49% in a Day. $3.6B Vaporized. Not Macro. Transition Physics.
On Oct 29, 2025, Varonis ($VRNS) fell ~49% ($63.00 → $32.34).
Management cited weak renewals across federal and non-federal on-prem subscription customers, sales process issues, and budgetary scrutiny. The public filings already show the core mechanism — one that surfaces in nearly every License-to-SaaS transition I analyze in PE diligence. In diligence, system-of-record exports let us test this signature in 48 hours.
The Failure Mode: Two Companies, One Number
Management said it themselves on the Q3 call: "Varonis is a story of two companies." But they meant it as framing — SaaS momentum masking on-prem drag. What they didn't say was that the decline rate was worsening every quarter.
Look at the growth rate chart. SaaS QoQ growth held in a tight band through Q3: +19%, +18%, +14%. Healthy. Consistent. Not the problem. Now look at the other line. Non‑SaaS decay worsened each quarter into Q3: -14%, -17%, -20%. That’s the signature: the leak rate deteriorates as the renewal base shrinks. And yet the headline metric — total ARR — still printed double-digit year-over-year growth.
The Compounding Factor: Vertical Concentration
Federal was only ~5% of total ARR — explicitly confirmed on the Q3 earnings call. But the hypothesis is that federal was disproportionately represented in the remaining non-SaaS renewal pool. If true, even a small federal cohort could punch far above its weight when renewals collapsed in the final two weeks of the quarter.
This is a testable inference, not a confirmed fact from public filings. The table that proves or disproves it: renewal cohort by vertical × delivery model (SaaS vs. non-SaaS). That single cross-tab tells you whether you're buying a geographic/vertical problem or a structural transition problem — two very different valuations.
The Private Equity Application
If you are buying a platform "migrating to cloud" or "moving up-market," top-line ARR growth is the wrong number. It hides the decay rate in the legacy base.
You must apply Behavioral Revenue Forensics (BRF) to run a Cannibalization Waterfall: strip out migrated revenue and isolate organic net retention by cohort.
If the non-SaaS renewal pool is bleeding faster than SaaS bookings can replace it in absolute dollars, the forward ARR assumptions are wrong — and the valuation multiple built on them is wrong too.
The public filings already showed the signature. In diligence, system-of-record CRM exports let us test these renewal mechanics in 48 hours — before you are committed to the price.
I just finished a forensic backtest of 52 public SaaS forecast failures — $314B in market cap losses. 36 of 52 (69%) were detectable from CRM data before the close. Varonis is Case 1.