Welcome to the SaaSpocalypse

Methodology

The dataset tracks the lowest-priced monthly subscription published on each vendor's default pricing page in January 2020 and April 2026. 2020 prices were taken from the closest available Wayback Machine snapshot to January 2020. 2026 prices were taken from live vendor pricing pages captured in April 2026.

Inclusion rules: a tool was included only if it published a monthly-billed entry-tier price on its default pricing page in both years, and if its pricing model remained directly comparable between the two dates. Tools that moved from seat-based to consumption-based pricing (or vice versa) were excluded as non-comparable. Tools where the 2020 Wayback snapshot was incomplete, missing, or ambiguous were excluded.

Price selection: the lowest-priced paid standalone subscription, on monthly billing (not annual-billed-monthly-equivalent), exact figure as published. Free tiers, trials, usage add-ons, SMS and AI credit packs, and per-user fees beyond the base plan were excluded. Where a vendor showed both monthly and annual billing, the monthly figure was taken in both years. Annual discounts have widened at many vendors during the period, which would obscure underlying price movement.

Category assignment: each tool was assigned to a single category based on its primary product function in 2026. Categories with fewer than three tools were excluded, with one exception: Visual Design (n=2, Framer and Venngage) was retained as a standalone category because both tools directly represent the creative design software segment most exposed to generative AI substitution.

AI-replicability classification: categories were labelled as "Replicable" or "Resistant" based on whether the core product function could reasonably be approximated by a small team using 2026 AI coding assistants and generative AI tools. This is a judgement label, not a measurement. Marketing is labelled "Mixed" because the category contains tools with materially different replaceability profiles.

Inflation comparison: US Consumer Price Index (all urban consumers) rose 27.1% between January 2020 and March 2026, per Bureau of Labor Statistics data.

Limitations

List price is not always the price paid. Enterprise discounting, partner pricing, and promotional credits sit outside this dataset and can shift effective realised price in ways the published pricing page does not show. The list-price comparison is the cleanest like-for-like read available without privileged access to vendor billing data.

The AI-replicability label is a judgement, not a measurement. Where a category sits on the replicable-to-resistant spectrum will continue to shift as AI capability advances. Today's resistant category can become tomorrow's replicable one. The current classification reflects an informed view of 2026 capability, not a settled framework.

Pricing-page design itself changed across the window. Several vendors moved from clear three- or four-column pricing pages to "contact sales" gating, particularly at higher tiers. The dataset only includes tools whose entry tier remained publicly listed in both 2020 and 2026, which biases inclusion toward smaller and mid-market vendors and away from vendors that retreated into private quoting.

Sample size by category varies. Categories with three or four tools (Comms, Productivity, Analytics) carry more single-point sensitivity than categories with seven or eight (Project Mgmt, Ecommerce, Marketing). The directional finding (AI-replicable categories cluster at or below zero, AI-resistant categories cluster well above CPI) is robust to small-sample noise at the category level.

The 65-tool sample is a sample, not a census. Coverage skews toward SMB and mid-market SaaS with publicly listed pricing. Enterprise SaaS, infrastructure platforms, and vertical-specific tools are under-represented. Conclusions describe the publicly priced mid-market layer of SaaS, not the full software economy.