The Expansion Engine — Why Retention Slowed, and the Bet to Restart It
The Expansion Engine — Why Retention Slowed, and the Bet to Restart It
The deceleration that halved monday.com's stock is the number the whole case hinges on, and chapter 1 stated it without explaining it. This chapter takes the engine apart. The finding that matters: the slowdown is not churn and it is not a broken model. It is a maturing seat-expansion machine whose decline is concentrated in the long tail of small accounts, while the enterprise core — customers paying more than $50,000 a year — has expanded at a near-constant 115%–116% for three straight years [1]. The seats-plus-credits pricing pivot, launched in 2026, is management's wager that AI consumption can do what added human seats no longer can. Whether it works is the swing factor behind everything that follows.
How the engine works: land cheap, then expand four ways
monday.com lands small — a single team on a per-seat plan — and grows the account along four axes: more users (seats), more products (work management, CRM, service, dev), more departments, and migration upmarket into larger contracts. The gauge for all of it is net dollar retention: take the recurring revenue from a cohort of customers, look at the same cohort twelve months later — after every upsell, contraction, and cancellation — and divide. Above 100% means the average customer spends more each year without the company adding a single new logo [2].
When that number ran hot, it was the entire growth story. For customers with more than ten users — the company's core target — net retention was over 135% at the end of 2021 and over 130% at the end of 2022; across all customers it was over 120% [3] [4]. A customer base that reliably spends a third more every year, compounding, is a growth annuity. That is the machine the market paid a hyper-growth multiple for.
Two retention numbers, two different stories
The single most important distinction in this chapter is that monday.com now reports two retention rates, and they say opposite things. The overall rate — every customer, including the vast self-serve tail — has fallen hard and flattened in the low 110s. The enterprise rate, for accounts above $50,000 in ARR, has barely moved.
Sources: FY2021 and FY2020 figures [5]; FY2022 [6]; FY2023–FY2025 overall rates [7].
The overall line tells the scary version: from 121% at the end of 2022 to 110% at the end of 2025 — eleven points of expansion gone in three years [8] [9]. But look at the same window through the cohort the company actually sells to, and the panic recedes.
Source: FY2025 Annual Report (Form 20-F), Net Dollar Retention Rate disclosure [10].
The enterprise base expanded 115%, 115%, then 116% across 2023, 2024 and 2025 — essentially flat, and flat at a healthy level [11]. The gap between the two lines is the long tail of small, self-serve accounts that adopt cheaply, expand little, and churn more. As the company scales, that tail weighs more heavily on the blended average even as the part of the business that pays the bills holds firm.
Why it slowed: weaker expansion, not a leaky bucket
monday.com has been unusually candid about the cause, and the cause is not customers leaving. As early as the 2022 annual report, management attributed the dip to "a slowdown in expansion of existing customers," adding that it saw "very healthy traffic of new customers" while "macro-economic factors are leading to slower expansion in some existing customers" [12]. In plain terms: the bucket is not leaking; customers are filling it more slowly. In a per-seat model, expansion is largely seat growth, and seat growth tracks customers' own hiring. When the broad software market stopped adding headcount in 2023–2024, monday.com's most powerful lever — selling more seats into accounts that keep growing — lost its force.
The seat data shows exactly this fatigue. The number of paid customers with more than ten users — the seat-expansion base — grew just 7% in the year to March 2026, to 65,016 [13]. The high-value cohort, by contrast, compounded several times faster over the same year.
Source: Q1 FY2026 results, Recent Business Highlights [14].
Two further forces are arithmetic rather than fundamental. First, monday.com changed its headline retention metric: through 2022 it led with the "more than ten users" figure (130%–135%); from 2023 it leads with the lower "overall" figure, so part of the apparent collapse is a definitional step-down, not a business one [15] [16]. Second, the law of large numbers: holding 116% expansion on a $1.2 billion base is a far heavier lift than holding 135% on a $300 million base. Neither force implies the model is impaired. Both imply it is maturing.
The mix is getting better even as growth gets slower
Here is the part of the engine the bear case tends to skip. While the headline rate fell, the quality of the revenue improved, because the upmarket migration is real and accelerating. The enterprise tiers are growing their account counts far faster than the base, and they now command a rising share of total recurring revenue.
Sources: Dec 2024 and Dec 2025 counts, Q4 FY2025 results [17]; Mar 2026 counts, Q1 FY2026 results [18].
The $500,000-plus tier nearly doubled in a year, from 50 accounts to 99 — Q1 2026 set a record for net additions in that bracket [19] [20]. As that happens, the revenue base concentrates in larger, stickier, multi-product contracts — the opposite of a business hollowing out.
Sources: Sep 2024 and Sep 2025 shares, Q3 FY2025 results [21]; Dec 2025 share, Q4 FY2025 results [22]; Mar 2025 and Mar 2026 shares, Q1 FY2026 results [23].
Customers above $50,000 in ARR now represent 42% of all recurring revenue, up from 34% eighteen months earlier; the above-$100,000 tier reached 29% [24] [25]. In its 2025 results, management named the growth drivers explicitly: "increased seat penetration, multi-product adoption, and broader adoption of AI capabilities" [26]. The first lever is tiring. The company is betting the second and third can carry the load.
The seat ceiling — and the consumption-pricing bet to break it
This is where the chapter meets the through-line. A per-seat model has a structural vulnerability that AI sharpens: if revenue scales with the number of human users, then software that lets a team do more with fewer people threatens the very mechanism of expansion. A skeptic's one-line bear case is that AI caps the seat count — and monday.com's overall NDR drifting toward 110% is the first tremor.
Management's answer, delivered alongside the "AI Work Platform" relaunch in early 2026, is to change how customers pay. New customers now buy on a seats-plus-credits model: a seat base plus consumption-based credits that meter actual AI usage. Existing customers can opt in, and enterprise accounts receive complimentary AI packages to seed adoption [27]. The full economics of that pivot — whether credits can monetize at scale without cracking monday.com's prized 89% gross margin — belong to chapter 6, the report's home for the monetization case.
The bet is unproven, and the same facts read two ways. The bull sees a maturing enterprise engine — stable 116% cohort retention, a doubling $500K tier, a rising ARR mix — bolted to a new consumption lever that turns the AI threat into a tailwind. The bear sees consumption pricing as exactly what a company adopts once its seat model has hit a ceiling. What is no longer in doubt is the diagnosis: the deceleration is a seat-expansion problem, located in the long tail, arriving as the market stopped adding seats. Whether the credit meter restarts the engine — and whether competitors let monday.com charge for AI at all — is the question the next chapters must answer.