibility.
Pursuing large customers: lower probability of converting, longer sales cycles—opportunity cost to not convert smaller customers quicker.
Not always best for startups. Earlier mention of pursuing large customers may have been a mis-speak. Get smaller customers who decide quicker, especially while iterating product-market fit. Waiting for big customers might kill you.
Thank you for questions—keep them coming on Twitter and forums.
Next week: live design reviews with Aaron Abela. Post your website/landing page in the forum thread on Monday for review. Session after long session (2 hours), enough time to cover many sites.
Keep Y Combinator in your library
Save the videos and channels worth coming back to, and find them again in one place.
