YC SUS: Eric Migicovsky & Dalton Caldwell discuss pivoting & pitching
These notes were generated by AI and may contain inaccuracies.
Eric Migicovsky, Startup School course facilitator, hosts live Q&A with Dalton Caldwell, YC partner in charge of admissions.
Question from Danielle at Squid Bio: At what point does pivoting turn into shutting down and starting a new company?
Founders are too willing to start new companies. Change idea with same Delaware C-corporation using Stripe Atlas—no need to redo. Main reason for new company: legal liability in old company or strong reason. Keep same cap table, investors—waste of time, unproductive work like dealing with entities and lawyers.
No need to change corporate name immediately—Jawbone called Aleph for 10 years before renaming. For Danielle with grant money and incorporation: stick with it, no harm. Focus on product-market fit, not corporate shell game.
Question from Peter at Whiplash Gaming (interactive watch party platform for eSports): How to judge between product not good enough (pivot) vs sales efforts not good enough (different strategy/channel)?
Good startup ideas may not fit founders (founder-market fit). Doesn't matter if founders bad at sales or bad idea—same outcome. Avoid hypotheticals, diagnosing too much time (conversations, advice). Early sales = talking to customers, responding to direction. Can't get past that: idea not picking up steam.
Cases of founder-market misalignment: coders talking to programmers but idea requires selling to insurance companies—struggle. Great startups solve problem founders themselves have (they're the customer, know how to sell).
Question from Matthias at Estill: Pivoting from wrong-fit returns solution (40% of clothes returns) to discovery tool showing clothes on body for new/used. Focus on new/used or both (50% shoppers do either)?
Answer: What are your numbers/sales? Avoid conceptualism without customers/revenue/traction. Get anyone to pay $1 for one purchase.
Metrics-based decisions over intellectual arguments. Works across domains: find few people to buy first version, incrementally add. Best pivot time: fail to get one live customer using product—that's the bar.
Don't worry if MVP lacks everything—worry if you can excite couple people to use it live (e.g., 100 returns). Then iterate (used or new).
Question from Julia at Marble (collaborative learning solution for teams, automates learning from links to courses/videos/articles into plan): How know if pivot design/functionality or continue user search?
First: Is anyone using it? Does product exist? Continuum: constant pivots without building vs building live product for year not working.
Actionable: Set up week-long experiments (landing page, product changes, or user segments). Control: don't change both product and users.
Example: Product has 5-10 users, test new segment (teachers vs parents)—aim for 20 customers in week. Hit: good sign. Zero/two: not right. Middle: tougher.
Subtlety: Pitch excites, but testing finds no use case—mirages. Example: "Startup makes delicious food" excites but not startup. "Free money", "less work"—leading questions.
Vague pitch ("learning solution for teams") excites everyone—not validation. Need concrete: "$20/user/month Slack plugin with daily reminders/lessons"—then ask to pay.
Experiment: Of 20 casually interested, how many pay for MVP?
Remote work example: Everyone excited, but need specifics (what it does, how works)—not validation.
Question from Joey at Company Nurture: Tips for hard tech (e.g., mobile apps easier than rocket ships) to avoid wasting time/money on late pivot.
YC funds many hard tech. Software risk: no one wants product (main cause of death). Hard tech adds financing risk: need capital to even show up.
Plan financing from start—even with demand/pricing/Kickstarter, tricky without.
Financing = business model where customers buy. Example: Robotics for labor-saving (hot, many YC investments).
Tomato-picking robot company: Built arm/vision/faster-than-human lean ($200K, not millions). Talked to greenhouses: buying timeframe, budgets, P&L, ROI. Built better financial model. Sized robot cost/charge to fit line-item.
Consider customer revenue/pay models, financing/loans if big upfront.
Tesla: Elon put entire net worth in during financing crisis—almost died half-dozen times. No bootstrap for some ideas.
Avoid late pivot: Validate customers want to buy specifically (Tesla Roadster: high-margin, sold to friends at high price, validated market/tech). Scale later. Large vision OK, but MVP deliverable.
Question from Elliott at Unify: Good way to pitch market opportunity? Standard TAM slide (all users who take pictures).
Early stage, TAM doesn't matter. Example: Early Uber TAM = black cars in SF (~$10M). Airbnb = sleeping on floors at conferences.
Investor TAM question = "not interesting". Combat with exciting company overall, not TAM debate.
Better: Specific market—investor visualizes, sees domination, generates ideas. Founder: "Obviously, I've thought this."
YouTube comment from Createtray: Largest sources for before/after pivot case studies? Metrics/rules to track for pivot?
Startup School best equivalent (detailed talks). Triggers:
- Unfalsifiable idea needing 2-3 years no feedback.
- Every turn: zero customers.
Product trigger: Users always need 2-3 more features before regular use (3 cycles).
Consumer null hypothesis: 80% signups poke around, never return.
Good sign: People use despite flaws/bugs—pursue, improve (don't pivot).
Biggest trigger: Complicated narrative why OK it's not working (mental prison, keep coding 9 months).
Talk to founders: How far to product-market fit?
Question from Matt: If pivoting/considering, mention in YC app?
Yes, use "other ideas" section—even not serious, interesting. Primary app on main idea.
Subtleties like go-to-market (B2B?, industry?) not pivot—OK to say open to options. Avoid disjoint app.
When applications mention multiple disjoint ideas like vegan dog food and satellites, pick one—the one you like most—as the primary focus. Put other ideas aside. YC looks for founders, not perfectly formed startups, but disjoint ideas make it hard to understand the deal.
Don't submit two applications; it signals hoping YC picks the better one. Put your best foot forward with the idea you're excited about—that's what YC wants to fund.
PostHog (now in YC) applied with one idea, got funded, then pivoted mid-batch after a blog post detailed their process (on Hacker News last week). They discussed Startup School, the interview, and office hours where they decided to pivot—best pivot this batch.
PostHog built an open-source, self-hosted analytics tool run on first-party infrastructure. Unlike third-party analytics (e.g., Google) that are sketchy, slow, blocked by ad blockers, PostHog provides analytics benefits without downsides.
In interview: "This is great, how fast can you build it?" They built fast, had day-one customers, built it at prior job. Market agreed.
During YC: Funded for first idea with $150K, big fans of idea/founders. After 3-4 weeks working on early customers, hit wall. Had pre-YC traction but never charged—excitement evaporated when asking for money. Happens to dozens of YC companies; normal, sometimes a relief. Pivoted, felt like wall of water rushing over.
Not harder for social apps inherently—evaluates based on something different, smart founder insights. Social crowded; differentiation often only in founder's mind, not proven by users/traction.
Examples: Dating apps with meaningless features like hats; apps copying Facebook/Instagram/privacy/TikTok/with friends; blockchain versions. Founders convinced, but no external evidence.
B2B easier for first customers (e.g., paying $100K great). YC has 10-15 partners/visiting partners reading apps; expert in your domain seeks them out. System has keyword search for cool ideas.
Guarantee: Every possible startup has someone at YC interested. Don't stress B2B vs. social.
Bar for social: Successful founders shipped toys/side projects with millions of users (e.g., Zuckerberg's high school Winamp plugin with millions, Microsoft wanted to buy). Sophisticated about growth, million downloads, even non-monetized.
Advice for consumer builders: Ship anything all the time, develop philosophy on what works. Don't create one perfect world-takeover thing—doesn't map to reality. Successful social founders tried a few times, have internal strategy.
Sophisticated investors: "Great, behind you 100%, how can we help?" Context-dependent; talk it through. Legal: Convertible note/SAFE ramifications; can't usually ask money back if spent.
If money unspent: Investor might not want in (e.g., biotech investor, pivoting to consumer social). Offer money back. Odeo story: Podcasting pre-smartphones; offered money back—half took it, half didn't. Those who stayed made lots; scared people from asking back.
40% of current batch targets markets outside US exclusively—bread and butter, fantastic success.
Invest in experts in space/tech/geography; YC experts in Silicon Valley connections, raising money, pitching, network.
Examples: India—Razorpay (2nd largest payment processor, pre-launch funded); Latin America—Rappi (largest online/offline, in every city); Southeast Asia; Nigeria.
Help YC: Provide context on geography-specific problems (e.g., India's 2FA for payments explained need for local startup).
For B2C: Weekly active users, retention. Spot expensive; growing while losing money per user = negative economics.
Example: YouTube startup—video hosting costs exorbitant, sold to Google.
Ideal: Make money sooner, avoid faster growth = more losses. Facebook early: Always profitable via ads (Apple ads 2003-2004), offset burn, bought time.
Top: Easy to understand/parse. Follow Startup School advice.
First read: One-liner (what does company do?). If parsable, better than surprisingly large % of apps. Practice one-liner with others.
Don't hack system (cold emails, fake recs)—misplaced effort, doesn't endear. Getting customer weighs more.
Example: Obsessed with app video view counts—YC hid views on purpose (zero shown). Bit.ly links: Won't click (thousands of apps).
YC reads every application: Fully formed idea = dozens of minutes partner time.
Lesson: Do actual thing well (per Paul Graham)—secret to raising is good startup. YC teaches making investable company, not fundraising hacks.
Idea not in dropdown (e.g., biotech/hard tech): Fine, YC handles weird stuff.
Nonprofits: Interested, but harder. Not for hacking easier entry. Need tech advantage, life's work. Trend: Nonprofits with business models (generate income for ops/growth, iterate on metrics/KPIs + donations). Great success; apply.
Top 3 (soft skills): Behave like normal human, treat YC as normal humans—not theater.
Don't memorize scripts/monologues. Falls into natural convo in 1-2 min.
Mental model: Fun/engaging conversation. Not a pitch—normal people doing jobs.
In-person events affected (like CA fires, 2008 crash). Affected companies: Little could do—contingency, wait, spend less.
React quickly/flexibly (not spend much). Companies die from running out of money/ability to work—spend less for more iteration time.
Hard. Examples:
- Dev shops (contract dev): Fund if pivot to specific good idea, use resources to build, ignore dev shop (clean up by YC). Great: Team knows each other/skills, move fast. Downsides: Too large team post-pivot, confuse consulting vs. startup revenue.
- Two founders: Find customers paying to manually do future product (e.g., DevOps infra). Feels like selling time, but funds product dev to sell to others. Good success—not pure service.
Average international app accepted at same rate as average US app.
International applications get accepted to YC at the same rate as U.S. applications, with percentages lining up and no wild difference between them. A large percentage, almost a majority, of companies in YC have business outside the U.S., which is totally normal.
YC encourages companies to come to the Bay Area for three months to network and build community, especially if from places like Indonesia or Japan. However, companies can return to their original HQ right after YC and raising money—that's totally fine. In the old days, there was a meme that YC required moving to the Bay Area, but that's not the case.
Having a provisional patent for a mobile app provides little value when applying to YC or pitching investors for initial funding rounds. Patents are more relevant in hard tech, business processes, or biotech. Software patents are generally problematic and not impressive to investors or YC. Investors often ignore them as a line item.
Sophisticated IP strategies can feel productive but are not valuable without customers. Focus on MVP and initial customers instead—99 times out of 100, that's way better use of time.
Dalton, thanks very much, really appreciate your time. That's it for today, thanks.
Keep Y Combinator in your library
Save the videos and channels worth coming back to, and find them again in one place.
