Ask a room of CS leaders what counts as a good retention rate and you’ll get a number back with total confidence attached to it. Ask what ACV tier that number describes, and half the room goes quiet. That gap, a benchmark stated with certainty and no context for who it applies to, is where retention rate does its most damage. Not because the formula is wrong. Because the number travels alone when it should never leave the building without its segment attached.
Retention rate is the inverse of churn rate, and it inherits every one of churn’s blind spots while adding a few of its own. It gets reported as a single figure in board decks, compared against benchmarks that describe a different business entirely, and treated as proof of health when it’s really just one angle on a much bigger picture.
The Same Retention Rate Can Hide Very Different Results
The formula itself is not the problem. Customers retained divided by customers you started with, times 100, or the revenue version of the same math. Anyone can compute it in a spreadsheet in thirty seconds. The problem is what happens after the number exists: it gets treated as a verdict instead of a starting point.
A 90% retention rate sounds like a passing grade. But 90% retention compounds very differently depending on what’s inside it. Ten businesses can each report 90% and be having ten different years, because the formula collapses information the same way churn rate does. It tells you how many stayed. It says nothing about which ones, what they were worth, or whether the ten percent who left were your best-fit accounts or your worst.
Logo Retention and Revenue Retention Pull in Different Directions
This is the split that a single blended retention number hides, and it’s not a rounding error. Logo retention counts accounts. Revenue retention counts dollars. A business can hold onto 95% of its logos while revenue retention slips because the accounts it kept downgraded, and a business can post strong revenue retention while quietly bleeding smaller accounts that would have mattered in three years.
Dean Curtis, CEO of Ingage, made this exact tradeoff explicit on the Across the Funnel Podcast, talking about why his platform’s tiered pricing exists:
So that retention is so important. Logo retention, revenue retention. Well, if I can maintain the logo, but for a lower dollar, I’m happy to do that, if that’s the value I provide.
That’s a deliberate trade, not an accident, and it only makes sense once you stop treating logo and revenue retention as the same metric wearing two names. A downsell that keeps the logo and drops the dollar is a very different outcome than losing the account outright, and a retention report that blends the two into one number can’t tell your board which one actually happened. For the fuller breakdown of why these two numbers diverge and what each one is actually built to catch, gross retention and net retention deserve to be read side by side, not averaged into one figure.
A 92% Retention Rate Can Describe Two Completely Different Businesses
Enterprise SaaS with six-figure ACVs and multi-year contracts can report 92% retention and be underperforming, because at that tier, losing even one or two large accounts is a material event that a healthy-sounding percentage can mask. Self-serve SMB, sold with a credit card and no onboarding call, can report the same 92% and be running well ahead of what’s normal for the segment, where 75 to 85% annual retention is often the realistic ceiling given the price point and the lack of switching cost.
The number is identical. The story underneath it is not even in the same category. Reporting retention rate without naming the ACV tier it came from is like reporting a temperature without saying which scale you used, and yet it happens constantly, because the formula doesn’t force anyone to disclose the context that makes it interpretable.
Why Retention Benchmarks Need ACV Context
This is where most public retention benchmarks quietly lie. A blog post that tells you “good SaaS retention is 90%” is describing an average across companies that have almost nothing in common with each other operationally. Enterprise motions with named CSMs and quarterly business reviews retain differently than PLG motions where nobody has ever spoken to a human at the vendor. Blending them into one benchmark produces a number that’s true for almost no one reading it.
Paul Staelin, former Chief Customer Officer at Vercel, described what real enterprise retention work looks like on the Across the Funnel Podcast, talking about turning around gross dollar retention for Vercel’s largest accounts:
GDR went from low seventies up to high eighties. We actually put in some differentiation. Actually, it ended up getting up to 90 for the last three quarters.
Notice what’s missing from that story: a single industry-wide benchmark. What mattered was the segment, the starting point, and the specific motion that moved the number, not whether 90% cleared some universal bar. A 20-point swing in gross dollar retention at the enterprise tier is a multi-quarter turnaround project. The same 20-point swing at a low-ACV self-serve tier might happen from a single pricing change. Same math, entirely different level of difficulty, and a benchmark that doesn’t say which one it’s describing isn’t useful to either.
Net Revenue Retention Can Hide a Shrinking Core Business
Net revenue retention gets the most airtime of any retention metric because it’s the one that can go above 100%, and a number above 100% photographs well in a board deck. But NRR blends expansion into the same figure as retention, and a big expansion quarter from a handful of large accounts can cover for real erosion happening everywhere else in the base. A business can report 115% NRR and still be losing its smaller, newer cohort at a rate that will matter in eighteen months, because the expansion dollars from a few accounts are large enough to swamp the signal from everyone else.
This is the same distortion that shows up in blended churn numbers, just flipped. If you’re only watching the metric that can go up, you’re structurally unable to see the part of the business that’s going down. Gross revenue retention, the version that strips out expansion and only measures what you kept, is the number that can’t hide behind a good quarter from your biggest accounts, and it belongs next to NRR in every retention report rather than left out of it. For the mechanics of how a strong headline number can still be masking a weaker core, why a high net dollar retention rate can be misleading walks through the pattern in more depth, and understanding net revenue retention is worth the read if the distinction between gross and net still feels academic rather than operational.
Retention Rate Is Just Churn Rate Wearing a Different Sign
Retention rate and churn rate are mathematically inverse, subtract one from a hundred and you get the other, and yet teams routinely report one without the other as though they’re separate stories. They’re not. Every caveat that applies to churn rate applies to retention rate in reverse. Cohort age distorts both. A young, fast-growing base drags retention down the same way it inflates churn, because newer accounts haven’t been filtered for fit yet. A structural floor exists for both, the ceiling your retention rate can realistically hit depends on your category and your buyer, the same way a churn floor exists that no CS motion will get you below.
Reporting retention rate as if it’s an independent, standalone health signal, separate from the churn conversation, is how a business ends up with two departments telling two different stories about the same underlying reality. The SaaS churn rate benchmarks by ACV tier and the retention benchmarks by ACV tier should be the same document, viewed from opposite ends, not two separate slides prepared by two different teams.
What a Retention Number Needs Before It Goes to the Board
None of this means retention rate is a bad metric. It means it’s an incomplete one until it’s paired with the context that makes it interpretable. A retention number in a board deck needs its ACV tier stated, its logo and revenue splits shown separately rather than blended, and a note on whether the figure includes or excludes expansion. Skip any one of those and you’ve handed the board a number that invites the wrong comparison, usually against a public benchmark from a business with a completely different cost structure and buyer.
The teams that get the most credit for their retention motion aren’t the ones with the highest raw percentage. They’re the ones who can explain, in one sentence, why their number is wha
t it is and what would need to change to move it. Customer retention management done well starts with that explanation, not with the percentage itself, and a platform like Hyperengage is built around surfacing the segment-level and account-level signals that explain a retention number before someone has to defend it cold in a leadership review. Measuring loyalty by the formula alone only gets you halfway there, and what the retention rate formula actually reveals versus what it conceals is worth sitting with before the next board cycle.
Negative net churn, where expansion revenue more than offsets what you lose, is the ambition behind most retention conversations in SaaS right now, and it’s a real and achievable target. But negative net churn as a goal only makes sense once you’ve separated logo retention from revenue retention and stopped letting one metric’s good quarter hide the other’s bad one.
Conclusion
Retention rate doesn’t need a better formula. It needs the same discipline every other headline metric in post-sales requires: name the segment, split logo from revenue, and say whether expansion is baked in or stripped out. A retention number presented without that context will always sound more reassuring than the business actually is, and the gap between how a number sounds and what it means is exactly where boards get surprised months later by a problem the raw percentage never showed them. The formula was never the hard part. Knowing what to ask before you trust it is.
