Title: Rob Liu — How to Start a Startup

Session context

  • Host: Lisa (According to Cloud)
  • Speaker: Rob Liu (Founder, Athena VC; Co-founder, ContactOut; ex-Groupon)
  • https://www.linkedin.com/in/robliu/
  • Topic: Crash course on starting a startup by “stealing” ideas, learning fast, and validating through customers
  • Company metrics shared: ContactOut ~60 people, ~$14M revenue, ~$10M profit

Core thesis

  • Good artists copy, great artists steal: Study winning companies and industries, extract what works, recombine into a 10x better solution.
  • Learn fast, test fast, and iterate relentlessly; prioritize inputs you control (learning, interviews, outreach) over vanity outputs.

How to find ideas and pick markets

StepWhat to doWhy it mattersTools/resources
1Survey investor-backed spaces (YC batches, Crunchbase, Signal)See what top investors believe is the futureYC directory, Crunchbase, Signal
2Look beyond obvious SaaS (energy, ag, healthcare, finance, marketplaces)Biggest markets aren’t always softwareIBISWorld, Gates Foundation initiatives
3Benchmark winners (top YC cos: marketplaces/FinTech dominate)Counterintuitive: giant categories can be easier to winYC Top Companies
4Prefer enormous marketsEasier to hit meaningful outcomes with small share“Big idea > small idea” logic
5Learn at high velocityOutlearn competitors to spot unique insightsChatGPT, textbooks, papers, podcasts

Learning workflow Rob uses

ActivityTacticNotes
Map problemsAsk GPT: “biggest problems in X industry?”Use it to branch to sub-problems
Scan solutions“Which startups tackle these?”Build a landscape
Read research“Papers addressing Y”Hunt pre-commercializable ideas
Go deepPull MIT/Stanford textbook lists; audio-TTS on walks1 hour/day at 3x = ~2 books/week
Close gapsScreenshot confusing bits → ask GPT for explanationsTreat GPT as tutor

Competitive teardown process

AreaWhat to doExample outputs
Product UXClick through 10+ competitors; screen-by-screen strengths/weaknessesBuild your merged UX
SalesBook their demos; record calls; capture pitch decks and sequencesBuild a best-of sales flow
MarketingSpy on SEO keywords, ad libraries, referral mechanics, socialReplicate and adapt (e.g., “James Clear email” SEO pages)
Inside infoDM ex-founders/ex-employees for mentoring (offer $500)Learn data sourcing, sales playbooks (e.g., ZoomInfo address book strategy)
AdvisoryUse intro.co to book unicorn foundersHigh ROI insights quickly

Customer discovery and validation

GoalTacticsTargets/Signals
Understand workflows100+ interviews; map jobs, pains, current tools, gaps“What would you pay $50k for next month?”
Validate demandPresell, paid pilots, LOIsCash or clear willingness to pay
Build smallest thingStart with the highest-ROI feature onlyContactOut started with email lookup
Manual firstDo it by hand behind an API if neededLearn before automating
Iterate fast10–100 tests across product/sales/marketingLearn every week (LIFT: Learnings, Ideas, Fast Tests)

Growth channels and playbooks mentioned

ChannelHow to try fastNotes
SEOCopy proven intent pages from competitorsContactOut: ~2M visitors/mo via SEO
OutboundSDR motions, cold email, LinkedInTest targeting hypotheses quickly
Ads$5–$100 experiments to find ICPsLet CPC/conv data guide
Referrals/partnersCopy competitor referral designs; sign ecosystem partnersLeverage existing CRMs/platforms
Content/socialMirror competitors’ channels (TikTok/YouTube/etc.)Learn what resonates in niche

Sales philosophy

TopicGuidance
ToolsDon’t over-tool early; focus on “one most important thing” daily
ClosingNo magic lines; great product wins. Process helps, but value wins.
Learn by osmosisBook competitors’ demos, record and study
Word of mouthUltimate growth; driven by superior product

Pivoting and perseverance

TopicTakePractical approach
When to pivotOnly you know; iterate based on weekly learningsRun small tests; update hypotheses weekly
Time horizonThink 10 years; many flail for 3–4 before lift-offDon’t quit early
FundraisingTake as little as needed; optimize for founder liquidityRevenue > investment “problem”
CompetitionA YC competitor is not a deterrentBe 10x better for average user

Hiring, management, productivity

AreaPracticeDetail
Hiring questionsMost impressive accomplishments; books read last year; role-specific build questionAssess thinking via writing
MeetingsStack on Mondays; mostly group sessionsAsynchronous written updates
FocusOne most important task/day; track productive hoursUse a stopwatch method
LearningRead constantly; leverage audio + GPTMake learning part of lifestyle

Selected Q&A takeaways

QuestionAnswer summary
Tools for PM?Keep it simple (notes/stickies). Don’t overcook tooling early.
Closing fence-sitters?Product superiority + clear problem solving > tricks.
Sharing ideas publicly?Share freely; ideas cheap, execution hard; public builds audience.
IP concerns (toys)Revenue first; defend later via brand/acquisition. Patents secondary early.
AUS vs US (hardware/software)Sell to US early, leverage Asia supply chains; Aussies have US cultural advantage.
Vision vs MVPShow both: the big vision and today’s proof. Revenue traction attracts investors.
Freemium vs paidTry both; measure conversion/engagement; parallel small bets.

8-week execution tracker (Notion/Markdown-ready)

WeekOne Most Important ThingCustomer (interviews/demos)Product (milestone/tests)Growth (SEO/outbound/ads)Sales (demos/pilots)Learning (papers/books)Result/Learning
1
2
3
4
5
6
7
8

Milestone template

AreaBaselineTarget (8 weeks)Success criteria
Customer discoveryX interviews/weekY total; Z decision-makersClear top-3 pains; pricing signals
ProductMVP live to N users≥B% activation; time-to-value ≤ T
Revenue$0 or $MRR1–3 paid pilots or $A MRR10% WoW post-first dollar
Growth0SEO pages P; outbound Q/wk; ads test R≥ S signups/wk or cost/conv threshold

Action items for you

  • Pick 1–2 markets from YC/IBISWorld; list 10 competitors each.
  • Book 5 competitor demos; record and capture collateral.
  • Schedule 15 customer interviews this week; use 3 core questions about pains, current tools, and willingness to pay.
  • Build the smallest valuable slice; ship within 2–4 weeks.
  • Run 3 growth tests next week: 1 SEO page cluster, 1 outbound sequence, 1 micro-ad test.
  • Track in the 8-week table; update learnings weekly.