Ben Horowitz On What Makes a Great Founder
In a Nutshell
Ben Horowitz identifies great founder CEOs as blunt truth-tellers who ask aggressive questions, think independently, inspire curiosity, and decisively tackle problems without hesitation or over-deference, exemplified by Zuckerberg, Elon Musk, and Ali Ghodsi. Common scaling pitfalls include lacking confidence (leading to decision debt and paralysis), botching executive hires like sales VPs (prioritize blind references, discipline over enthusiasm, playbook writers for complex products), and cultural mismatches between engineers and salespeople. Success demands constructive confrontation, clear behavioral rules for culture, and building confidence through reps to manage seniors effectively without micromanaging or pretending omniscience.
These notes were generated by AI and may contain inaccuracies.
Really good companies, the very best companies, tend to have founders and CEOs who ask pretty aggressive questions. Zuckerberg, Larry Page—those guys who have gotten all the way to the mountaintop—are pretty blunt. If you're running away from the truth to preserve feelings, that's a very dangerous thing in a tech company. The corollary is that bad news travels fast. As CEO, you need to find out if something's wrong.
Today's guest is Ben Horowitz of A16Z. Reasons he's fun: behind-the-scenes look on why Andre passed on HubSpot back in the day (funny story). He's seen so much, backed amazing CEOs and some that went down in dust. What do great ones have in common? Patterns? Same with failures. Read his book The Hard Thing About Hard Things (published 2014) when running HubSpot—really good. Wanted updated view on what's changed.
Ben is completely unfiltered and gets after it. Lots of nuggets.
You probably don't remember when we first met. HubSpot pitched you on the Series D. Three came in: newly hired COO JD Sherman, myself, co-founder Dharmesh (Dash). Walked in, nice welcome—your sales guy Mark Cranny was former colleague. Just written book Inbound Marketing, two of your marketers had it. Felt like hometown advantage.
Sat down: JD barely got a sentence out—“wait, you're the COO, tell me about your background.” Spent 15 minutes on JD, then 15 on me—“you're such a knucklehead, why did you hire a COO?” then Dharmesh. You guys passed, as did everyone. Two weeks later read your book, saw you're not a fan of hiring a COO. Put a ginormous chip on JD Sherman's shoulder—incredibly beneficial. Chips on the shoulder are really valuable.
How do you feel about CEOs, COOs these days? Same thing or changed your mind?
Generally when a company is small, flatter is better. In early 2000s, popular construct: Mr. Outside, Mr. Inside. If really scaling and not wrestling with product-market fit, it can work for a time on a product cycle. But when you hit the end of the product cycle, so much about a tech company is the communication architecture—it makes that worse generally. Two people in charge. Not to say it can't work—depends on definition of COO. If COO is big title for sales guy, fine. COO as “I'm running the company and you're Mr. Thought Leader outside”—not great for a startup.
Invested in tons of CEOs, coached tons. What are you looking for? Major green flag, red flag? How's it changed? Your filter on a CEO—Founder CEO. Looking at company, market, but the person, not co-founder, the CEO.
No canonical founder CEO. Mark Zuckerberg, Ali Ghodsi, Elon Musk—extremely different types. Don't get into look, feel, sound, extroverted/introverted—none of that matters.
Few things in common:
- Anybody great really thinks for themselves. Don't feel like reading the room or influenced by what I'm going to say.
- Can figure out if they're not that during the pitch.
- Truly original thinking on things.
- Leadership: Coen Pal definition—“leadership is ability to get people to follow you, if only out of curiosity.” Would I want to work for this person? How interesting are they? If not, won't bring in super high-end talent. Momentum thing: can't hire top talent, unlikely to be great company. Talent density matters.
At Sequoia offsite, discussion with Shawn Maguire (outspoken partner). He said: look for founders that won chess olympiad or math olympiad in high school, surrounded by those folks. Articulate, good point, many nodded. But HubSpot did pretty well—not Tesla, but co-founder and I Sloan people, hired tons from Northeastern. Worked out. Early cycle maybe need PhDs, chess/math champions, but as it moves up the stack, Salesforce/ServiceNow look like us.
Take: Depends on era and market. HubSpot was marketing/sales idea as much as technological—so great people on that side, not math olympiads. Math Olympiad very specific. Very best companies founded by exceptionally smart people—no question. Agree on that. How smart is Elon? Extremely smart. Math Olympiad? Probably not. Raw horsepower important. Larry Page one of smartest people in world. Matters if building Google-sized. Hard to imagine without somebody like Larry Page who could think that big.
Who's the best? Probably best personally working with now: Ali Ghodsi at Databricks. PhD in computer science, extremely smart—meets Shawn Maguire test. Exceptionally good at go-to-market for enterprise software CEO, competing with Snowflake (very good GTM org). Absolutely paranoid—only person as paranoid as Ali: Elon, Andy Grove (wrote book about it). Comes from being refugee: grew up in Iran, fled to Sweden when Ayatollah came in. Having everything taken away informed psychology—beneficial for founder.
Work with founders whose companies growing fast—new to this. Pattern of mistakes? Common traps?
Universal one: confidence (in a word). Nobody knows what they're doing—you invent something, no idea how to run company. Make damaging mistakes, feel terrible—hired people, sold great idea, then hurt everybody. Psychologically challenging (almost no founder been in that position).
Two dangerous reactions:
- Overly defer: Hired smart people, open to input to point of not deciding myself—just polling. But only CEO has full context. Leads to bad decisions + dangerous politics (“vacuum, I can step in”).
- Hesitation: Don't know right answer, 52/48 suspicion but afraid of mistakes, so don't decide. Or avoid: see need to fire head of sales, think about press/board instead of “can this person do the job?”
Those cause founders to fail at CEO job: lack of confidence, hesitation. Analogy: NFL linebacker fast but doesn't trust eyes—gets cut. Won't get to ball carrier. As CEO, trust your eyes, run at problem, make decision—or get replaced.
Resonates. At HubSpot: culture debt, product debt, decision debt. Decision debt worst—paralyzes company. Had it, fell in trap: company slowing, lots of debate/stuff on desk. Every 4-6 months: “need to make some f***ing decisions.” Then everything moved/flowed.
Other mistakes: team building. Hire VP of sales—people f*** up VP sales hires all the time, more than anything.
Hiring executives: don't know how. Often don't know what job is (e.g., CFO? control structure? systems?). Can't smell it out. Like hiring Japanese interpreter without knowing Japanese—all sound good, hard to distinguish.
Preparation important: talk to CEOs—what's difference good/great CFO? How hire? What ask? See? Understand job. Try to do job yourself for feel of challenges. Want CFO for your business, not generic.
Engineers running company hire head of sales—culturally different, down to how they talk.
Engineers: ask question, 100% try to think correct answer. Salesperson: first thought “why the f*** are you asking?”—clue.
Ask engineer “Does product have this feature?” Yes/no. Sales guy: “What competitor planted trap? Their weakness? How get to that?”
Engineer talking to good sales guy: upset, not answering question—figuring out why asked. Good sales guys get rejected because you don't like them. Terrible ones hired. Mark Cranny: CEOs want guy who failed engineering test, clean shirt, make head of sales.
Engineers upset paying sales commissions? Tell them: if want commission, grab bag, sell software—if don't hit number, fire them. Otherwise shut up. Engineers write software weekends for fun; no sales guy sells software as hobby. Prize fight—no prize, no fight.
Cultural difference vast—what are you really looking for?
I'm engineer CEO of your brand new company—likely screw up hire. Advice?
Example: Octa (Sequoia + a16z invested, did very well; Pat Grady on board). Todd almost made big mistake hiring head of sales. Two candidates: one Adam Erenss (PTC guy, worked for Cranny, McMahon—knew tree, called all: “this is the guy”). Other: front-door references okay, not great. Todd wanted other—more enthusiastic about Octa. Adam: “Don't know, let me talk to customers.”
Advice to Todd: Don't want sales guy enthusiastic—you want them qualifying you. Good sales guys qualify customer: money? Need? Shot? Or politically boxed?
Plus: guys who owe me favors (more than Adam) say he's great; other guy, guys who owe him favors say B. Do not f*** up. “If this hire doesn't work, end of company.” Put pressure—hired Adam, made the company. If hired other, Octa bankrupt.
Takeaways: blind references matter. For sales: who bringing with you? Call them, see if coming. Great sales leader has big followers; not great: nobody. Salespeople savvy about leaders—even mediocre knows good leader.
PTC equivalent now building sales leaders for Silicon Valley? What about PTC mafia worked? PTC great company, '90s fastest growing for a while—sold CAD/CAM, PLM to manufacturers, hard sell, very good. Diaspora compelling.
PTC: product wasn't that great (windshields/hoop didn't work)—anybody can sell easy product. Hire sales guy with hard sell.
E.g., Google Adwords: “How much do you want?” Not building culture.
Databricks VP: brand/mojo—not early Oracle salespeople (different from today). As established, harder to tell—survive on brand.
PTC amazing: culture/attitude (unique, probably illegal today). Replicable if difficult product: discipline. Into account with competitors: systematic traps for competitors, comprehensive technical/business case, chart everyone in decision process (complicated for PTC), assurances all lined up. Complicated, requires courage/effort/competitiveness—translates anywhere.
Easy sell: skip steps, get away with s***—predisposed (e.g., now OpenAI/Anthropic, everybody wants AI).
PTC: “Who the hell are you? PTC? Never heard. Why buy? Can't work? Let me walk through.” Invaluable.
Ryan Gabrisco, original head of sales at Databricks: from public company selling secure FTP—I'd never heard of it. How good to make number public company selling FTP? Discipline forced, extremely smart—take that all day.
Successful company fine, but playbook harder to write, product harder to sell.
Playbook runner vs. writer: Joined VMware on fire, ran playbook? Or wrote playbook, figured how sell complicated software, lock out competition, train/discipline salesforce? Much harder find. Depends on product complexity.
Being very disciplined on both sides and very process-oriented, managing the process tightly. Very tight on the profile of sales rep, and that works very well.
When interviewing sales guys, always ask what's the profile of your rep. Senior people in big jobs sometimes have loose profiles, whereas those who compete have extremely tight, specific, often surprising profiles.
At HubSpot, got the process to interview a rep right. Good learners did well. Liked people on their second sales job, not first, fifth. Gone to half-decent state school, B average, competitive athlete. If they could fog a mirror and met those, brought them in for half hour.
Gave scenario: "Ben, welcome." Acted as customer, said "sell me HubSpot," gave 12 minutes, literally watched 12. Then "that's very good, Ben. Here's my feedback," 12 minutes feedback. "Think about it, let's do it again." If they internalized feedback and sold better second time, hired 99% of the time. Scaled well.
Tests how good they are at listening, not talking. Subtle but important difference.
Favorite email in history: Had big product launch, press tour in couple weeks. Mark Andreessen front-ran it. Sent him email: "guess we're not going to wait until the 5th to launch a product."
Response from Mark: "Apparently, you don't understand how serious the situation is. We're getting killed, killed, killed out there. Our current product is radically worse than the competition. We had nothing to say for months. As a result, we've lost over three billion in market cap. We're now in danger of losing the whole company. And it's the server management's fault. Next time, do the [] interview yourself. [] off, Mark."
Are we out of the [__] off era? Emails like that at PTC. Elon’s not soft. In Cranny's interview with McMahon, McMahon: "What would you do if I hit you in the face right now?" Cranny: "Well, first of all, you better knock me out." McMahon: "But what would you do?" Cranny: "Is this a question? Are you testing my intelligence or my courage?"
First interview, walked in with suit, Harrison (original head of sales): "I wouldn't wear that suit to a [__] fight."
Very best companies have founders and CEOs who ask aggressive questions. Zuckerberg, Larry Page are blunt.
Critical culturally: give direct feedback, have it out. Mark was wrong on that feedback, but important he could say it, and I could stand up to it. Otherwise truth doesn't come out. Running away from truth to preserve feelings is dangerous in tech company.
Corollary: bad news travels fast. As CEO, find out if something’s wrong. Need bluntness.
Andy Grove called it constructive confrontation. Anecdote: somebody late to meeting at giant Intel, Andy: "All I have in this life is time and you're [] wasting it." Meanest thing, but resets culture as company grows. Everybody tells the story: we're not [] late to meetings. Culture setter.
Paul Graham's founder mode memo now standard. Part right: Brian Chesky (originally Brian Jessie?) hired extremely senior people, overtly deferred, created thefts, politics. Took company back by going dictatorial "founder mode."
Danger: people take it as "don't hire any senior people." To compete with worldwide sales organization, growing from scratch, senior person knows profile, territory splits, international expansion, years of experience and relationships. Can't learn all on your nickel.
Founder mode doesn't mean avoid hiring experienced; means manage them confidently. Learn enough about the job to manage, not have them tell you what to do. Overcorrection from before.
Jensen Huang: 60 direct reports, feedback in public. Most CEOs don't do that. Not like Elon's playbook or algorithm. Founders lack confidence to run like that: requires CEO-caliber direct reports, not developing them. CEOs can't develop executives; they either can or can't.
Need knowledge to evaluate company. Elon most competent/confident. Jensen too. Zuckerberg developed over time, deferred to Sheryl Sandberg until confident. Normal path. Hard from zero to Jensen. Jensen's 30+ years in.
Elon got fired early. Learning curve even for best.
Zuck excellent at first principles thinking, not overly influenced by others' ideas. Disciplined at data lens: what does data say? Grew fast, overcame competition. Stays extremely curious—all great CEOs interested in everything.
People underestimate Zuck, Elon, Jensen's people's skills—astoundingly good. Zuck introverted early, but always had it inside. Great psychologist in M&A. Mom in that profession. Needed confidence. Started company as little kid.
People don't know what culture is—wrote book for that. Cultural issues hit when big. Early startup: manage culture by hand unconsciously, sowing seeds.
Culture isn't values platitudes (integrity, have each other's backs). Actual: behaviors that put you where you want, give advantage.
Venture capital: everyone loves entrepreneurs, but entrepreneurs say VCs arrogant, talk down. Behaviors: on time for pitch? On phone during? Get back if not investing?
What you do is who you are. To be company treating money like own: systematic—hotels, travel.
Cultural gotcha: brilliant 100x engineer who's an []. More in VC than engineering. Define parameters: what [] means. Can't silently rewrite code at night. Expect super high standard?
Engineers great at following rules. Make clear parameters.
VC: spiky personalities. Steve Jobs mean because brilliant, or got away with it? Latter—he didn't have to, but could. Founders modeling Jobs/Elon: if not them, people walk. Hard to have highest talent without spiky behaviors.
Outlawed: can't make yourself look smart by making others look dumb. With entrepreneur on Twitter: "that business sells dollars for 85 cents. Haha"—fired. Love you're smart, don't do it that way. People fine once know.
Harder: "don't be an [__]." Define over the line. Spiky people can live in that context.
Founders pretend they know. Everybody pretends. Took while: went public 18 months old at LoudCloud—grew up fast. Felt knew what doing 3.5-4 years in, after rebuilding sales force, bringing in Cranny.
Never as confident then as now at Andreessen Horowitz. Need reps: more confident judgment = faster decisions, less care what people think or mistakes.
Jensen/Elon magical: sit in room, know exactly what want. Didn't start that way—build it.
Entrepreneurship harder than VC. VC more luck (join a16z/Sequoia, one of three deals trillion-dollar). Entrepreneurship: one shot, pay directly for mistakes, clock ticking, quarters. VC no quarters, many funders vs. one. Can lay off thousands—brutal.
Host bug as CEO: conflict averse—many first-time founder CEOs. a16z/Sequoia back many first-timers who haven't managed. Confrontation uncomfortable unless grew up with it.
Borrow constructive confrontation. Need with most in organization.
Decision debt: kick can—fire VP of sales hoping next quarter, fix pricing later.
Lack of confidence, want to be popular CEO.
Relates to hiring: first-time CEO hires more senior, nervous managing, defers—problems start.
Founder mode: from deferring to micromanaging. Constructive confrontation productive. If risk-averse, want popular, be more constructive in confrontation.
Being CEO uncomfortable, especially hypergrowth—everything breaking. Fine not knowing what doing: Ben didn't at LoudCloud, host didn't at HubSpot.
I didn't think I knew what I was doing when I was running HubSpot. So, if you're feeling that as a CEO, I'm here for you. Hope you liked it. Let's keep the convo going. I'm Behag on X. See you over there.
Keep Sequoia Capital in your library
Save the videos and channels worth coming back to, and find them again in one place.
