ServiceNow's Record 27% Drop Priced the App-Software Discount: Its Premium to Salesforce Fell From 2.1x to 1.5x

ServiceNow reported third-quarter results after the close on October 28 and closed the next session at $109.67, down 26.95% — its biggest one-day drop on record. Total revenue was $4,106 million, subscription revenue $3,988 million (up 20.5% in constant currency), and the company raised full-year guidance and still fell 26.95%. Full-year 2026 subscription revenue guidance went to $15,780–15,790 million, the free-cash-flow margin target was lifted to 36%, and the only outright miss was adjusted EPS of $1.00 against roughly $1.02 expected. At least 12 brokerages cut price targets anyway, with Raymond James pointing to lagging subscription revenue and backlog growth while AI initiatives were still yet to materialize.

Revenue guidance and revenue bookings are telling two different stories, and the market only prices one of them. ServiceNow raised the current year and pointed the next booking window lower: Q4 current remaining performance obligations are guided to grow 18.5% in constant currency, against 21% in Q3 and 21.5% in Q2. Because roughly 46% of the $29.7 billion total remaining performance obligation converts into revenue within twelve months, a 250-basis-point step-down in the book is a direct forecast of slower subscription revenue two to three quarters out.

ServiceNow ended October 29 with a market capitalization of $113.4 billion and an enterprise value of $119.4 billion against trailing revenue of $15.43 billion — 7.7 times sales. On the October 28 close of $150.13, the same trailing revenue was capitalised at 10.4 times. In one session the market removed 2.7 turns of sales from a company whose revenue guidance it had just raised, lifting the free-cash-flow yield from about 3.1% to 4.2% and cutting the premium ServiceNow commands over Salesforce from 2.1 times to 1.5 times.

The cleanest way to see what the market did is to divide the multiple by the growth it buys. On that measure Salesforce is now the most expensive seat-priced app per point of growth at 0.47 times, ahead of ServiceNow at 0.38 times and Adobe at 0.29 times, with Oracle at 0.26 times on 29.6% revenue growth. On cash flow the gap is just as wide: ServiceNow trades at 25.1 times trailing free cash flow against 14.7 times for Salesforce, 17.6 times for Workday and 9.3 times for Adobe.

A beat-and-raise can lose $42 billion when the company is renegotiating its own revenue model in public and the new model is harder to forecast. Management now describes a hybrid of seat subscriptions and consumption pricing, with the chief financial officer noting that more than half of net-new business already comes from non-seat models, while customers still prefer seat-based predictability and are not asking for outcome pricing. The bull case is arithmetic — management has floated the potential for a 5x improvement in revenue per user. The bear case is equally arithmetic: consumption revenue is lumpy where seats were recurring, and a 5x revenue-per-user prize on an unmetered consumption base does not support the same multiple.

The GAAP figures investors looked through on the quarter are not a rounding error; they are the third leg of the same story. ServiceNow closed its $7.6 billion cash acquisition of Armis on April 20, 2026, funding it with a $4.0 billion term loan repaid in May through $4.0 billion of senior notes, and it still carries $2.1 billion of commercial paper at 4.05% with a weighted-average 62 days to maturity. Goodwill of $5,323 million and intangibles of $2,530 million came with it — together more than the $7,637 million of net assets actually acquired — so the deal converted almost the entire purchase price into amortization: $198 million in Q3 alone against $33 million a year earlier, with $783 million scheduled for 2027. Interest expense went from $6 million to $81 million in the same quarter.

The same session that took 27% out of ServiceNow left the agent economy's other two layers almost untouched, which is the clearest evidence that the market repriced a business model rather than a quarter. NVIDIA fell 0.87%, Microsoft slipped 0.16%, and Accenture rose 0.91%. Money moves from the layer being disintermediated to the layers that capture agent spending.

Oracle's 48% drawdown came with 29.6% revenue growth, and it still trades at 7.5 times sales — the same multiple ServiceNow carries after the crash. That is the sorting mechanism at work: the market pays for growth and discounts seat-based terminal value. ServiceNow's own path proves the point from the other direction — it fell 60% from its 52-week high to $81.24 on April 10, 2026, recovered to $142.35 by September 14 and $150.13 by October 28, and then gave back more than a third of that recovery in a single session without a single revenue miss.

Over the next two quarters the debate is narrow and testable. If ServiceNow is going to defend a 7.7 times sales multiple, the Q4 print has to show the book stabilising: the 2026 AI contract value target is $1.7 billion, production agentic consumption is up more than 13 times year to date, and Control Tower net new annual contract value grew more than sixfold. Any slip in that target, or another step down in current remaining performance obligations toward the 18.5% guided for Q4, hands the bears their next leg.

ServiceNow still renews 98% of its revenue, grew subscription revenue 20.5% in constant currency, guides to a 36% free-cash-flow margin and trades on a 4.2% trailing free-cash-flow yield. Nothing in the print says the business broke. Everything in the tape says the market stopped paying for its terminal value, and the discount has only partly landed: 7.7 times sales is still 53% above Salesforce and 103% above Adobe, and 25.1 times trailing free cash flow against their 14.7 and 9.3. For the next two quarters this is not a debate about whether software gets disrupted. It is a debate about how much of the disruption is already in the price of the company that sells the workflow layer, and how much of it still has to be paid for in Salesforce, Workday and Microsoft multiples that barely moved.