Looking back at Silicon Valley in 2001, it’s hard to avoid a sense of contradiction. The dot-com bubble had burst, but the ecosystem itself was, if anything, accelerating. While many high-profile failures grabbed headlines, an undercurrent of new startups persisted—some on borrowed time, others already laying the groundwork for the internet’s next phase. For economists and statisticians, measuring the contours of this landscape poses a challenge. ISIC 6201—computer programming activities—is the statistical tool at hand. Yet, for all its precision as a global category, it remains just a rough proxy when trying to single out “dot-com” companies from the sea of software and IT ventures that always seem to populate Silicon Valley.
The starting point is surprisingly mundane: extracting the set of firms registered under ISIC 6201 within the relevant California counties. Business registries, state incorporation records, and, when available, local economic development databases collectively offer a population that is, frankly, overwhelming in size and diversity. Within this, a small but significant subset are the “dot-coms”—companies that built their business on web-based delivery or commerce, often prioritizing growth over immediate profit.
Identifying which firms to count as genuine dot-com startups requires filtering. A close reading of business descriptions, founding press releases, and even early marketing copy is essential. Many startups, in 2001, still positioned themselves as “solutions providers” or “internet platforms,” shying away from the “dot-com” label after the high-profile collapses. Yet the underlying model—scaling fast by leveraging the web—remained unchanged. Web archives, such as the Wayback Machine, are invaluable here, preserving site snapshots and business models long since erased from official memory.
The next step is to integrate venture capital (VC) data, which both complicates and clarifies the picture. Firms that raised funding rounds from recognizable Valley investors—Sequoia, Accel, Kleiner Perkins, and a handful of others—tend to be the most visible. Public VC databases, press coverage, and S-1 filings for the few that made it as far as an IPO, all help tie funding events to registered entities. In some cases, the VC round predates formal company registration; in others, a pivot or merger means that a business exists on paper long before attracting capital. Disentangling these timelines often requires comparing dates across multiple sources—state filings, funding announcements, and internal company documents if accessible.
Many dot-com startups, especially in the lower funding tiers, attracted angel or seed capital that was less formally documented. Some received convertible notes, others operated on founder savings until later rounds. It is tempting to restrict the dataset to companies that closed a Series A or higher, but this risks omitting much of the true early-stage experimentation that defined the period. For a more inclusive approach, smaller VC rounds, incubator involvement, and participation in high-profile demo days (such as Y Combinator’s earliest cohorts) all serve as secondary signals.
Mismatches abound. Company names and legal entities sometimes change within months of formation. Acquisitions, mergers, and even stealth mode launches can obscure the true scope of the sector. There is no perfect algorithm for reconciling all these records. Instead, analysts need to document each assumption: which sources were used to resolve conflicts, which companies were included or left out, and why. Erring on the side of inclusion, with careful notes, can allow for more meaningful future revision as new sources become available.
Connecting registrations to funding isn’t only about chronology—it is also about mapping momentum. Tracking the time lag between firm formation and first VC round offers insights into market conditions, founder networks, and perhaps even prevailing investor sentiment. In the 2001 climate, some firms launched and raised funding in a matter of months, while others lingered in a sort of limbo, adjusting business models until capital could be secured.
Patterns begin to appear, though not always neatly. Startups with experienced founders or strong VC backing moved quickly. Others fell away or reconstituted themselves after initial stumbles. Industry subcategories—B2B marketplaces, web publishing, SaaS tools—showed varying degrees of resilience or fragility, and the data, when cross-referenced with funding rounds, gives a sense of both churn and endurance.
None of this work is entirely satisfying. Gaps in the records persist, and some of the most important future successes of Silicon Valley came from companies barely visible at this time. Still, through ISIC 6201, venture capital records, and a willingness to embrace uncertainty, it’s possible to trace the faint but unmistakable outlines of a sector reinventing itself—less a graveyard than a seedbed for the internet economy to come.