What is the curve, quarter by quarter?

The playbook starts at 80% blended gross margin, reaches 65 to 68% in quarter three, bottoms at 58 to 62% in quarter four with the sunset date announced twelve months out, recovers toward 65% in quarter five, and is back above 70% in quarter six. The migration I ran at a $40M ARR SaaS took 21 months from decision to sunset and was at 71% by month 22, when outcome revenue reached 91%.

Why is ServiceNow's 70.5% a migration margin?

Because its call, as Lemkin reports it, puts 50% of net new business on non-seat pricing. That is the middle of the road on my curve, the quarter where margin is still falling and the trough is ahead, and 70.5% is about where my blended margin sat in quarter three. The problem is not the number. Nothing on the call said where the bottom is or when the line turns.

What goes on the board slide?

Three slides every quarter for 24 months: the change, the curve with the trough drawn in before it happens, and the leading indicators, which are the share of net new on the new model, the share of revenue, and disputes against plan. Write the trough as a range and name the quarter the line turns. A company at 20% growth with those two numbers published is a different story from the same company without them.

How Far Gross Margin Falls in an AI Pricing Migration

How far should gross margin fall during an AI pricing migration?

THE SHORT ANSWER

On the migration I ran, from 80% blended gross margin to a trough of 58 to 62% in quarter four and back above 70% in quarter six, and the number that matters is not how far but whether the trough and the month were presented before they arrived. Per Jason Lemkin's October 31, 2026 post, ServiceNow's shares fell 26.9% after a quarter with subscription revenue up 21%, gross margin at 70.5% from 77.3% a year earlier, and 50% of net new business now on non-seat pricing, while Sierra posted $400M+ ARR with $150M+ added in the quarter and no margin line at all. Lemkin's rule is that investors accept compression when growth accelerates and punish it when growth is flat. The variable is not growth. Half of net new on the new model is the middle of a pricing migration, where my blended margin sat around the high sixties with the trough still ahead, and my board saw that curve every quarter for 24 months with the 58 to 62% bottom drawn in before it happened, so the number was a waypoint when it arrived. A falling margin with no stated bottom reads as permanent, and that is what ServiceNow's call left out.

From 80% to a trough of 58 to 62% and back above 70%, on my curve. The number that decides how the market reads it is whether the trough and the month were on a slide before they arrived.

ServiceNow's quarter, Sierra's post, and the three board slides are in ServiceNow Fell 27% on a 70.5% Gross Margin. Name the Trough and the Month.. The full curve by quarter and the 21-month bill are in Lovable Isn't Ready for Outcome Pricing. Here Is the 21-Month Bill., and the case for a margin line on every AI product is in Anthropic's S-1 Has No Gross Margin Line. Your AI Product Needs One..

Last reviewed 2026-11-02 · 1 min read

How far should gross margin fall during an AI pricing migration?

SOURCES

THE LONG VERSION

RELATED ANSWERS