NRR by ACV Tier: How Six Enterprise Accounts Hide a Failing Mid-Market

One number, three completely different businesses

A $40M ARR company put a 110% blended NRR on the cover of its information memorandum and called it the proof point for a repeatable model. When we pulled the renewal ledger apart by contract size, that 110% was almost entirely the work of six enterprise accounts expanding around 150%. Underneath them, the $20k-ACV mid-market book the company named as its primary growth engine was retaining at 88% and shrinking every quarter. Same metric, two opposite stories, and the buyer's Quality of Earnings team found the second one in an afternoon.

This is the specific failure of a blended NRR: it is a weighted average that lets your largest accounts vote on behalf of your smallest. Net Revenue Retention measures what a cohort of existing customers is worth this year versus last, after upgrades, downgrades, and churn. Roll every ACV band into one figure and a handful of seven-figure expansions can drag the line above 100% while the bulk of your logos bleed out. The average looks like health. The distribution is a warning.

The reason this matters at exit is mechanical, not cosmetic. An acquirer is not buying your aggregate; they are buying the durability of each segment's cash flow, and they discount the segments that don't hold. If your mid-market engine is the thing the deck promises will scale, and that engine is at 88%, the model isn't repeatable — it's enterprise concentration wearing a growth-story costume. We unpack why a sub-100% reading is a structural signal, not a soft one, in our diagnostic on why sub-100% NRR points to a broken customer success function.

What "good" actually is at each contract size

The mistake is measuring a $10,000 account and a $250,000 account against the same target. They retain through completely different mechanics, so the benchmarks diverge sharply by tier — and the 2026 data makes the spread hard to ignore.

SMB, ACV under $25k. Median NRR sits at roughly 97%, per Optifai's read across more than 900 B2B SaaS companies [1]. Below 100% means the existing book contracts on its own every month, so you are buying new logos just to stand still. The reason it's so hard here isn't your team — it's the segment. Small customers are price-sensitive, they fail at higher rates, and the contract is small enough that a single budget cut ends the relationship. Expansion at this tier has to be product-led and nearly free to deliver; a human save motion costs more than the account is worth. If you're clearing 105% in SMB, the expansion is built into the product, not into a renewal call.

Mid-market, ACV $25k to $100k. This is the band that tells you whether you have a real business. Median NRR climbs to about 108%, though SaaS Capital's retention work puts the lower $25k–$50k slice closer to a 102% median [2] — useful context, because "mid-market" is wide and the bottom of it behaves more like SMB than people admit. You can't lean on credit-card upsells here, and you can't afford enterprise-grade white glove. Retention comes from structured value realization: did the customer actually reach the outcome they bought? A mid-market book stuck at 95% is usually a product-fit problem dressed up as a customer-success problem, and it's the single most expensive thing to discover during diligence rather than before it.

Enterprise, ACV above $100k. Median lands near 118%, with top-quartile performers past 125%. These contracts are sticky by construction — embedded in core workflows, multi-year, built to expand by seats and departments. Fullview's tier-by-tier benchmarks track the same widening gradient from SMB up through enterprise [3]. The trap is using this number as your headline. A genuine 125% enterprise figure papering over an 85% SMB book doesn't reassure a buyer — it tells them your durable revenue is concentrated in a few accounts, and concentration is exactly what gets discounted. For how these figures move across segments, see our guide to NRR benchmarks.

Three things to do before the next board meeting

First, retire the single NRR line and report it by ACV tier. Put SMB, mid-market, and enterprise in three columns, and show Gross Revenue Retention next to Net for each. The GRR-vs-NRR gap is the tell most decks bury: if a tier's NRR looks fine but its GRR is well under it, expansion is masking churn, and that gap closes the moment expansion slows. Splitting the columns forces the room to own which segment is actually carrying the number.

Second, stop running one customer-success playbook across all three. A motion built for $80k accounts will quietly bankrupt you on $12k ones, and the reverse leaves your best expansion candidates under-served. SMB retention has to be triggered off product usage and automated — alerts and in-app nudges, not calls. Mid-market needs prescriptive onboarding aimed at a measurable outcome inside the first 90 days, because that early proof is what carries the renewal. Enterprise needs named account management working a whitespace map and multi-threading past the original champion. Three segments, three cost structures, three definitions of a good month.

Third, follow the renewal money, not the health scores. The pattern we see most often is senior CS time pouring into $15k accounts that were always going to leave, while $80k accounts sitting on obvious expansion go untended. Audit where your most expensive people actually spend their hours against where the dollars renew and grow. If the answer doesn't match, you have found real margin without touching headcount.

The acquirer is going to segment your retention data regardless. The only question is whether you've already done it — while you still have the time and the leverage to fix what it shows — or whether you find out the mid-market was at 88% the same week they do.