a16z called Workday “arguably the most important and least loved product in enterprise software.” I’m not sure I’d put it quite there, but the data they used to make the case includes a striking chart: Workday at 97% Gross Revenue Retention, ServiceNow at 98%, both sitting at the top of the public B2B stack rank.
So how can a product “nobody loves retain” 97% of its revenue every year?
It’s solving big enterprise problems. It’s structured data lock-in. And, importantly, it’s contract structure. And it makes GRR in many cases a lot more impressive, and predictive, than it sounds. Especially today, when so much budget is going into truly agentic products. Long term contracts are just masking that because there’s no exit. For now.
GRR is Not The Lighthouse Metric It Used To Be, At Least in The Enterprise
Both Workday and ServiceNow default to 3-year initial contracts.
Workday actively pushes 4, 5, even 6-year terms on renewal. ServiceNow charges roughly a 10% premium if you want a 1-year term instead of 36 months. Their pricing models are built to lock customers in for as long as possible.
In any given year, only about 33% of a 3-year contract base is up for renewal. Even if half of those churn at the renewal table, reported GRR only takes a ~17% hit. You can post 90%+ GRR while losing half your renewals every year.
For Workday with a heavier 5-year mix on renewal, only 20% of the base is up at any given year. Lose 30% of those and GRR still prints 94%.
97% GRR on multi-year contracts is not the same metric as 97% GRR on annual contracts. They are not comparable. One measures how much customers love you. The other measures how well your legal team wrote the MSA.
This Is a Category-Wide Pattern
Before getting to why ServiceNow and Workday are specifically at risk, the broader benchmark matters. Every major enterprise SaaS leader defaults to 3+ year contracts:
- Salesforce: Headline default is 1-year, but every meaningful enterprise deal is pushed to 3 years, and the largest customers (SELAs) are on 3-5 year lockups
- Oracle Fusion: 3-year minimum on Fusion SaaS, 4-year minimum on Cloud at Customer, with Oracle Master Agreements running 5 years as standard
- SAP RISE: 3-5 year subscription terms with 50-75% early termination fees on remaining contract value
- ServiceNow: 3-year default with a 10% uplift on shorter terms
- Workday: 3-year initial, 4-6 year renewals proposed by default
Multi-year contract duration is the universal posture at the top of public B2B. Which means every retention number sitting above 92% in the enterprise SaaS cohort is structurally inflated by contract duration, not just NOW and WDAY.
So why are ServiceNow and Workday more at risk than the rest?
Three Reasons NOW and WDAY Are Specifically Exposed
1. They sit at the absolute top of the GRR chart, with no reset yet.
Salesforce, Oracle, and SAP have all had their GRR resets over the last decade. Best-of-breed tools peeled off slices of their suite. CRM lost ground to HubSpot in the mid-market. Oracle and SAP have been losing finance and HR workloads to Workday, NetSuite alternatives, and a long tail of vertical players. Their retention numbers reflect a market that has already adjusted.
ServiceNow and Workday haven’t had that reset. They’ve had near-monopoly positions in their categories for a decade. 98% and 97% GRR is what you print when the market hasn’t yet figured out how to replace you. The fall is bigger when you’re starting from the top.
2. The AI displacement vector hits them harder than the others.
ServiceNow’s ITSM and workflow surface area is exactly what AI agents are best at automating. Ticket triage, request fulfillment, knowledge management, runbook execution. The whole product category is structurally exposed to “what if an AI agent just did this.”
Workday’s situation is different but worse in another way. The a16z framing of “least loved” is the leading indicator. There’s pent-up customer willingness to switch the moment something credible appears. The HR and finance use cases are sticky for record-keeping but not for workflow. AI-native HCM and finance tools don’t need to replace Workday completely. They just need to take the parts users actually touch.
Salesforce has Agentforce as defense and is moving aggressively into the AI agent layer. SAP and Oracle have multi-decade ERP lock-in that is hard to displace because the cost to migrate is enormous regardless of what the AI alternative looks like. NOW and WDAY are in the worst middle position: modern enough to lack deep entrenched lock-in, horizontal enough to be exposed across multiple AI threats.
3. They’ve trained the market that 97% GRR is normal.
When NOW and WDAY GRR compresses from 97-98% to 92-93%, that is a public market re-rating event. Their multiples are built on the assumption of permanent near-perfect retention. Salesforce trading at lower NRR doesn’t move the stock the same way Workday losing 4 points of GRR would. The expectation has been priced in for the legacy enterprise leaders. It hasn’t been priced in for NOW and WDAY.
The Oracle and SAP Precedent
Oracle and SAP have spent two decades building 5-year master agreements with 75% early termination fees and aggressive auto-renewal mechanics. They thought they had built unbreakable contracts.
What actually happened: a third-party support market emerged. Rimini Street, Spinnaker, and others built billion-dollar businesses helping customers exit Oracle and SAP commitments earlier than the contracts allowed. Customers who couldn’t terminate found ways to stop paying for what they didn’t need. License audits became weapons that customers learned to defend against.
The contracts didn’t actually hold. They just delayed the exit by a few years and turned the customer relationship adversarial.
That’s the trajectory NOW and WDAY are on if customer dissatisfaction outpaces contract duration. The current GRR numbers don’t tell you whether that’s already happening. They can’t. The contracts are designed specifically to prevent that signal from reaching the income statement until it’s too late to do anything about it.
What to Watch Instead
If you’re an investor, board member, or operator trying to read NOW and WDAY’s actual health, GRR is the wrong place to look. Four better indicators, ordered from earliest signal to latest:
1. Net new customer growth. This is the earliest signal in the chain, and the cleanest. New customers have choice. They aren’t locked into anything. If the category is shifting to AI-native alternatives, the first place it shows up is in new logo acquisition, not the existing base. ServiceNow discloses customer counts by $1M+, $5M+, $10M+, and $20M+ ACV bands. Workday discloses total customer count. Both have been decelerating on net new adds. When new logo growth flattens at a category leader, the category has started moving without them. By the time it shows up in GRR, the company is already 24 months into the slide.
2. Contract duration mix. ServiceNow has already disclosed in its 10-K that it’s seeing more 12-month deals, particularly out of US Federal. That’s the second leading indicator. When the average new contract drops from 36 months to 24 to 18, GRR compression is locked in 18-24 months later. The contract length is what customers are willing to commit to today, which tells you what they think of the product two years from now.
3. NRR before GRR. Net Revenue Retention compresses first, because expansion slows before churn shows up. Customers stop adding seats and modules when they’re losing confidence, well before they actually leave. Both companies’ NRR has already been trending down. That’s the canary.
4. Renewal-cycle deltas, not full-base GRR. What percentage of customers up for renewal in a given quarter actually renewed at full contract value? That’s the number that tells you whether the product is winning at the renewal table right now. Almost no public B2B company discloses this. They report blended GRR because the blend looks better than the underlying truth.
Why the AI Era Breaks This Specifically
The 3-year contract was built for a different market. It assumed product replacement cycles measured in years and a procurement function that wasn’t actively trying to preserve optionality. Both of those assumptions just broke.
AI-native B2B is doing 1-year and monthly deals as a structural feature, not a concession. When the new generation of B2B prices in 12-month or monthly increments, that’s not generosity. It’s a recognition that the product is improving every quarter, the customer’s use cases are evolving every quarter, and the competitive set is reshuffling every quarter. Their 95% GRR on 1-year terms is a much harder number than legacy 97% on 3-year lockups. It means customers are choosing to stay every year, not contractually obligated to.
AI compresses replacement cycles by 3-4x. A product category that took 5 years to disrupt in 2015 takes 18 months in 2026. ITSM workflows, expense reporting, recruiting automation, contract management, sales enablement. The list of categories where an AI-native tool has reached production credibility in under 2 years keeps getting longer. Contract duration was a 5-year defense in the SaaS era. In the AI era, it’s a 5-year liability. By the time the renewal comes up, the market has moved twice.
Procurement teams have figured out the math. This is the part that should worry NOW and WDAY boards the most. CIOs and procurement leaders are now actively shortening contract terms specifically to preserve the optionality to switch to AI-native alternatives that ship in 2027 and 2028. ServiceNow’s own 10-K disclosure of more 12-month Federal deals is the public version. The private version is happening at every Fortune 500 procurement table where someone has said “we’re not signing a 3-year deal in this market.”
When the customer side starts demanding 1-year terms and the vendor side has to choose between losing the deal and breaking the duration model, the duration model loses. That has already started.
High GRR is The Enterprise Doesn’t Mean You Have the Right Agentic Product
If you’re building B2B today, the takeaway is simple:
95% GRR on 1-year contracts beats 97% GRR on 3-year contracts. Every time.
Don’t hide behind multi-year terms to make your retention look better than it is. Your board will see through it. Your acquirers definitely will. And in 18 months, your customers will too.
The best products of the AI era will be the ones that win every renewal because the customer chose to renew, not because the MSA forced them to.
ServiceNow and Workday are great companies that built great products. They’re going to be very large for a very long time. But “97% GRR” and “least loved product in enterprise software” being simultaneously true tells you everything about how to read these retention metrics in 2026.
The contract is hiding the truth. The AI era is going to expose it. And the companies sitting at the top of the GRR chart are the ones with the most to lose when it does.
