David Sacks - How to Operate a SaaS Startup - [Invest Like the Best, EP. 234]
What this covers
Patrick talks with David Sacks, a general partner at Craft Ventures and PayPal's founding COO, about how software companies are built and run. They begin with what lets a product explode. For consumer products, Sacks says, that means finding the single atomic transaction users will repeat, such as PayPal's email-plus-amount send or Uber's map pin. He argues that enterprise and consumer playbooks have converged now that all software ships through the cloud. A dashboard-style hook then kicks off a race to completeness.
The middle of the episode is diagnostics and operations. Sacks explains the burn multiple and treats gross margin well below 80 to 90 percent as a structural warning, especially when humans do work the software should automate. On sales, he says experienced managers enforce good selling behavior rather than just hitting the number, because reps paid on ARR may promise features that surface as churn later. His Cadence splits product and marketing from sales and finance, with the launch event placed mid-quarter so the pressure peaks are staggered by half a quarter.
The most contested material is regulation. Founders should obey clear law, but in gray areas they should explain their value openly and lobby, as Uber and Airbnb did. He also warns about platform dependence, citing PayPal's reliance on eBay. The final stretch profiles the PayPal mafia, including Levchin, Botha, Rabois, and Peter Thiel. It closes on founder psychology: "wild stallions," culture as a mirror of the founder, and leading from the front, with Michael Jordan's Bulls and Elon Musk as examples.
Sacks argues that SaaS companies win by turning consumer-style virality into multiplayer seat expansion, and that disciplined operating cadence, capital efficiency checks, and founder psychology determine whether that growth compounds or breaks.
- Team products with seat expansion beat single-player products on net revenue retention because expansion offsets logo churn
- A quarterly cadence staggered by half a quarter aligns product, marketing, sales, and finance without overload
- Company culture mirrors founder psychology, so unbalanced founders need boards and co-founders to check strategy
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
Building on a platform for distribution is dangerous because the platform can shut the app down or launch a competitor, and PayPal's dependence on eBay, which reached about 70 percent of payments, made it existentially vulnerable.
“The reason why it spun out is because eBay became a small fraction of PayPal's total payments. But at the time when we had first created the company, it was like 70%”
Pure software and SaaS businesses should have near-perfect gross margins of about 80 to 90 percent, since each incremental copy costs almost nothing and the main cost is hosting such as AWS, so much lower margins signal a structural problem.
“These should be just about perfect gross margin businesses. 80%, 90%. The main cost is usually just in hosting on AWS”
Founders must follow clear, black-and-white laws because non-compliance eventually gets caught at scale, but in gray areas where laws never contemplated the new technology, founders should build consumer value and evangelize publicly, since law tends to accommodate popular services.
“founders need to abide by those rules. It's very stupid not to.”
In gray regulatory areas, founders should not hide but openly explain their consumer value and social benefit and lobby for favorable rules, as Uber and Airbnb did, rather than trying to stay hidden as long as possible.
“founders make the mistake of just trying to hide as long as possible.”
Founders must comply with clear-cut legal rules, since violations such as Napster's may go unnoticed while the company is small but will inevitably be enforced once the company reaches scale.
“founders need to abide by those rules. It's very stupid not to.”
The Bulls won their first title only after adding Phil Jackson, Scottie Pippen, and role players like Dennis Rodman, and Jackson's triangle offense turned Jordan from a lone scorer into a team player, because a simple give-the-ball-to-Michael offense was too easy to double-team.
“what Phil Jackson did with the triangle offense was create a system in which Jordan could either take the shot, or if he gets double-team, dish it off to the open man”
Once a SaaS company has a singular product hook, usually a dashboard, it enters a race to complete the offering, because enterprise customers tolerate and demand complexity and the product must become comprehensive enough to justify payment against competitors.
“SaaS customers have a much greater willingness and need to kind of engage with a more comprehensive product.”
Startups that use humans behind the product to perform work the software should automate, as when no APIs exist for legacy insurers, lower gross margins and create an internal constituency that resists automation, so a company with many bodies is a red flag.
“once you develop a dependency on humans doing the work, it's actually very hard to change that.”
An inexperienced sales manager's job is to hit the number, whereas an experienced sales manager's job is to enforce good selling behavior while still hitting the number, including not burning prospects who say no, because incentives miss these subtleties.
“an experienced sales manager understands that their job is to enforce good selling behavior without missing the number.”
The Cadence has two main rhythms: product and engineering on a quarterly or seasonal release calendar paired with marketing around a big launch, and sales and finance on quarterly plans mapped to fiscal quarters with board meetings at quarter end.
“there's actually two main cadences or rhythms that's happening in the company.”
The PayPal mafia combined a playbook of pattern recognition developed at PayPal with favorable timing after the 2002 eBay sale and the dot-com crash, which led them to found companies or invest in ones using the same playbooks.
“there was a lot of pattern recognition that we learned at PayPal about what's successful”
Enterprise and consumer software playbooks have converged over the last dozen years because all software is now delivered through the cloud, whereas in the early-to-mid 2000s enterprise companies were top-down, CIO-gated, and culturally distinct from consumer companies.
“now that all software is being delivered through the cloud, you're seeing a convergence of playbooks from consumer and enterprise.”
Highly talented but wild founders, which Sacks calls wild stallions, can drive a company off the rails unless they learn to control their psychology, drawing on Jerry Seinfeld's comparison of talented comics who died young to unbroken horses.
“we have a name for those, we call them wild stallions.”
Sacks offloaded problems onto his board and exec team by putting the problem on the table for them to turn, then reabsorbed the partially solved problem, because billion-dollar companies are too hard for a founder to solve alone.
“I would take the Rubik's cube out of my head, put it on the table and I would let my board start turning the Rubik's cube”
Peter Thiel replaced PayPal's gray-haired CFO about three months before the IPO with Roelof Botha, roughly 28 years old and freshly out of Stanford GSB, because Botha knew the numbers and built a revenue cohort model that Sacks first saw there.
“Peter took that person to lunch and said, you're fired.”
Startups typically race onto a platform like Facebook for distribution and virality, then race off once the platform changes rules or competes, and those that cannot diversify before being cut off face very high risk.
“startups will race onto a new platform to get distribution... then racing off them to basically create a more secure position.”
The CEO's core obligation is to ensure the strategy is correct, because rowing hard in the wrong direction does not help, and an intense founder should subject their strategy to inspection by experienced people.
“if you're rowing in the wrong direction, rowing faster is not going to help you.”
A leader can push a team very hard only if they lead from the front and work at least as hard as the team, which gives them the credibility that Jordan and Elon Musk used to get people to respond.
“he wasn't asking anybody to work harder than he was working.”
Max Levchin served as PayPal's CTO, built the engineering team by hiring former University of Illinois classmates, and developed fraud-prevention systems that did not previously exist to stop fraudsters draining millions from PayPal.
“he kind of figured out the fraud problem at PayPal. PayPal was being taken for many millions of dollars by fraudsters.”
Sales reps are paid to close ARR, so without quality control they may promise features the product does not deliver, creating downstream churn that surfaces later in customer success rather than with the salesperson.
“A salesperson's job is just to close ARR. You have to essentially have quality control on your sales reps.”
Company culture reflects the founder's psychology because founders role-model behavior, so an aggressive founder produces a warlike culture, a cutting-corners founder produces a negligent culture, and an unbalanced founder produces an unbalanced company.
“company culture is a macrocosm of the founders psychology. Anything happening in the founders head... will ultimately be writ large across the company.”
To avoid everyone being under pressure at once, schedule the big product launch event in the middle of each quarter, when product and marketing peak, and let sales hit its deadline at quarter end, staggering the two by half a quarter.
“What I like to do is to stagger those things by half a quarter so everyone's hair not on fire at the same time.”
Keith Rabois was PayPal's bad cop who spun up efforts to alert antitrust authorities such as the FTC and DOJ to anticompetitive actions by monopolies like eBay, Visa, and MasterCard, which kept them from stomping on a small competitor.
“He spun up an effort to make sure that the antitrust authorities like the FTC, like the DOJ, were aware of any potential anti-competitive actions”
For consumer products, the founder must identify the single atomic transaction or interface that users will repeat, such as PayPal's email-plus-amount send, Uber's map pin summon, or Google's search field, because consumers will not engage with complex behavior before a simple one.
“It has to be something that consumers are willing to engage in over and over again.”
OpenPhone's individual product, though retaining well against benchmarks, still churned about 50 percent of revenue annually, meaning a company effectively rebuilds from scratch every two years, while its team product grew 100 percent annually instead of churning.
“if you have say 50% revenue churn over the course of a year, you're effectively rebuilding your company from scratch every two years.”
For enterprise SaaS, explosiveness is defined around lead generation: whether a company can sustainably create leads, because retained SaaS revenue lets companies afford customer acquisition spending, and consumer virality tactics can build a rapidly growing top of funnel.
“in the enterprise context, the explosiveness is more around lead gen. Do you have a way to sustainably create leads?”
PayPal grew virally on eBay, which was about 70 percent of its payments early on, but when eBay introduced its own payment solution the threat was existential, leading PayPal to sell to eBay, a dependency that later reversed when eBay spun PayPal out.
“PayPal went viral on eBay first. That was the first market that we really exploded in. And then eBay introduced its own payment solution”
Individual consumer subscription products typically churn 5 percent or more per month, so by year end a company keeps only about half its customer base and struggles to upgrade users, because single-player products offer no seat expansion.
“usually 5% plus per month churn. So what that means is by the end of the year, you're lucky if you still have half your customer base”
Burn Multiple is net burn divided by net new ARR over a period; for example, burning 5 million to add 2.5 million of net new ARR yields a Burn Multiple of 2, and a value under 2 is considered healthy.
“it's simply your net burn over your net new ARR in a given period of time.”
Burn Multiple was created so founders cannot pursue growth at any cost, because a 1 million dollar ARR milestone reached by spending 1 million is very different from one reached by spending 20 million, and the metric reveals how impressive growth really is.
“we want to make sure people aren't cheating on growth. Growth is still the most important thing.”
Pro-founder boards are not enough; the founder needs directors with relevant experience building startups who can test whether the strategy is correct, since all VCs now claim to be pro-founder.
“does your board have enough relevant experience, having built startups that they can actually give you advice on whether your strategy is correct?”
The Cadence is an operating system for SaaS companies that becomes necessary around 50 to 100 employees, when functional silos form and the founders can no longer coordinate everyone directly, and it works well through about 500 employees.
“The Cadence is an operating system for companies, specifically SaaS companies. It becomes necessary once you get above, say, 50 or 100 employees.”
Having co-founders helps because no single person has all the personality traits a healthy culture needs, so pairings like Jobs and Wozniak supply a yin and yang that balances company psychology.
“one person is unlikely to have all of the personality traits that you'd want reflected in a company culture.”
Sacks's firm requires SaaS companies to grow about 15 to 20 percent month over month below 1 million dollars ARR, tripling year over year between 1 and 5 million ARR, and at least doubling annually above 5 million ARR.
“Below a million dollars of ARR, we really want to see about 15 to 20% month-over-month growth. Between say one and 5 million of ARR, we want to see a company that's at least tripling year over year.”
B2B team products typically have about 1 to 2 percent monthly logo churn and net seat expansion, because companies that stay add seats faster than lost logos remove them, so the subscriber base keeps compounding.
“typically call it 2%, 1 to 2% logo churn a month. You actually have net expansion because you're getting more seats sold”
Companies with individual subscription products should pursue a team product as quickly as possible, because the team product economics are simply better even though the multiplayer experience is often unclear from a product standpoint.
“what I always encourage companies to do when they have more of an individual subscription product is figure out how do you get the team product as quickly as possible?”
Companies should bundle modest announcements such as financing, new apps, partner integrations, customer wins, and milestones into one launch event, since combined they look impressive and pierce the clutter, rather than dribbling out separate press releases over a quarter.
“you save up a bunch of announcements that, by themselves, wouldn't be that impressive, but when you combine them, actually it looks very impressive.”
Peter Thiel's superpower was identifying the handful of strategic decisions each year that would matter 10x or 100x more than everything else, delegating the rest, and recruiting strong people to run those roles.
“Peter's superpower was that he could identify the handful of strategic decisions every year that were going to matter 10X or 100X more than everything else.”
Product and engineering should release on a quarterly or seasonal calendar rather than weekly once a company is large, because frequent changes disrupt customers, while code releases can still happen continuously.
“as you get bigger, it becomes actually quite disruptive to your customers if everything's changing on a weekly basis.”
Products can add complexity over time to users who have already been hooked by a simple core interaction, but adding complexity first causes users to leave.
“You can graduate users up to complexity over time but only if you get them hooked on some aspect of the product.”
Sacks frames the future as a race between technological acceleration, which creates opportunity, and sociopolitical deterioration marked by division into warring political tribes, and he does not know which force will win.
“My line on this is that we're in a race for the future between technological acceleration and sociopolitical deterioration.”
When a software company has a physical-world component, it incurs real cost of goods sold and must build supply chains and physical operations, which is a much harder challenge for entrepreneurs, though Sacks is not categorically opposed to such investments.
“As soon as you're operating in the physical world, you're going to have real COGS, real cost of goods sold.”
Sacks begins founder evaluation with a product demo, following the principle to show a product, not a PowerPoint, because the same slides can describe many different products and the product reveals what the founder is actually building.
“Show us a product, not a PowerPoint.”
Sales teams work best on quarterly plans mapped to fiscal quarters, because monthly plans are too disruptive to predictability and annual plans leave too few opportunities for mid-course adjustment.
“What I find is that the best sales plans to put your team on are quarterly.”
Customer panels at launch events carry more credibility than the vendor's own claims, because prospects hear customers selling the product rather than the company selling itself.
“There's nothing better than customer testimony, because then prospects and other potential customers will hear your customers selling your product instead of you.”
It is much easier to start a company today than 20 years ago because of abundant VCs, angels, no-code and low-code tools, and standardized legal instruments like the SAFE note, which together streamline fundraising and remove the earlier need to figure out how to reach Sand Hill Road.
“It's much easier to start a company today than it was 20 years ago.”
Yammer's thesis was that enterprise software could be adopted bottom-up by end users, bypassing IT and top-down sales leaders, a model later called the consumerization of the enterprise.
“go appeal to sort of this consumerization of the enterprise is what it got named.”
Bill Gates' insight was that mass-market software has near-zero marginal cost per copy, so he could scale by selling more copies, making him the richest person in the world.
“The insight that Bill Gates had that made him the richest person in the world is that he could create mass-market software.”
Consumer subscription products have high churn and low willingness to pay, so companies cannot spend heavily on customer acquisition and must instead build large user bases through viral, organic growth.
“For consumer, you almost have to be explosive because you basically have to go viral as a consumer product because the typical consumer”
Over the last 20 years the tech economy has expanded from concentric circles around Stanford to spread across the Bay Area and then to every city, accelerated partly by COVID and remote work, alongside a large growth in venture capital.
“So you had tech takeover the whole Bay Area from San Francisco to the East Bay. And then it's now spilling out to every city, and partly because of COVID.”
Bessemer's Efficiency Score is the inverse of Burn Multiple (net new ARR over net burn), and the Hype Ratio is total capital raised divided by ARR, both related to capital efficiency measures.
“Bessemer has one called the Efficiency Score, which is basically just ... It's the flip side. It's net new ARR over net burn. There's one called the Hype Ratio”
SaaS businesses are highly benchmarked because everyone tracks the same metrics such as MRR, ARR, month-over-month growth, churn, expansion, and CAC, allowing comparison to industry averages, whereas consumer success is more binary and recognized by feel.
“Everyone's looking at the same metrics.”
On a return visit around 2000, after Mandela became president, Sacks found South Africa transformed, with apartheid gone and a widely optimistic mood, and he visited Robben Island where he impressed a tour guide by knowing Mandela's heroes from Long Walk to Freedom.
“Mandela had become president, and they completely transformed the country. Apartheid was gone, and things were very optimistic.”
Reid Hoffman ran PayPal's business development relationships with platforms such as eBay, Visa, and MasterCard, persuading them to avoid switching PayPal off, which mattered because those platforms could cut PayPal's access.
“Reid's job was kind of to make sure that all the platforms that could switch us off didn't”
Sacks's parents immigrated from South Africa to America when he was five, a decision he sees as a gigantic unlock for his opportunities, and they chose it deliberately after discussing where their kids would want to live.
“Immigrating to America was obviously a gigantic unlock for me to have opportunity in my life.”
Uber likely did not need taxi medallions because it did not pick up passengers on the street but was summoned like a town car, so it obtained TPS licenses instead.
“Did Uber need to go get taxi cab medallions? I think the answer was no, because they weren't stopping to pick people up on the street.”
After eBay acquired PayPal in 2002, eBay's corporate culture clashed with PayPal's freewheeling culture and drove out the PayPal group, scattering them into a diaspora that created new companies.
“eBay had a very different company culture. PayPal was very freewheeling. It was kind of like the old west and eBay was very corporate”
Aligning product and marketing works because product development generates the news that marketing communicates, so organizing product, engineering, and marketing around a shared seasonal release creates coordination.
“most news in a company is product created. This creates the alignment between product and marketing”
Talented teams want to be pushed and, even when a maniacal leader creates hard feelings, members rarely regret being part of a great franchise, as the Last Dance interviewees did not regret their time on the Bulls.
“I didn't hear anybody who was interviewed for that documentary regret their time on the Bulls.”
Burn Multiple is a can't-hide metric because it is high level and built from non-controversial numbers, making it hard for a founder to obscure the efficiency of growth.
“it's a 'can't hide' metric. You can't hide from this metric at all if you're a founder”
Even a franchise founder like Elon Musk needs a surrounding team that acts as a force multiplier, similar to how Jordan's team made him better.
“you still need a team around that person to be the force multiplier of that person.”
Founders find it easier to build a single-player mode before a multiplayer mode, so most products start individual and add team features later, though occasionally the order reverses.
“It's a lot easier for founders to conceive of and build a single-player mode very often before a multiplayer mode.”
Patrick notes that public equity CIOs and PMs rarely have been great entrepreneurs, so the PayPal group's crossover is unusual.
“in the public equity markets, you almost never find a great CIO or PM that was like a great entrepreneur.”