Neither wins outright — and anyone telling you otherwise is selling something. Usage-based pricing boosts net revenue retention by squeezing more money from customers who stick around, but flat-rate pricing often keeps more logos on the books during the exact moments usage-based models fall apart.

That distinction matters more than most pricing content admits. Almost every "usage-based is the future" article quotes net dollar retention (NDR) numbers and calls it a day. NDR tells you how much revenue grew from your existing base. It says nothing about whether the same number of customers renewed. Those are different questions, and founders keep answering the wrong one.

The Metric Everyone Confuses: NDR vs. Logo Retention

Here's the sleight of hand baked into most "usage-based wins" content: it cites revenue retention, then lets you assume that means customer retention. It doesn't.

Net dollar retention measures money. Have 100 customers, watch 10 of them triple their usage while 20 churn completely, and your NDR can still look fantastic — even as your logo count drops by 20%. Usage-based pricing is structurally built to produce this exact pattern: your best customers expand hard, and your worst customers can shrink to near-zero spend before they ever formally cancel.

Flat-rate pricing doesn't have that flexibility. A customer paying $500/month either renews at $500/month or churns entirely. There's no slow bleed. Worse for revenue smoothness, sure, but flat-rate churn is more visible, more binary, and — critically — tied to a renewal event rather than a usage dip that happens silently between invoices.

The data that gets cited to support usage-based pricing is almost entirely revenue data. Companies using usage-based models average roughly 10 percentage points higher net dollar retention than subscription-only peers, according to the State of the Cloud 2021 report cited via BillingPlatform. Subskribe found that seven of the nine SaaS IPOs with the best net retention in recent years used usage-based pricing. Both stats are real, both are impressive — and neither tells you whether those companies kept more logos than a flat-rate competitor would have.

Where Usage-Based Pricing Genuinely Wins

Adoption is accelerating fast enough that ignoring the model entirely is its own risk. According to Metronome's State of Usage-Based Pricing 2025 report, cited via Advisable, 78% of companies currently using usage-based pricing adopted it within the last five years, and nearly half adopted it in just the last two years. OpenView Partners' SaaS Benchmarks tracking shows usage-based adoption climbing from 27% of SaaS companies in 2023 to 38% now.

The growth case is strong too. OpenView found that companies with primarily consumption-based pricing grew revenue roughly 8 percentage points faster than peers on other models. Three things drive that: low entry friction, natural expansion, and fair scaling across customer sizes.

Low entry friction matters because usage-based pricing behaves like a permanent free trial. A developer can start with an API integration costing $12/month and never have to go through procurement to scale it to $12,000/month. No renegotiation, no upsell call, no plan-migration friction — usage climbs, the invoice climbs with it. That's why usage-based pricing dominates transaction-heavy, API-first, developer tooling categories: Twilio, Stripe itself, and infrastructure players all lean into it because the value metric (calls, messages, compute) maps directly to what the customer is actually getting.

That expansion-without-friction dynamic also solves a structural problem flat-rate pricing can't: extracting proportional value across segments. A flat $99/month plan either underprices your enterprise power users or overprices your SMB trial accounts. Usage-based pricing self-adjusts, letting a five-person startup and a 5,000-person enterprise sit on the exact same pricing structure without a single contract renegotiation.

The Contrarian Case: Bill Shock and Budget Freezes

Here's what the usage-based cheerleader content skips entirely: what happens when usage drops.

Flat-rate revenue is sticky by design. A customer signs a 12-month contract at $2,000/month, and even if their usage of your product falls off a cliff in month four, you keep collecting until renewal. That's a full contract term of runway to win them back, prove more value, or at minimum bank the revenue. Usage-based pricing offers no such buffer. A customer facing a Q3 budget freeze, a seasonal dip, or a parent-company consolidation simply stops sending traffic — and your revenue from them collapses in the exact same billing cycle. No grace period. No lag. It's instantaneous, and it's brutal during downturns specifically, which is exactly when you need retention the most.

There's a second, quieter problem: bill shock. Unpredictable invoices are consistently cited as a top billing complaint in usage-based SaaS, and it's an underexplored retention killer because it doesn't show up as "churn" in your dashboard — it shows up as a support ticket, then a downgrade, then a slow fade. A customer who gets hit with an invoice 3x higher than expected doesn't think "great, I'm getting value." They think "I don't trust this vendor's billing," and they start rationing usage defensively, which caps the exact expansion revenue usage-based pricing is supposed to generate.

Does month-to-month usage billing remove the lock-in effect of annual contracts?

Largely, yes. Annual flat-rate contracts create structural retention through the renewal cycle itself — a customer has to actively decide to leave, and that decision point often comes with a win-back conversation, a discount offer, or a contract renegotiation that surfaces problems before they become churn. Pure usage-based billing, especially month-to-month, removes that checkpoint entirely. There's no renewal event forcing a conversation — the customer just quietly reduces usage until spend approaches zero, and you may not notice until it's too late to intervene.

This is also where finance teams start pushing back. Usage-based revenue looks great in an NRR slide but creates forecasting chaos — nobody can reliably predict next quarter's revenue from a base whose spend fluctuates with customer usage patterns you don't control. That forecasting uncertainty doesn't just annoy your CFO; it indirectly hurts retention, because unpredictable billing erodes the trust that keeps customers from shopping around when a competitor pitches a simpler, flatter price.

The Hybrid Verdict: Layer, Don't Choose

Most fast-growing SaaS companies aren't picking a side. They're running a subscription floor with usage on top — call it the Floor-Plus-Consumption model. You get a predictable MRR base that survives a bad quarter, plus uncapped upside when customers scale.

Zapier is the textbook example. Its pricing scales by tier and by usage volume — the number of "Zaps" run each month — so a customer pays a stable base fee for their plan tier, and consumption above that threshold triggers upgrades or overage. That structure gives Zapier the forecasting stability of flat-rate pricing and the expansion mechanics of usage-based pricing, without forcing a customer through a full contract renegotiation every time their usage grows.

This hybrid approach solves the retention gap directly. The subscription floor absorbs short-term usage dips — a customer having a slow month doesn't disappear from your revenue overnight, because they're still paying the base fee. The usage layer on top still lets you capture expansion from your power users without capping their spend at an arbitrary seat count. It's the closest thing to a genuine "best of both" in SaaS pricing, and it's why hybrid models are becoming the default recommendation for companies that already have a usable consumption metric.

Which Model Actually Fits Your Product?

The honest answer to "which model retains customers longer" is: it depends entirely on whether your value metric is consumption-linked.

Run what we'll call the Value-Metric Fit Test: ask whether the thing you'd charge for scales linearly with the value the customer gets. If a customer sending 10x more API calls is genuinely getting 10x more value, usage-based pricing tracks that fairly and retention follows value delivery. If the value is fixed per user — a project management tool, a learning platform, a note-taking app — usage-based pricing makes no sense, because more "usage" doesn't mean more value; it might just mean the product is less efficient.

That's why per-seat and flat-rate pricing still dominate collaboration tools, personal productivity software, and anything where value is delivered per individual rather than per transaction. Nobody wants to be charged more for clicking around a dashboard more often.

There's also a market nuance worth flagging for founders selling into both the UK and US: US enterprise buyers have grown more comfortable with consumption billing thanks to cloud infrastructure norms set by AWS and Twilio. UK mid-market buyers, particularly outside London tech hubs, still lean toward the predictability of an annual flat-rate quote for budgeting reasons — procurement cycles there often want a fixed number to put in front of finance, not a forecast range.

Frequently Asked Questions

Does usage-based pricing increase churn during economic downturns?

It can, because usage-based revenue drops in lockstep with customer usage the moment a budget freeze or seasonal slowdown hits. Flat-rate contracts hold revenue steady until the renewal date regardless of usage changes, giving vendors more runway to intervene before a customer fully churns.

Is net dollar retention the same as customer retention?

No. Net dollar retention measures revenue growth across your existing base and can look strong even while your actual logo count falls, because a handful of expanding accounts can mask a larger number of shrinking or churned ones. Always check logo retention separately before trusting an NDR figure as a loyalty signal.

Should an early-stage SaaS startup use usage-based pricing?

Only if there's a clear, defensible usage metric that maps to customer value — API calls, records processed, seats active. If that metric doesn't exist yet, a flat-rate or per-seat model is usually easier to sell and forecast in the early stages, with a usage layer added once product usage patterns stabilize.

What causes "bill shock" in usage-based SaaS pricing?

Bill shock happens when a customer's invoice spikes unexpectedly due to a usage surge they didn't anticipate or track closely, often from an integration, automated process, or seasonal traffic increase. It's a leading cause of quiet churn because customers respond by cutting usage defensively or shopping for a more predictable-priced competitor.

Can flat-rate and usage-based pricing be combined?

Yes — it's increasingly the default approach among fast-growing SaaS companies. A subscription floor covers baseline costs and stabilizes revenue, while a usage component on top captures upside from heavy users without forcing a full plan renegotiation.

If you're mid-decision on your own pricing model, don't start by picking a side. Start by mapping your actual value metric, check whether it scales with consumption, and prototype a hybrid floor-plus-usage structure before ripping out your existing flat-rate plan entirely.