In this podcast, venture capitalist Eric Vishria explains the evolution of SaaS (software you pay for monthly, not buy once). He says high SaaS valuations come from multiple expansion, not profit upgrades—investors pay more for the same revenue. Key examples: Stripe (payment API, praised for elegant design), Salesforce (early SaaS giant, 170x bigger than old rival Siebel), and Twilio (API-based, charges per usage).
Eric Vishria (Partner at Benchmark Capital) discusses the past, present, and future of the SaaS and software industry on the program. The core view is that the high valuations of current public software companies stem from their predictable recurring revenue and high customer retention rates, and th
Eric Vishria (Partner at Benchmark Capital) systematically reviewed the three generations of SaaS evolution in the episode and pointed out that the core reason for the current high valuations of SaaS companies is not upward estimate revisions but multiple expansion—carrying a valuation of 10–11x forward revenue, the entire Bessemer Emerging Cloud Index (100+ SaaS companies) delivers a "Rule of 40" performance (35% growth + 7–8% free cash flow margin) sufficient to support this multiple. He believes that COVID pushed digitalization from "nice-to-have" to "life-or-death," and SaaS penetration (approximately 15% of total IT spending) still has significant room for growth.
Vishria believes that the surge in the SaaS sector (Bessemer Emerging Cloud Index up 50% YTD) during the pandemic was not driven by broad upward revisions in corporate forecasts—with few exceptions like Zoom, most companies’ 2020 guidance either remained flat or was withdrawn. What actually occurred was multiple expansion: these companies currently trade at an average of 10–11x forward revenue, while simultaneously exhibiting a 35% revenue growth rate and a 7–8% free cash flow margin. The entire index collectively has surpassed the “Rule of 40” (growth rate + profit margin ≥ 40%).
He explains this phenomenon using the “unicorn business model”: SaaS creates a win-win between sellers and buyers. For sellers: software-level margins plus compounding effects; for buyers: elimination of the hassle of operations, upgrades, and maintenance, allowing them to focus on their own customers and core competencies. Vishria also notes that traditional software companies expend roughly 30% of engineering resources on backward compatibility and supporting legacy versions, whereas the SaaS model allows all engineering resources to be directed forward—creating an “invisible” acceleration effect.
Vishria cautions readers that he is a partner at a venture capital firm, and his perspective is naturally biased toward the sectors in which his firm invests. The above analysis should be interpreted as his interpretation after adopting a “long-term bullish on SaaS” stance.
Vishria proposes a three-generation SaaS framework, with the core difference lying in the shift of the “competitive frontier”:
| Generation | Time Window | Core Innovation | Pricing Model | Customer Characteristics | Representative Companies |
|---|---|---|---|---|---|
| Gen 1 | 1999–2005 | Better delivery model (vendor-hosted) and economic model (subscription replacing one-time license + 18% annual maintenance) | Seat-based | Large enterprises, large sales, single instance | Salesforce, Workday, ServiceNow (mostly re-created by the same founders) |
| Gen 2 | ~2005–2015 | Better adoption model: from “macro adoption” (enterprise-wide decision) to “micro adoption” (individuals or small teams can try, try-and-toss) | Seat-based, but lower initial price | Mid-market, SMB, marketing-driven | Zendesk, New Relic, Wix, Asana, Twilio, DocuSign |
| Gen 3 | ~2015–present | Business “encoded into software,” software interacts with software via API, not GUI with humans | Transaction/usage-based | Developers, embedded in other products | Stripe, Twilio, Contentful, Modern Treasury, Duffel |
Vishria emphasizes that the “delivery + economics” advantage of Gen 1 made Salesforce two orders of magnitude larger than Siebel (approximately 170x). Gen 2’s “micro adoption” triggered the “Cambrian Explosion” of SaaS—hundreds of SaaS applications emerged within companies, and any repeatable function was “SaaS-ified.” For Gen 3, the competitive frontier has shifted from UI experience to API design quality. Using Stripe as an example, the ease of use and design sophistication of its API form its core moat.
Vishria cites a blog post by Confluent founder Jay Kreps, arguing that enterprises are moving from “using software” to “being encoded into software.” When business processes are encoded, software needs to call upon other software, making APIs (Application Programming Interfaces) the necessary interface—rather than GUIs.
He illustrates with a loan approval process: from a weeks-long manual process (“submit documents → loan officer operates → credit committee meets”) to “user clicks a button online → software automatically retrieves bank and mortgage company data → machine scores → instant feedback.” APIs are the “Lego bricks” of this automation chain; a decade ago there were only a few pieces, but now they are proliferating.
The pricing model has also shifted: the seat-based pricing of Gen 1/2 (charging per person) has moved to transaction/usage-based pricing (charging per call or per transaction amount). Vishria believes that all traditional SaaS companies need to build an API strategy, or they will lose the ability to communicate within the ecosystem—GUIs will not disappear, but APIs will become the primary consumption path.
Vishria believes that although Gen 2 SaaS achieved multi-tenancy, it remains in “customer silos” in terms of data utilization—Customer 1,000 does not derive any value from the data of Customers 1–999, aside from the product’s own maturity. He identifies an underrealized opportunity: anonymizing and abstracting multi-tenant data so that each customer can see industry benchmarks and compare their performance, thereby generating a quasi-network effect.
He cites Stripe Radar as an example: when an IP address is flagged as fraudulent by one customer, Stripe can automatically apply it to all customers. Vishria notes this is a prototype of the “cross-customer perspective,” but the vast majority of SaaS companies have not yet systematically developed this asset.
At the same time, he references partner Sarah Tavel’s “Marketplace Hierarchy” framework, arguing that there are significant crossover opportunities between SaaS and Marketplaces: HackerOne (a bug bounty platform) is a classic case—it embodies both the supply-demand matching logic of a Marketplace and charges in a SaaS manner. He suggests that every SaaS entrepreneur should study the moat logic of Marketplaces, and every Marketplace entrepreneur should study the operational logic of SaaS.
Vishria believes that the long-term opportunity for SaaS is far from tapped, with two core arguments:
Vishria believes that “every company is becoming a software company”—this trend has not reversed, and COVID has only accelerated it. Digitalization has shifted from “nice to have” to “a matter of life and death.”
| Position | Analyst View | Key Data |
|---|---|---|
| Salesforce | Historical Reference (Gen 1 Representative) | Approximately $170B market cap, compared to Siebel ~$3B (approx. 170x) |
| Workday | Historical Reference (Gen 1 Representative) | Replaced PeopleSoft, valuation approx. 5x |
| ServiceNow | Historical Reference (Gen 1 Representative) | Replaced Peregrine, valuation approx. 150x |
| Shopify | Case Study (Gen 2 → Cross-network Effects) | Currently adding Marketplace elements, connecting end customers |
| Twilio | Case Study (Gen 3 API Representative) | API as the primary consumption path, transaction volume pricing |
| Stripe | Case Study (Gen 3 API Design Benchmark) | The sophistication of API design is its core moat |
| Confluent | Investment Case (Benchmark Early Investment) | Citing founder Jay Kreps' blog "Business Coded into Software" |
| Amplitude | Investment Case | Quantifying product management, measuring user behavior |
| Chainalysis | Investment Case | Blockchain data analysis, quantifying forensic analysis |
| HackerOne | Investment Case (SaaS+Marketplace Hybrid) | Bug bounty platform, SaaS pricing + Marketplace matching |
| Zendesk | Historical Reference (Gen 2) | No specific data provided |
| New Relic | Historical Reference (Gen 2) | No specific data provided |
| Asana | Historical Reference (Gen 2) | No specific data provided |
| Zoom | Case Study (The only exception with clearly raised forecasts) | No specific multiple provided |
| Anaplan | Case Study (New Specialization) | Approximately $7B market cap |
| Zuora | Case Study (New Specialization) | Subscription billing, multi-hundred-million-dollar market cap |
| Salesforce (Platform Strategy) | Case Study (SaaS+Marketplace Cross) | Built an ecosystem moat via AppExchange |
| Nike | External Case Study | In 2020Q2, digital + direct sales revenue accounted for 30% (originally a 2023 target), now raised to 50% |
1. “Multiple expansion, not estimate revisions, drove SaaS valuations” (Vishria): Benchmark analysis finds that beyond Zoom, most SaaS companies did not raise their 2020 guidance, yet the entire Bessemer cloud index rose 50%, supported by the fact that the entire index had surpassed the Rule of 40 (35% growth + 7-8% free cash flow margin).
2. “SaaS is a unicorn business model—win-win, not zero-sum” (Vishria): For sellers—software-grade margins plus compounding; for buyers—no operations, no upgrade burden, focus on core business. Traditional software allocates 30% of engineering resources to backward compatibility; SaaS frees all of that for forward progress.
3. “Three generations of SaaS evolution: delivery model → adoption model → embedded model” (Vishria): Gen 1 replaced traditional software with better delivery and economics (Salesforce, etc.), Gen 2 opened the mid-market/SMB through “micro-adoption” and try-and-toss (Shopify, etc.), and Gen 3’s competitive frontier lies in API design quality.
4. “API means software interacting directly with software, not GUI with humans” (Vishria, citing Jay Kreps’ blog): When enterprises are “encoded into software,” APIs become the necessary interface. Pricing accordingly shifts from per-seat to transaction/usage-based—who is using the software changes from “people” to “software.”
5. “Cloud and SaaS represent only 15% of total IT spend, with 85% still penetrable” (Vishria, citing Tomasz Tunguz’s data): Even after the COVID acceleration, this ratio has vast room to rise, and moving from 15% to 60% is conceivable.
6. “SaaS companies have not yet fully exploited the ‘cross-customer perspective’ of multi-tenancy” (Vishria): Customer 1,000 ought to benefit from the data of customers 1-999 (anonymized, benchmarked), but current practice is extremely limited. Stripe Radar’s cross-account fraud detection is one of the few early examples.
7. “Marketing goes from ‘Mad Men’ to ‘Moneyball’—this quantitative transformation is permeating all functions” (Vishria): Product management (Amplitude), legal analytics (Chainalysis) all undergo the same qualitative-to-quantitative shift; this is the foundation for SaaS creating new markets.
8. “The boundary between SaaS and Marketplace is blurring, and entrepreneurs should learn from each other’s frameworks” (Vishria): HackerOne charges like a SaaS company but operates on marketplace logic; Salesforce builds a platform ecosystem through AppExchange; Shopify is adding an end-customer marketplace feature. Recommended reading: Sarah Tavel’s “Marketplace Hierarchy” framework.