Why your SaaS expansion revenue stalled while new logos climb
New logos land and gross churn looks normal, but your net revenue retention has slid for four quarters. That one number is four different machines, and the two fixes founders reach for first make it worse.
By Stacey Tallitsch | August 3, 2026
Your board deck still looks healthy. New logos are landing on schedule, gross churn is inside the range it has always been, and the top-line number is up and to the right. But one line has been sliding for four straight quarters, and you have finally started watching it: net revenue retention. Eighteen months ago it printed 116%. Last quarter it came in at 99%. Nobody on your team is alarmed, because every number they personally own still looks fine. Your Customer Success (CS) lead points at a steady renewal rate. Sales points at a full pipeline of new deals. And you are left holding a single figure that says your existing customers, taken as a whole, have stopped growing. The instinct in the room is to hand CS an upsell quota, or to ship a new pricing tier. Hold on.
That one number is lying to you about what is actually wrong. Not because it is inaccurate, but because it is four different machines wearing one gauge.
The metric no one in your company owns
Net Revenue Retention (NRR) is the percentage your revenue would grow or shrink from your existing customer base alone, before you add a single new logo. It is three things stacked into one figure: the revenue you keep, plus the revenue you expand, minus the revenue that contracts. Above 100% means your installed base grows on its own. Below 100% means it is shrinking, and every new deal your salespeople close is spent replacing revenue you already had instead of adding to it.
Here is the structural problem, and it is the reason your team is calm while you are not. NRR is a composite that no single person in your company owns. Retention lives with Customer Success. Pricing and contraction live with finance. Expansion lives with sales. Each of those leaders is optimizing their own slice, each slice looks acceptable in isolation, and the revenue is leaking out of the seams between them. You are the only person in the building positioned above all three. That is not an accident of your org chart. McKinsey's 2025 research on the metric found the same split across nearly a hundred B2B software companies: responsibility for the three buckets of NRR almost always falls under different leaders, which is precisely why so few companies move the number at all.
And the number matters more than your team may realize. In that same McKinsey analysis, top-quartile-valued SaaS companies posted NRR of 113% while their bottom-quartile peers sat at 98% — and the top quartile carried a median revenue multiple of 24x against 5x for the bottom. The gap between 113 and 98 is not a rounding error. It is the difference between a business that compounds and one that runs on a treadmill.
So before you touch anyone's compensation, you need to know which of the four machines under that gauge is the one that broke.
Machine one: your new cohorts are activating worse than your old ones
Expansion is downstream of activation, and activation happens months before the expansion ever shows up. A customer who onboards well, reaches the moment your product actually pays off, and builds it into their week is the customer who adds seats, adds usage, and upgrades a year later. A customer who signed, logged in twice, and never crossed that threshold will renew once out of inertia and then quietly leave.
The trap is timing. The NRR you are reading today reflects the cohorts you onboarded twelve to eighteen months ago. If your activation quality has been decaying — because onboarding got busier, because the product got more complex, because the people who used to hand-hold new accounts got reassigned — you will not feel it in the retention number for a year. Then it arrives all at once, disguised as an expansion problem. It is not an expansion problem. It is a time-to-value problem that aged into one. This is the same mechanism that shows up earlier in the funnel, and I have written before about why a SaaS trial-to-paid rate slides while signups hold steady — same disease, earlier stage.
Machine two: your customers already bought everything you sell
Some expansion stalls are not effort problems. They are ceiling problems. Your best customers have bought every seat they need, every module you offer, and the top tier of your plan. There is nothing left to sell them, and no amount of quota or QBR pressure conjures a purchase that your packaging does not contain.
This one hides because it looks like sales underperformance. It is a product and pricing design failure. Companies that build expansion into the packaging — predefined upgrade paths, add-ons, tiers tied to more than one variable so growth in any dimension triggers a move up — grow their installed base structurally, without asking a human to push. The McKinsey research put a number on it: companies with best-in-class pricing and packaging discipline ran roughly 16 percentage points higher NRR than peers who left it to improvisation. If your top accounts have topped out, no comp plan fixes that. Your roadmap does.
Machine three: a few whales are hiding a shrinking base
This is the most dangerous of the four, because the blended number looks calmest right before it collapses. Your NRR can sit at a respectable 104% while two enormous accounts double their spend and the entire long tail underneath them quietly downgrades, drops seats, and slips toward the exit. Two opposite motions cancel out in the average, and you read a plateau where you actually have a fault line.
The tell is the gap between net and gross. Gross Revenue Retention (GRR) strips out all expansion and shows you only what you kept. If NRR looks steady but GRR has been falling, expansion is doing all the work and the base beneath it is eroding. That is a business propped up by a handful of accounts you do not control, which is a fragility problem wearing a growth costume. I unpacked the non-software version of this in why revenue can drop while your customer count holds steady — the mechanic is identical, only the vocabulary changes.
Machine four: last year's logo targets poisoned this year's expansion pool
Look at who you sold to over the past year. If new-logo pressure pushed your team down-market — smaller accounts, lower-fit buyers, companies who signed because the deal was cheap rather than because your product was built for them — you bought your growth number with customers who will never expand. Low-fit customers use less, upgrade rarely, and churn on schedule. They pad the acquisition slide and then drain the retention one.
This is the cruelest of the four because the decision that caused it felt responsible at the time. You had a number to hit. You hit it. And the bill came due four quarters later as a stalled NRR that has nothing to do with your Customer Success team and everything to do with who you let into the base. Your Ideal Customer Profile (ICP) is not a marketing slide. It is a retention filter, and loosening it is a loan against your future expansion revenue.
The two fixes that make it worse
Now the turn, because the room's two instincts are both wrong, and they are wrong in a specific way.
The first instinct is to give Customer Success an upsell quota. Understand what that does. Your CS team is the one function your customers still trust to be on their side, and the moment you put a bag on their back, every renewal call becomes a sales call and the trust that drove your organic expansion evaporates. The McKinsey data is blunt here: for customer success and support, sticking to solid foundational practice is often enough, and piling on more pressure does not correlate with higher NRR. You would be spending your most valuable relationship asset to chase a number that lives somewhere else entirely.
The second instinct is to ship a new, higher pricing tier. If your problem is machine one or machine three — activation decay or a shrinking base — a new tier expands no one, because the customers you have are not using enough of what they already bought to want more of it. You will have added packaging complexity, confused your sales team, and moved the gauge zero percent. A new tier only works against machine two, and only after you have confirmed machine two is the one you actually have.
That is the whole point. Every one of these four machines needs a different repair, several of them are opposites, and the blended NRR number cannot tell you which is broken. Reaching for a fix before you have decomposed the metric is how founders spend a quarter making the wrong machine louder.
What to do before you touch comp or a roadmap
Do this today, before your next leadership meeting. Take your single NRR figure and break it into its three parts — gross retention, gross expansion, and contraction — and then cut each of those by acquisition cohort and by customer segment for the last six to eight quarters. You are looking for one thing: which line moved, and in which group. If contraction is climbing in your newest cohorts, you have an activation or ICP problem. If expansion has gone flat across your healthiest accounts, you have a packaging ceiling. If GRR is falling while NRR holds, you have whales hiding a leak. The decomposition takes an analyst an afternoon, and it converts one anxious number into a specific, fixable diagnosis. The recurring-revenue plateau I described for an MSP signing clients while its recurring revenue stays flat yields to exactly this move: stop staring at the composite, and go find the component.
The number is not your problem. The number is a symptom. Go find the machine.
— Stacey Tallitsch, Stronghold CMO
About the Author
Stacey Tallitsch is the President of Stronghold CMO, a Fractional AI CMO service operating under Talisman Capital, Inc. He is a 30-year tech veteran and the author of 21 books on systems thinking, operator-grade decision-making, and personal sovereignty, with more than 30,000 students across his Udemy course catalog.