S01.2 · Technology & Digital Infrastructure
Industry-specific SaaS for sectors like healthcare, construction and legal, sized between $500B and $800B depending on scope.
Vertical Software is SaaS built around one industry's workflow — healthcare, construction, legal, restaurants, field services — rather than one function sold everywhere. Sizing for 2025 runs from $500B (Vertical Market Software Global Market Report, core licensing) to ~$800B (Bessemer, including embedded fintech); the spread is scope, not disagreement. Either way the segment is outgrowing horizontal SaaS, and for a reason horizontal SaaS cannot copy: the vendor that owns a vertical's workflow also sees its transactions, and increasingly gets paid on them.
Growth has run ~15-18% and should hold in the mid-teens. Demand concentrates in the US, a function of the fragmented, SMB-heavy industries the segment serves. The fragmentation is by design, not immaturity: Toast, Procore, Veeva, ServiceTitan and Clio each dominate a narrow niche instead of contesting one category, with a long tail of sub-scale vertical players underneath. Regulatory gating varies sharply by vertical — HIPAA in healthcare, financial-services rules elsewhere — and that variance is part of what keeps the category fragmented; compliance depth is a moat that does not travel between verticals.
The economics carry a premium to horizontal SaaS, but the premium comes from distribution, not product: niche dominance keeps customer-acquisition cost low, and embedded payments add take-rate revenue on transaction volume the vendor already sees. For a growing share of these businesses, the take rate on processed volume matters as much as the per-seat license price — which changes what an acquirer should be underwriting. Upstream, the segment rides horizontal rails it did not build (payments, communications APIs); downstream it extends into fintech and data monetization, where the incremental margin now sits.
The two expansion paths differ in kind. Payments attach is new money — take-rate revenue on volume already flowing through the platform. Moving into an adjacent vertical that shares a workflow (field service into construction) is defense: it pre-empts the neighbor doing the same in reverse, and opens little that is genuinely new.
Agentic AI attacks the segment's original pricing logic. Scheduling, dispatch and back-office data entry are labor-arbitrage functions — the seats existed because staff did this work manually — and as agents absorb the work, the seats it justified go with it. The deeper exposure is the system-of-record moat itself. Vertical software has won by being the one place a business's operational data lives; agents that read and write across systems lower the switching cost that position protected. Compliance and workflow depth slow the substitution — the segment's own sources put material agentic substitution two to five years out — but slowing is not blocking, and the vendors that convert workflow ownership into transaction economics before then will care much less when it arrives.