Should You Buy ServiceNow Now? A Beat That Went Nowhere

📅 Written: 2026-07-23 (during US market hours) · 🏢 Company: ServiceNow (Ticker: NOW) · 💰 Price: $95.97 (~₩141,000) · 📈 7/22 Q2 earnings beat · ⚡ Gapped up to $99.76, gave it all back to $93.88 in 16 minutes

🎯 The One-Liner

"Last time they got grounded for saying 'my next test might come in a bit lower.' This time they actually did better on that very test. And the allowance stayed exactly the same."

Three months ago (April 22), ServiceNow aced every subject, and still the stock dropped -17.75% in a single day, all because of one line: "next test might come in a bit lower."

That "next test" was this July 22 report. The result?

It didn't drop. It went up. They scored well above the number they had warned about.

And the stock? After hours it popped +5%, and at the next open it jumped to $99.76, then gave back the entire gain within 16 minutes, falling to $93.88. It now sits at $95.97, essentially flat.

Let's unpack why this happened, and whether it's an opportunity or a trap.

💰 Cheap or Expensive Isn't About the Share Price

Whether a stock is cheap or expensive isn't about "the price of one share." You judge it by dividing the company's whole price (market cap) by its revenue.

📉 In 2025 this number was 22x; today it is about a quarter of that. Revenue actually grew 24%, yet the "price tag on the company" got cut to one-fourth.

That's the crux. The company got bigger, but the price tag went the other way. The stock halving from $201 to $95 isn't "because the company broke," it's "because people slashed the value they assign to it."

🏢 What Does ServiceNow Actually Do?

You know this situation at work.

"My laptop won't turn on" → who do I tell? → where's the request form? → when will it get fixed?

ServiceNow sells the system that runs this entire process automatically. It started as an IT-department tool, but now it handles the whole company's workflow, including HR, customer service, and finance.

Lately it has been bolting on AI workers (agents), finishing in 20 minutes what used to take a person two days.

🌍 Market Power Check

How it makes money: Software subscriptions are 97% of revenue. It comes in month after month, year after year, so it's very stable.

🆚 Rivals

On growth alone, ServiceNow is 2.5x Salesforce. Yet the premium the market used to pay for it is gone.

📊 Did Our Last Prediction Hold Up? 🟡 Half Right

Three months ago we saved a prediction file in advance. Let's grade it.

What we got right: The revenue call was 0.2% off. That isn't a lucky guess, it's a hit.

First miss: We had EPS at $1.10 while the actual was $0.90. The cause was that the "consensus estimate" we referenced was a wrong number. The real consensus was $0.86, so $0.90 actually beat it. In other words, the direction (a beat) was right and the scale was off. The real mistake was not accounting for the costs from three recent acquisitions.

Second miss: We got the up-move right, but we didn't see that it wouldn't hold the gain. The hint was that the stock had already fallen -6.5% the day before the print, and we missed it.

Next time we fix these two. (The prediction file keeps getting saved, and we grade it like this every time.)

📈 This Report Card, What Went Well

🌟 The One Line That Matters Most Here

"cRPO 21.5%", remember just this one number.

The only reason the stock crashed -17.75% three months ago was the company's warning that "next-quarter cRPO looks set to slip to 19.5%."

And the actual print came in at 21.5%. It didn't slip, it rose above the prior quarter's 21%.

It hasn't dropped once in four quarters! In other words, the "growth is rolling over" fear has never once been true in the actual numbers. The only thing that rolled over was the company's own "guide number."

📉 So Why Is the Stock Flat?

If the results are this good, why did it rise and then come right back down? Three reasons.

1️⃣ They Raised the Allowance by Only $15 (the biggest reason)

This is the crux. Let me explain with an analogy.

The kid scored 47 points higher than expected on this test. But the parents said, "Then let's raise this year's target score by 15 points." Wait, they scored 47 higher and the target goes up only 15? That means the parents expect 32 fewer points in the second half.

By the numbers:

The market read this as "the company isn't confident about the second half," and pulled the stock back down after the pop.

2️⃣ They Lowballed Next Quarter's Guide Again

Here's how the company guided Q3:

Once again they said "it looks set to come down." That is the exact same phrasing that erased -17.75% in April.

3️⃣ Part of the Score Was 'Borrowed'

Asked about it, the CFO answered:

"We had a bit more on-prem software this quarter, but that's purely a timing issue."

On-prem software is recognized all at once as revenue. So revenue that was going to land in Q3 got pulled forward into Q2, which means Q3 will be that much emptier. It's like spending next month's allowance early.

🧠 Big Picture: Opportunity or Trap?

This is the most important part of the piece. Read it slowly.

🔍 Finding 1: The company always guides low and always beats

This is the genuinely interesting find. Look at the recent record.

Two quarters in a row of doing better than they said. And by more each time.

Yet the market keeps taking the guide number at face value, getting scared, and then going "oh, that's good" when the real number lands. The Q3 guide is 20% again, and if the pattern holds, the actual is likely 21~22%.

💡 Analogy: There's a kid who keeps saying "I'll probably get around 80" and comes home with a 95. The parents keep freaking out at "80?!", then look at the report card and say "oh, nice work." By the third time, even the parents should learn.

🔍 Finding 2: There's genuinely good news nobody noticed

The most important number that didn't make the news is the company's cash-generation (FCF) target.

Remember the second reason the stock crashed in April? It was "buying a company called Armis worsens cash generation by 2pp."

And this time they raised it by 5pp. That's better than the thing that was supposed to get worse, in a single quarter. Yet no article that day put this number in a headline. This is what the market is missing.

🔍 Finding 3: The 'AI kills SaaS' fear is cracking

The reason software stocks all crashed together in 2026 was this scary story:

"Once AI robots replace employees, software companies that charged by headcount go under."

This quarter ServiceNow rebutted that with numbers for the first time.

The CEO also said this:

"2.2 billion AI agents are coming into enterprises worldwide. That means 2.2 billion new identities (accounts)."

The point being, each AI robot is also something to be managed, and managing it is our business. The more AI grows, the more there is to manage. It flips the threat into a revenue line.

🔍 Finding 4: But there's a real weakness (no hiding it)

Accounting profit is shrinking.

Revenue grew 24%, yet real accounting profit fell 21%. The two numbers differ by 3.1x.

That's because the company recently bought three others (Moveworks, Veza, Armis) at very high prices. Those costs are eating into profit right now.

💡 Analogy: Your salary went up, but you bought three houses on loans, so your bank balance actually shrank. If home prices rise, it was a smart call; if not, it was overreach. The verdict comes around 2027.

🔍 Finding 5: Breaking down the 1-year halving

The stock fell -50% over one year. Break down the cause and it looks like this:

The part driven by results is essentially 0. That's the conclusion of this analysis.

🔍 Fundamentals Traffic Light

Verdict: 🟢 7 / 🟡 3 / 🔴 2. Clearly more green, but the two reds ("GAAP profit" and "price trend") aren't light ones.

🎯 Action Guide

Verdict: 🟢 Buy, but not all at once

Last time (May) it was 🟡 Wait. Back then we set this condition:

"We need to confirm at the next earnings that cRPO clears the 19.5% guide."

That condition passed at 21.5%. So we upgrade the rating.

How to Scale In

Why scale in: the business is healthy, but the price trend itself is still a downtrend. It sits 25% below the 200-day line, and yesterday it ran to $100 and got pushed straight back. Even a good company can fall further before the trend turns.

💵 If You Put In ₩1M

At ₩141,402 per share, ₩1M buys 7 shares (₩989,814).

Probability-weighted expected return: +23.4%

👀 What to Watch Next, 5-Point Checklist

#1 matters most. They beat by 2pp this time; if they beat again in October, "this company just guides low" gets confirmed, and the market stops falling for it. That's when the stock re-rates for real.

🤔 FAQ

Q1. If results are good but the stock doesn't rise, isn't it just a bad stock?

Short term it can look that way, but a stock price is the product of two things.

Price = money the company earns × the multiple people assign to it

For ServiceNow the first number (earnings) grew 24%, while the second number (multiple) got cut 72%, from 22x to 6.1x. That's why the stock halved. As long as the first number keeps growing, the stock can rise sharply when the second returns to normal. The problem is nobody knows when, which is exactly why you scale in.

Q2. If I had bought during the -17.75% crash three months ago, where would I be now?

The day after the crash it closed at $84.78. At $95.97 now, that's +13.2%. 13% in three months isn't bad. But you would have had to sit through a further drop to the April 10 low of $81.24. The lesson: even a good decision comes with a stretch of being underwater.

Q3. Won't AI eventually wipe out software companies like this?

It's the most important question, and nobody can answer it for sure yet. But the rebuttal this quarter is fairly strong.

• 50% of new revenue is already unrelated to headcount
• Seats in use aren't shrinking, they're rising
• Customers actually using AI are up 9x in nine months

And the company's logic is "as AI grows, you also need a system to manage the AI." All 2.2 billion AI agents are things to manage. Whether that logic holds gets decided around 2027.

Q4. Buy now, or wait for the October print?

The answer is both. That's why we split it 40% / 30% / 30%.

Buy it all now and you have no dry powder if it falls further. Wait until October and you miss it if it rises before then. Scale in and you're at least half right either way. That's the standard playbook for a name where conviction is fuzzy.

Q5. Is GAAP profit at -21% really that big a problem?

Not right now, but it becomes one if it drags on. Plenty of institutional investors have a rule of "only buy companies with GAAP profit." If profit stays negative, that money stays out.

The company promised to "bring the headcount added by acquisitions back down in 2027." Keep it and this resolves; miss it and it turns into a trust problem. Note it as a Q1 2027 checkpoint.

🚨 Risk Warnings

1. The price trend is still a downtrend It sits 25% below the 200-day line ($127.57). Yesterday too it ran to $100.08 and got pushed back to $93.88. Even a good company can fall further when the trend is bad.

2. Cutting the second-half target is a real fact We argued "$32M is 0.2% of the total, so it's small," but you can't deny the company is looking at the second half cautiously. If Q3 actually disappoints, this call was wrong.

3. This quarter's score contains a 'borrowed' piece On-prem software revenue was pulled forward from Q3 into Q2. The company didn't disclose the size. Q3 could be that much emptier.

4. Microsoft keeps pressing Its strategy is to bundle Office, Teams, and Copilot and sell them cheap. No matter how good an individual product is, it can lose to the bundle.

5. Watch the CEO's rhetoric CEO McDermott tends to talk big. In April he said "Armis will be our Instagram," and the stock fell -17.75% that day. This time he called it "the foundation for a re-rating." Watch the numbers, not the words.

6. This piece was written during US market hours $95.97 is the price at 9:46 a.m. local on July 23, so it's not the day's close. It can change after the bell.

📚 Hard-Word Dictionary

📌 Three-Line Summary

1. The very fear (growth slowdown) that erased -17.75% in April got refuted by the numbers this time. Guide 19.5% → actual 21.5%.
2. But the company raised the full-year target by only $15, so the market read it as "not confident about the second half" and undid the pop.
3. The company keeps guiding low and beating, so one more beat in October could flip the trend. That's why we recommend scaling in: 40% now, 30% on a dip, 30% after the October print.

For the deeper data and how the 12M target is derived, see the April crash post-mortem in the analyst-version crash diagnosis; for the plain-English version of that same event, see the ServiceNow crash beginner guide. The underlying fundamental health check is in ServiceNow fundamental analysis. Comparing it with Datadog +31%, which surged in the same season, shows where the AI-era winner and loser line gets drawn.

⚠️ This is not investment advice. Predictions can be wrong, and we openly flagged two of our own that missed above. All investment decisions and their consequences are your own.