Q3 2026 SaaS Guidance Scorecard: Who Raised, Who Cut, and Why the Raisers Fell

Of 15 software companies that reported a July to September quarter between 20 and 29 October 2026, 10 raised a full-year guidance line. Four of the five stocks that fell more than 15% the next session were among those raisers. ServiceNow lifted its subscription outlook by $15 million and closed down 26.95%, its worst day on record. Q3 2026 SaaS earnings punished something, and it was not the guidance number.

Short answer. Which SaaS stocks fell after Q3 2026 earnings? ServiceNow, F5, Fortinet, Alkami and Twilio had the worst next-day moves among 15 filers, each down 15% or more. Four of them raised full-year guidance. Their common trait was a slower forward metric, such as backlog growth or next-quarter organic growth, rather than a shared business model.

The Q3 2026 SaaS guidance scorecard, 15 filers

Every row below is a company that reported a July to September quarter between 20 and 29 October 2026. The guidance columns compare the full-year range in the Q3 release with the range in the Q2 release. The reaction column is the close-to-close move on the first full session after the release.

Click any column heading to sort the table. It opens sorted by the next-day move, worst first, because that ordering is the whole argument of this post.

Ranges are taken from each company's Q2 and Q3 earnings releases filed as 8-K or 6-K exhibits. Midpoint change is our calculation. Moves are close-to-close on the first full session after the release; Roper and Commvault reported before the open, so their move is the same day. Price data is S&P Global Market Intelligence via stockanalysis.com, retrieved 31 October 2026.

How the Q3 2026 SaaS guidance scorecard was built

The selection rule was simple. We took software companies with a calendar July to September quarter whose results were filed with the SEC by 30 October 2026. Of 25 names we checked, 15 had reported, and the other 10 either report in November or are being acquired.

That rule excludes the companies most people mean by "SaaS earnings". Salesforce, Workday and Snowflake run fiscal years that put their next reports in late November and December. Atlassian reports on 5 November, Dynatrace on 3 November and Check Point on 2 November. Microsoft reported on 28 October but gives quarterly rather than full-year guidance, so it has no row.

What counts as a guidance change

We compared the full-year range in the Q3 release against the range in the Q2 release, using the filed exhibit in both cases. Where a company guides total revenue, that is the line we used. Where it does not, we used its headline line: subscription revenue for ServiceNow and Commvault, constant-currency growth for IBM and Twilio, cloud revenue for SAP, adjusted EPS for Roper.

The midpoint change is new midpoint minus old midpoint, divided by old midpoint. Fortinet is the clearest example of why one line is not enough. Its total revenue midpoint rose from $8,100 million to $8,170 million, while its service revenue midpoint fell from $5,200 million to $5,160 million in the Q3 release filed on 28 October.

What counts as the market's verdict

We used the close-to-close move on the first full session after the release. After-hours prints were unreliable in this cycle. IBM fell 6.5% before the open on 22 October, according to Reuters, then closed up 0.26%. Tenable dropped 6.9% after hours on 28 October and closed the next day up 6.28%.

A close-to-close move still mixes the company's news with the whole market's day. The IGV software ETF fell 3.52% on 29 October, the session that absorbed ServiceNow, Fortinet, Tenable, Tyler and Alkami. We report raw moves and note the sector move where it matters, because adjusting for a single ETF would add false precision to 15 data points.

Raising guidance did not protect software stocks this quarter

The 10 raisers averaged a next-day move of -5.6%. The four filers that held, cut or gave no guidance averaged +1.1%, leaving F5 aside because a new fiscal year guide is not comparable. The median move across all 15 was +0.26%, which tells you most stocks barely moved and a handful moved enormously.

Our view is that the raise had become the price of admission rather than news. When revenue growth across public SaaS has been steady at around 20%, a small upward nudge is what investors already assume. It confirms the year. It says nothing about next year, which is what the stock was trading on.

The five worst reactions were mostly raisers

ServiceNow, Fortinet, Alkami and Twilio all moved a full-year line up, and all four fell between 15% and 27% the next day. The fifth, F5, beat the top of its own fiscal 2026 range with 11% growth before guiding fiscal 2027 to 8% to 9%. It fell 20.43%, which Barron's reported as F5's biggest one-day drop in more than 13 years.

The working behind the correlation is short enough to show. Pairing each midpoint change with its next-day move across the nine dollar-range filers gives a Pearson coefficient of 0.34. Squared, the raise accounts for about 11% of the variation in the reaction. With nine points, even that weak link could be noise.

The only cut, and the only clear miss

Tyler Technologies took $10 million off the top of its revenue range in its Q3 release, a midpoint cut of 0.20%. Its shares fell 5.8% after hours, per Investing.com, and then closed the next session up 3.61%. Pegasystems missed revenue estimates by roughly $28 million, giving no full-year guide at all, and fell only 2.90%.

We would not read either case as investors rewarding bad news. Both stocks had sold off in the sessions before their reports, so expectations were already low. That is the point. The reaction was set by what investors expected relative to what the release implied about 2027, and the guidance delta was a poor proxy for both.

ServiceNow and IBM: two beats, two very different sessions

The two names that framed this season are worth separating, because the popular version of the story is half right. Both companies beat on revenue. Only one of them actually fell on a closing basis.

ServiceNow: a raise the market read as a stall

ServiceNow reported subscription revenue of $3,988 million, up 21%, and lifted its 2026 subscription range to $15,780 million to $15,790 million from $15,760 million to $15,780 million in its Q3 release. It also raised its free cash flow margin target to 36% from 35%. Adjusted EPS of $1.00 missed the $1.02 consensus by 2 cents.

The shares fell 26.95% close to close. CNBC reported that the results reignited concerns about AI disruption, and quoted Citi's Tyler Radke on "the lack of upside". The line that moved was the Q4 cRPO guide of 18.5%, against 20.5% growth in Q3. An earlier version of the same pattern, after Q1, is covered in the rebuild of the ServiceNow bear case from its own filings.

IBM: the premarket drop that did not hold

IBM reported revenue of $17.1 billion, up 5%, and software up 8%. Its release kept full-year constant-currency revenue growth at 4% to 5%. It now expects free cash flow to rise by more than $1 billion year over year. Reuters reported a raised software growth range of 7% to 8%, from 6% to 8%, given on the call.

The doubt was about quality. Reuters cited BofA analysts saying about $250 million of software growth came from deals pushed out of Q2. The stock fell 6.5% before the open and then closed up 0.26%, having already risen 3.94% on the day of the report. IBM belongs in the "held" column, and on a closing basis the market did not punish it.

Do the punished SaaS stocks share a business model?

The hypothesis going in was that investors were pricing one specific risk. If AI makes it cheaper for customers to build their own software, seat-priced horizontal platforms should suffer most. The table below tests that against the five worst and five best reactions.

The model column is our classification from each company's own segment and revenue disclosures, simplified to one line. Alkami's forward line is the company's Q4 guide; we did not find a source that stated whether it disappointed.

Seats, hardware and usage all lost

The answer is no. The five worst reactions include one seat-heavy platform, two companies with large hardware lines, one usage-metered API business and one vertical software vendor. If the market were pricing a single AI threat to seat pricing, Fortinet and F5 would not sit beside ServiceNow at the bottom.

The seat-pricing worry is real and is examined in the analysis of seat compression in SaaS pricing. This scorecard simply does not isolate it. Seat-heavy SAP and Tyler rose, and two of the three worst falls came from companies that sell boxes.

Vertical software mostly held up, with one exception

Roper, Tyler, Cadence and Manhattan, all of which sell deep vertical or engineering software, finished up. That fits the argument that vertical software is better defended than horizontal tools. Alkami spoils the pattern. It is as vertical as software gets, raised guidance, and fell 18.29%.

We would hold this one loosely. Four winners and one loser is not a finding, and the vertical names that rose also had the least exposure to the AI spending debate that dominated the ServiceNow coverage.

What the losers shared: a slower forward line

Four of the five worst reactions carried the same signal somewhere in the release. A forward-looking measure grew more slowly than the quarter just reported. The full-year raise was backward-looking by construction, because three quarters of the year were already booked.

Twilio is the cleanest case. Its Q3 release raised 2026 revenue growth to 20% from 18% to 18.5%, and Q4 revenue guidance of $1,595 million to $1,605 million. Proactive Investors calculated that the Q4 range implies 11% to 12% organic growth, against 16% in Q3. The stock fell 15.24% the next session.

Fortinet's signal was smaller and more specific. Bloomberg Intelligence, quoted by Yahoo Finance, said the service revenue cut implied lower attach rates for bundled services on its firewalls. The stock had also risen about 120% in 2026 before the report, per the Motley Fool, so expectations were stretched.

Our position is that this is the real scorecard. Backlog and next-period growth are where AI substitution would show up first, if it shows up at all. The broader pattern of backlog across vendors is laid out in the comparison of cloud backlog and RPO, and it is the line we would read before any full-year range.

Where this scorecard is weakest

This is a small, early and uneven sample, and the conclusions above should be read with that in mind.

Fifteen companies is not the software sector. The biggest SaaS names report in November and December, and any of them could break the pattern. The guidance lines also differ. Comparing a subscription range with an EPS range and a growth rate is a judgement call, which is why the correlation uses only the nine dollar-range filers.

Fifteen filers and one month is a small sample

The "slower forward line" explanation fits four of five losers, but we chose the forward lines after seeing the reactions. That is the textbook way to find a pattern that is not there. Several winners had softer forward lines too: AppFolio trimmed its margin ceiling and Tyler cut its range, yet neither fell hard.

Alkami is unexplained. Its release raised the year, and none of the coverage we opened attributed the 18.29% fall to a specific line. Consensus estimates in this post come from Investing.com, LSEG via CNBC and Bloomberg via Yahoo Finance, not from our own survey. Price moves include whatever the market did that day, and the software ETF fell 3.52% on 29 October alone.

The competing view deserves its strongest form. It holds that the forward lines are noise, and that a software stock's reaction is set by positioning. Fortinet's 120% run-up and ServiceNow's year of AI anxiety would explain their falls on their own. We cannot rule that out with this data, and we think the honest position is that positioning and forward lines both mattered.

Reading the next software earnings release

The November reporters will publish the same three kinds of number. A full-year range, a next-quarter range, and a backlog or growth metric. This season says to read them in reverse order.

Start with the forward measure the company itself emphasises: cRPO, net new ARR, organic growth or billings. Compare it with the quarter just reported. If it is slower, assume a full-year raise will not carry the stock, whatever the size of the raise.

Then check whether one line moved against the rest, as Fortinet's service revenue did. A single cut buried among raises was the most expensive line in this whole cycle. For the longer arc of how software valuations got here, the explainer on the SaaSpocalypse selloff covers the first-half drawdown this season extended.

Frequently asked questions

Where to start

If you own, sell to or compete with any of the November reporters, build a one-line tracker before their results land. For each company, write down the forward measure it emphasises and the last reported rate, then fill in the guided rate on the night. That single column would have flagged four of this season's five worst reactions before the bell.

If you buy software from these vendors, read the slowdowns as negotiating context rather than distress. A vendor whose backlog growth is slowing has a quarter-end it cares about, and your renewal conversation should start before it.

Related reading