B2B SaaS Billing & Metering 2026 — Deep Dive on Stripe Billing, Paddle, Lago, OpenMeter, Orb, Chargebee, Recurly, Metronome

Prologue — Billing Is Not Payments

The classic illusion a SaaS founder hits in 2026: "Payments? Just drop in Stripe Checkout."

Three months later: proration is miscalculated and refund requests pile up, EU VAT filings get missed and the accountant explodes, you try to move to usage-based pricing and Stripe alone cannot meter it, and a Korean enterprise customer asks for a Korean tax invoice (세금계산서) you have no idea how to issue.

Billing is not payments. Payments is a single component of billing. Billing means:

Subscription lifecycle — trials, upgrades, downgrades, pauses, cancellations
Proration — daily math when plans change mid-cycle
Usage metering — aggregating API calls, tokens, GBs
Invoicing — PDF generation, tax invoices, accounting integration
Tax — VAT, GST, US sales tax, Korean VAT
Dunning — retry failed payments, notify expiring cards
Revenue recognition — ASC 606, MRR/ARR, deferred revenue
Multi-currency, multi-payment-method, multi-entity

Trying to do all of this with one tool fails in 2026. Billing is full-stack, and each layer has different winners.

This guide dissects the 2026 B2B SaaS billing and metering stack — Stripe Billing, Paddle, Lago, OpenMeter, Orb, Metronome, Chargebee, Recurly, Zuora, plus Korean and Japanese payment gateways.

1. Why Billing Is Hard

Before tools, let us nail down why "just plug in payments" never ends there.

Proration. A user upgrades from Pro to Business on May 15. You owe them 16 days of credit on the old plan and 16 days of charge on the new one. Downgrades — refund, credit, or apply next cycle? That is a policy decision, and the billing engine must express it in code.

Tax. EU VAT differs B2B vs B2C; US sales tax differs across 45 states; Korea charges 10% VAT and mandates electronic tax invoices; Japan adopted the qualified invoice system in 2023. A global SaaS that gets this wrong becomes a tax evader by accident.

Currency. Charging in USD annoys EU customers when the rate moves. Charging in EUR/JPY/KRW makes reconciliation a nightmare.

Trials. 14-day free, card required or not, auto-charge on day 15, mid-trial upgrades.

Upgrades and downgrades. Immediate vs next cycle, seat add vs reduce, usage carry-over on plan switch.

Dunning. Retry cadence on failed cards, churn threshold, email tone, self-service card update page.

Refunds. Partial refunds, credit notes, accounting treatment.

Regional regulation. Korean mandatory e-tax-invoice, EU SCA (Strong Customer Authentication), US OFAC sanctions list, Japan invoice system (effective 2023).

Building this yourself eats six months. That is why billing became an industry.

2. Five Pricing Models — Flat, Per-Seat, Usage, Tiered, Hybrid

Before picking a tool, pick a pricing model. Tools differ in what they handle well.

Flat-fee. $99/mo fixed. Simple, predictable. Notion Personal Pro, Linear Starter — small-team SaaS. Stripe Billing handles this best.

Per-seat. $10/mo per user. Slack, Notion Team, Linear Business. Proration on seat changes is essential. Every tool supports it, but seat-handling details differ.

Usage-based. $5 per 10k API calls, $0.10/GB. Twilio, AWS, OpenAI API. The fastest-growing model from 2023 to 2026. This is why metering tools exist.

Tiered. First 10k calls $0.10 each, 10k–100k $0.08, 100k+ $0.05. Volume discounts. Twilio, SendGrid, Postmark. Or package tiers like Starter $29 (10 seats), Pro $99 (50 seats), Enterprise (custom).

Hybrid. Base fee plus usage overage. Pro $99/mo (10k API calls included, then $0.005 each). Effectively the PLG SaaS default. Datadog, Segment, Linear Business.

2026 trend: usage-based + hybrid is the PLG SaaS default. Per-seat is now enterprise-only, flat-fee is now indie-only. An a16z 2024 survey found 41% of SaaS companies adopted usage-based pricing.

3. Stripe Billing — The Industry Default

Launched in 2018, Stripe Billing is still the de facto standard for SaaS billing in 2026.

Core features.

Subscriptions — full lifecycle (create, upgrade, cancel, pause).
Invoices — auto-generated PDF, email delivery, hosted invoice pages.
Metered billing — usage_records API to report usage; auto-charge at period end.
Customer Portal — hosted page for users to self-serve card, plan, cancellation.
Stripe Tax — global tax calculation and filing assistance. Launched 2021, expanded 2024.
Webhooks — events like invoice.paid, customer.subscription.updated.

Strengths.

Best-in-class developer experience. SDKs, docs, test cards, dashboard are all first-rate.
Card processing (PSP) + billing in one — invoice creation and collection in one flow.
Global reach — 40+ countries, 135+ currencies, Apple Pay, Google Pay, SEPA, iDEAL, Bancontact.
Ecosystem — Zapier, Segment, HubSpot, Salesforce integrations are mature.

Weaknesses.

Usage-based is weak. usage_records is a simple counter; complex metering (dedup, dimensional, late events) you build yourself.
Not an MoR. Stripe is a PSP — tax filing, refunds, chargebacks remain your responsibility. Stripe Tax helps with calculation but you file.
Enterprise gaps. Complex revenue recognition (ASC 606), multi-step approvals, quote-to-cash are weak.

Pricing (2026). 2.9% + $0.30 per card, Stripe Billing adds 0.5%–0.8% of invoiced amount. Stripe Tax is 0.5% of transactions.

Notable users. Notion, Slack, Linear, Vercel, parts of Figma, Cursor, Anthropic Console.

4. Paddle — Merchant of Record (MoR)

Paddle is the canonical MoR (Merchant of Record). The biggest structural difference from Stripe.

What is MoR? Paddle becomes the seller on your behalf. The customer pays Paddle, Paddle handles tax, Paddle remits to you. Concretely:

Tax filing and remittance is Paddle's responsibility. EU VAT, US sales tax, AU GST — Paddle handles them.
Chargeback and refund liability is Paddle's. On Stripe you fight disputes; on Paddle, Paddle does it.
No PCI burden. Paddle holds the card data.

The price: higher fees. Around 5% + $0.50 per card vs Stripe's 2.9% + $0.30. Roughly 2x, but you trade tax and legal load for it.

Who is it for?

Indie SaaS, bootstrapped founders — global sales without hiring an accountant.
B2C SaaS, digital downloads — multi-country tax handling is the headache.
Small SaaS — solo operator with no time for filings.

Who is it not for?

US-domestic-only B2B — lower tax burden, the fee surcharge stings.
Enterprise — Net-30 invoices, ACH and wire are more important.

Paddle Billing (launched 2024). Old Paddle Classic was one-time-purchase oriented. New Paddle Billing covers subscriptions, metering, customer portal — a direct Stripe Billing rival.

Notable users. Userlist, Beamer, parts of BrowserStack, many indie SaaS.

5. Lago — Open-Source Usage-Based Billing

Lago is an open-source billing engine, Y Combinator W21. Started 2021, MIT-licensed.

Positioning. "Open-source Chargebee + Stripe Billing." Built around usage-based.

Core features.

Self-host or cloud — Docker self-hosting or Lago Cloud.
Events API — ingest usage events (API calls, GB processed, messages sent).
Plans and charges — define plans, tiers, packages via code or UI.
Invoices and credit notes — PDF invoices, credit notes, refunds.
Stripe, Adyen, GoCardless integration — payments delegated to your PSP.
Webhooks, GraphQL API, REST API.

Strengths.

Open source — read the code, modify it, self-host. Meets EU data-sovereignty needs and lets you audit metering accuracy.
Usage-based is a first-class citizen — exactly where Stripe Billing is weak.
Developer-friendly — clean API and SDK design, solid documentation.

Weaknesses.

Tax is delegated — bring your own Stripe Tax or Avalara.
Customer portal is basic — customization expected.
Enterprise features still maturing — quote-to-cash, multi-step approval are limited.

Pricing (2026). Self-hosting is free. Lago Cloud is revenue-tied with a free starter tier.

Notable users. Mistral AI, Pinecone, parts of Vercel, many metering-heavy SaaS.

6. OpenMeter — Kubernetes-Native Metering

OpenMeter is an Apache 2.0 usage-metering engine. Started 2022.

Positioning. "Not billing — metering." Leave billing to Stripe or Lago; focus on the accuracy and scale of usage aggregation.

Architecture.

Kafka backbone — events stream into Kafka, scaling to tens of thousands per second.
ClickHouse aggregation — time-series DB for real-time queries, dimensions, and dedup.
CloudEvents standard — event format is portable across systems.
Kubernetes-native — Helm chart and Operator for deployment.

Why a separate metering engine?

Late events — an event arriving an hour late is still aggregated correctly.
Idempotency — dedup on idempotency_key.
Dimensional aggregation — group by customer_id x region x model.
Throughput — Stripe usage_records caps around tens of writes per second; OpenMeter handles tens of thousands.

Who is it for?

LLM API companies — per-token metering across models, customers, regions.
CDN, cloud infra — GB, request, hour metering at scale.
IoT, telemetry — tens of thousands of devices firing events.

Lago vs OpenMeter. Lago is full-stack billing plus metering, OpenMeter is metering only. You can combine them — OpenMeter for aggregation, Lago or Stripe for invoicing.

Notable users. Multiple LLM infra companies, several cloud startups.

7. Orb — The PLG Star of Usage-Based Billing

Orb is Y Combinator W22, started 2022. Focused on usage-based PLG SaaS.

Positioning. "What Stripe Billing cannot do for complex usage pricing." Deep integration with Stripe Tax, Salesforce, NetSuite.

Core features.

Complex pricing — tiered + package + graduated, multi-dimensional, commit + overage.
Real-time usage portal — customer-facing dashboard for live usage visibility.
Salesforce + Stripe + NetSuite — connected to the RevOps full stack.
Revenue recognition — automated ASC 606, accounting integration.

Strengths.

Elegant complex pricing — fills exactly the gap Stripe Billing leaves.
RevOps-friendly — deep ties to Salesforce, NetSuite, QuickBooks.
Real-time visibility — customers see their own usage in real time.

Weaknesses.

Pricing is steep — enterprise tier, overkill for indie devs.
Not a PSP — bring Stripe or Adyen.
Self-serve SaaS optimized — invoice-based enterprise flows live elsewhere.

Notable users. Vercel, Replicate, Cohere, Anthropic API, many PLG SaaS.

Pricing (2026). Not public. Revenue-tied, typically 0.5%–1% of revenue.

8. Metronome — Used by OpenAI, Anthropic, Databricks

Metronome is Y Combinator, started 2019. Full-stack usage-based billing and metering.

Positioning. "Billing for AI and infra companies." Metering and invoicing for LLM APIs and cloud infrastructure.

The users are the positioning.

OpenAI — parts of ChatGPT Plus, Enterprise, and API billing.
Anthropic — Claude API and Console metering and billing.
Databricks — DBU usage billing.
Confluent, Nylas, parts of Twilio Segment.

Core features.

High-scale metering — OpenAI-level traffic.
Complex commit + overage — $1M/yr commit, $0.01/1k tokens overage enterprise deals.
Revenue recognition — ASC 606, multi-step approval.
Salesforce, NetSuite, QuickBooks integration.

Strengths.

Battle-tested at AI and infra scale — handles OpenAI- and Anthropic-grade traffic.
Enterprise RevOps full stack.

Weaknesses.

Very expensive — enterprise pricing only.
Setup is complex — a real RevOps team is required to extract value.

Orb vs Metronome. Similar positioning, different segments. Orb is PLG SaaS mid-market; Metronome is AI and infra enterprise. In real RFPs they show up on the same shortlist.

9. Chargebee — Traditional Subscription Billing Leader

Chargebee started in Chennai in 2011. As of 2026, it remains a traditional leader in subscription billing.

Positioning. "Between Stripe and Salesforce." Subscription full stack for mid-market and enterprise SaaS.

Core features.

Subscription management — full lifecycle.
Smart dunning — AI-driven retry optimization.
Revenue recognition — ASC 606, GAAP.
Quote-to-cash — quotes, contracts, invoicing, collection in one flow.
Multi-PSP — Stripe, Adyen, Braintree, Authorize.Net, more.
Multi-currency, multi-entity.

Strengths.

Battle-tested at enterprise scale — Freshworks, Calendly, parts of Okta, tens of thousands of customers.
PSP-agnostic — pick and switch payment gateways freely.
RevOps full stack — Salesforce, HubSpot, NetSuite, Xero, QuickBooks integrations.

Weaknesses.

Usage-based is average — weaker than Orb, Metronome, Lago.
Pricing is steep — mid-market SaaS pay thousands of dollars per month.
Setup is complex — RevOps headcount needed.

Notable users. Calendly, Freshworks, parts of Okta, parts of Pinterest.

10. Recurly — Another Traditional Subscription Billing Leader

Recurly started in 2009. Like Chargebee, it sits in the mid-market-to-enterprise tier in 2026.

Positioning. Nearly the same segment as Chargebee. Differences are fine details and pricing negotiation.

Core features.

Subscription lifecycle.
Revenue recognition.
Smart retries on failed cards.
Tax automation (Avalara, Vertex integrations).
Multi-PSP, multi-currency.

Strengths.

North American and European media and entertainment leader — Sling TV, AccuWeather, FabFitFun.
Polished dunning logic.
Clean customer portal.

Weaknesses.

Nearly identical to Chargebee. Usage-based remains average.

Chargebee vs Recurly. Same market, similar features. Sales negotiation, pricing, and existing integrations often decide the choice.

11. Maxio — Chargify + SaaSOptics Merger

Maxio was born in 2022 from the merger of Chargify (billing) and SaaSOptics (SaaS metrics and finance).

Positioning. "Full stack for the B2B SaaS CFO." Billing + SaaS metrics (MRR, ARR, LTV, churn) + financial reporting in one.

Core features.

Advanced Billing (formerly Chargify) — subscription, usage, hybrid billing.
Advanced Financial Reporting (formerly SaaSOptics) — MRR, ARR, churn, cohort analysis, ASC 606.
Accounting integration — NetSuite, Sage Intacct, QuickBooks, Xero.

Strengths.

CFO-friendly — billing and financial reporting in a single tool.
SaaS metrics full stack — MRR and ARR dashboards without a separate BI tool.

Weaknesses.

Usage-based is average — the Chargify-era weakness remains.
Developer experience is average — falls short of Stripe and Lago.

Notable users. Many mid-market SaaS, especially adoptions driven by the finance team.

12. Zuora — The Enterprise Subscription Monolith

Zuora started in 2007, IPO'd on NYSE in 2018. The absolute leader in enterprise subscription billing.

Positioning. Subscription infrastructure for Fortune 500 media, telecom, and SaaS. An enterprise tool comparable to Salesforce.

Notable users. Zoom (pre-IPO), Box, DocuSign, NCR, GE, Honeywell, Schneider Electric, Sky, The New York Times.

Core features.

Zuora Billing — subscription lifecycle.
Zuora Revenue — full-stack ASC 606 revenue recognition.
Zuora CPQ — quotes, contracts, pricing.
Zuora Collect — collections and dunning.

Strengths.

Battle-tested at enterprise scale — millions of customers, billions in revenue.
Deep Salesforce integration — sales, billing, revenue recognition flows are seamless.
Complex business models — B2B2C, marketplaces, usage + commit.

Weaknesses.

Very, very expensive — six figures USD and up.
Setup is very complex — typically 6 to 18 months, requires external consulting.
Not developer-friendly — a sales and consulting product.

Who is it for? SaaS at $50M+ ARR, or large media, telecom, hardware companies moving to subscription.

13. Accounting Integration — Sage Intacct, NetSuite, QuickBooks Online

Billing is not the finish line. Billing data must flow into your accounting system (ERP).

QuickBooks Online (QBO). Default for indie and small SaaS. Number one share in US and Canada. Stripe, Chargebee, Lago all integrate natively.

Xero. QBO competitor, strong in Australia, New Zealand, and the UK. Popular with indie SaaS.

Sage Intacct. Mid-market SaaS standard. AICPA-recommended. Strong at ASC 606 for SaaS. Deep integration with Chargebee, Maxio, Zuora.

NetSuite (Oracle). The mid-market-to-enterprise ERP heavyweight. Covers SaaS, e-commerce, manufacturing. Effectively the default for $50M–$500M SaaS.

SAP S/4HANA. Fortune 500 ERP. Massive SaaS companies eventually end up here.

FreshBooks. Focused on freelancers and solo operators.

Strategy. When picking a billing tool, ask first: "Is integration with our ERP native, or do we build it?" Self-built integrations always break.

14. MoR vs PSP — Who Owns Tax and Refunds

The single most important distinction in this guide, restated.

PSP (Payment Service Provider). Stripe, Adyen, Braintree, Worldpay. Card processing only. You are the seller — tax filing, refunds, chargebacks are yours.

MoR (Merchant of Record). Paddle, LemonSqueezy, FastSpring, 2Checkout, Cleverbridge. They are the seller. They take on tax filing, refunds, chargebacks. The fee is higher.

Item	PSP (Stripe)	MoR (Paddle)
Who is the seller	You	The MoR
Tax filing	You	The MoR
Refund disputes	You	The MoR
Chargeback cost	You	The MoR
PCI compliance	You (Stripe handles most)	The MoR
Fee	2.9% + `$0.30`	5% + `$0.50`
Flexibility	High	Lower (MoR rules apply)

When MoR?

Indie SaaS, one- or two-person operation.
Global B2C, multi-country digital downloads.
Stage where hiring an accountant is a burden.

When PSP?

US-domestic-only or home-market focused.
Enterprise B2B — invoices, wire, Net-30 matter more.
Revenue large enough for a finance team.

15. LemonSqueezy — Acquired by Stripe in 2024

LemonSqueezy is an indie MoR launched in 2021. Acquired by Stripe in July 2024.

Positioning. "MoR for indie SaaS and digital products." The SaaS version of Gumroad.

Why did Stripe buy it? Stripe was a PSP, and the MoR market (Paddle, FastSpring) was growing on indie SaaS. The LemonSqueezy acquisition let Stripe ship MoR features directly — in 2025 they announced the Stripe Tax + MoR integration.

Today (2026). LemonSqueezy brand remains, Stripe backend integration is in progress. New indie SaaS founders typically compare LemonSqueezy with Paddle.

vs Paddle. Paddle is stronger for B2B SaaS and mid-market; LemonSqueezy is more indie- and solo-developer-friendly in UX.

16. FastSpring, 2Checkout, Cleverbridge — Other MoRs

The MoR market has older players beyond Paddle and LemonSqueezy.

FastSpring. Started 2005. SaaS, games, digital products. Global tax handling. Used by Atlassian, Cisco, JetBrains.

2Checkout (Verifone). Started 2006, acquired by Verifone in 2017. Global payments + MoR. 87 currencies, taxes in 45 regions.

Cleverbridge. Started 1995 in Germany. Global B2B and B2C e-commerce. Strong with security and utility software vendors like Acronis, Avira, Corel.

Strategy. Paddle and LemonSqueezy are more modern, but FastSpring, 2Checkout, and Cleverbridge are older and broader. For something like JetBrains, selling in 90+ countries, FastSpring is the natural fit.

17. Tax Stacks — Stripe Tax, TaxJar, Avalara, Vertex

The tax stack matters as much as the billing tool choice.

Stripe Tax. Built into Stripe, automatic rate calculation, registration help. Covers EU VAT, US sales tax, Canadian GST, Australian GST, Japanese consumption tax. Price: 0.5% of transactions. Default for indie and small SaaS.

TaxJar. Started 2013, acquired by Stripe in 2021. US sales tax specialist with automated filing (AutoFile). Covers all 50 US states. Stripe acquired it but kept it as a standalone product.

Avalara. Started 2004, NYSE IPO (taken private by private equity in 2022, fully private by 2024). The enterprise tax automation heavyweight. AvaTax, Returns, CertCapture. Fortune 500 users. Integrated by Chargebee, Recurly, NetSuite, Zuora.

Vertex. Started 1978. Enterprise tax specialist. Direct Avalara competitor. Strong with large manufacturing and retail.

Sovos. Born from a 2014 merger. Global tax and invoice compliance. Strong with EU e-invoice mandates.

Sphera. Sovos subsidiary. Environmental and regulatory compliance specialist.

Strategy. By revenue stage — $0–$5M Stripe Tax; $5M–$50M Avalara or Vertex; $50M+ full-stack Avalara plus an in-house tax team.

18. Korean Payments — Toss, KG Inicis, NICE, KCP, KakaoPay

Korean SaaS payment flow differs from global SaaS. PG (Payment Gateway) is separate from the card networks, and tax invoices are mandatory.

Toss Payments. Subsidiary of Viva Republica (Toss). Rapidly rising share among Korean PGs as of 2026. The best developer experience among Korean PGs — Stripe-level SDK, docs, and test environment. Supports recurring, authenticated, easy payments, and overseas.

KG Inicis. Started 1998, first-generation Korean PG. Used by many indie and small commerce sites. APIs are dated, integration cost exists. Still has the most stable coverage of all Korean card networks.

NICE Payments. Part of NICE Group, a Korean info-infrastructure heavyweight. Card auth, recurring, easy payments, overseas. Many enterprise and finance customers.

KCP (Korean Cyber Payment). Started 1996, NHN subsidiary. With KG Inicis, the two pillars of first-generation PGs.

KakaoPay. Started 2014. Strong in B2C easy payments, expanding into B2B recurring. A big option for Korean SaaS before Stripe Korea arrived.

Stripe Korea. Started entry work in 2022, launched officially in 2024. Simplifies Korean billing for global SaaS. But Korean tax-invoice issuance is separate — typically integrated with KISA's electronic tax-invoice system.

Standard Korean billing pattern.

Global SaaS: Stripe + separate tax-invoice flow for Korean entity sales.
Korean SaaS: Toss Payments + in-house recurring + tax-invoice system (e.g., Bill36524).

19. Japanese Payments — GMO, SoftBank, Stripe Japan, PayPay

Japan resembles Korea but has its own quirks. Consumption tax + qualified invoice system (introduced 2023) is the wildcard.

GMO Payment Gateway. Started 1995, number one PG share in Japan. Covers cards, convenience-store payment, bank transfers, e-money. Stable across card networks.

SoftBank Payment Service. Part of the SoftBank group. Direct GMO competitor.

Stripe Japan. Launched 2016. Full JCB support, Japanese-language dashboard and docs. Brought global SaaS payment standards to Japan.

PayPay. 2018 joint venture between SoftBank and Yahoo Japan. Dominant QR-code B2C payment share. SaaS usually uses cards; PayPay is for B2C and retail.

LINE Pay. LINE-based, used in Japan, Taiwan, Thailand. Rarely used by Japanese B2B SaaS.

Pay.JP. Japanese indie SaaS PG. Strong JCB support.

Standard Japanese billing pattern.

Global SaaS entering Japan: Stripe Japan + qualified-invoice system (MoneyForward, freee).
Japanese SaaS: GMO Payment + in-house systems.

20. Japanese SaaS Billing — freee, MoneyForward, smartHR, Yappli

Japanese SaaS billing and finance stack also differs from Korea and the US.

freee. Started 2012, listed on the Tokyo Stock Exchange. Number one cloud accounting for Japanese SMEs. Full-stack billing, invoicing, tax filing, and payroll. The de facto standard for automating qualified-invoice issuance.

MoneyForward. Started 2012. Direct freee competitor. Accounting, billing, payroll, expense management. Strong in mid-market and enterprise.

smartHR. Started 2015. Number one cloud HR in Japan. Not a billing tool, but a pillar of the Japan SaaS full RevOps stack alongside freee and MoneyForward.

Yappli. Started 2013, no-code mobile app builder. Built its own billing and subscription management tailored to the Japanese market.

Strategy. SaaS entering Japan must integrate with freee or MoneyForward for invoicing. Korean and US accounting tools cannot produce Japanese qualified invoices.

21. Korean SaaS Examples — Channel.io, Allganize, NinjaCart, Toss Payments

How Korean SaaS solves B2B billing in practice.

Channel.io (Channel Talk). Started 2014, expanded to Japan. Seat-based + usage-based hybrid. In-house billing + Toss Payments for KR, Stripe for overseas.

Allganize. Started 2017, enterprise LLM. Enterprise quote-to-cash flow, in-house billing + Stripe.

NinjaCart. Indian-origin but global, agriculture SaaS. Multi-country payments.

Toss Payments. Although a PG, it also builds SaaS-billing tooling — the Toss Payments recurring API. Becoming the de facto standard for Korean SaaS.

Korean B2B SaaS billing patterns.

Seed to Series A: Toss Payments + custom billing code (regretted in three months).
Series B+: Stripe Korea + Chargebee or Lago + Toss Payments as auxiliary.
Going global: Stripe (global) + Toss Payments (KR) + tax-invoice system.

22. Webhooks and Idempotency — The Heart of Billing Reliability

Where billing breaks most often: webhook handling.

The problem. Stripe, Paddle, Lago send invoice.paid. Your server marks the user as paid in the DB. But:

The same event can arrive twice (Stripe retries).
If your server dies mid-processing, Stripe retries.
Order can flip (invoice.paid before subscription.created).

The fix: idempotency.

Store event.id in the DB. Skip events already processed.
Make every handler idempotent. Same event 10 times must produce the same result.
Queue webhooks for async processing. Return 200 OK fast, do real work in background jobs.
Design jobs to be retry-safe. DB transactions + idempotency-key patterns.

Tools. Stripe and Paddle stamp events with an id. Inngest, Trigger.dev, Hatchet, Temporal are 2026's standard pattern for queueing webhooks and processing idempotently.

23. RevOps — ASC 606, MRR, ARR, Churn

Billing is not just collecting money; it feeds into financial reporting. This domain is RevOps (Revenue Operations).

ASC 606 (Revenue Recognition). US accounting standard. Cash collected is not revenue. Revenue is recognized as you deliver the service. Annual subscription cash collected upfront is recognized at 1/12 per month. The core of SaaS accounting.

MRR (Monthly Recurring Revenue). The most important SaaS metric.

ARR (Annual Recurring Revenue). MRR x 12. The standard investor headline.

Churn. Logo churn (customer count), revenue churn, net churn (after expansion).

Expansion revenue. Upgrades and seat additions from existing customers. A healthy SaaS has net revenue retention above 110%.

Tools. Maxio, ChartMogul, ProfitWell (a Paddle subsidiary), Baremetrics. Stripe Sigma lets you SQL-query the raw data.

Strategy. Seed: Stripe Sigma + a spreadsheet; Series A: ChartMogul or ProfitWell; Series B+: Maxio or NetSuite full RevOps.

24. Open-Source Alternatives — Killbill, Crater, InvoiceShelf

When commercial billing is too expensive, open source is alive in 2026.

Killbill. Started 2010, Apache 2.0. A Java billing engine built by ex-JBoss engineers. Enterprise scale. Can handle larger sizes than Lago, but Java + ops overhead is heavy.

Lago. Covered above. Modern open-source metering + billing.

Crater. Laravel/PHP invoicing tool. Not full-stack SaaS, but solid for simple invoicing.

InvoiceShelf (formerly InvoicePlane). PHP invoicing. A simple freelance tool.

Cosmos.js, Laravel Cashier — framework-native billing helpers. Stripe and Paddle integration as code.

Strategy. Go 100% open source when (1) EU data sovereignty applies, (2) you have government or defense compliance, or (3) policy requires self-hosting. Otherwise commercial billing tool prices are usually less than your time cost.

25. Decision Matrix — Which Billing Tool Should You Pick

The summary table.

Stage / Model	Recommended Stack
Indie SaaS, B2C, 1–2 person	Paddle or LemonSqueezy (MoR)
Indie SaaS, B2B	Stripe Billing + Stripe Tax
PLG SaaS, Seed–Series A, usage-based	Stripe + Lago or Stripe + Orb
PLG SaaS, Series B+, AI/infra	Stripe + Metronome or Stripe + Orb
Per-seat mid-market SaaS	Chargebee or Recurly + Stripe/Adyen
Mid-market SaaS + finance full-stack	Maxio + Stripe
Enterprise (`$50M+` ARR)	Zuora + Salesforce + NetSuite + Avalara
Korean SaaS, B2B	Toss Payments + in-house billing, or Stripe + tax-invoice
SaaS entering Japan	Stripe Japan + freee/MoneyForward (qualified invoices)
Government, defense, EU sovereignty	Self-hosted Lago/Killbill + your own PSP

Three core principles.

Match the stage. Adopting Zuora at seed eats 12 months. Start with Stripe Billing and swap as revenue grows.
Usage-based means a separate metering engine. Stripe usage_records alone is not enough. Pick from Lago, OpenMeter, Orb, Metronome.
Take MoR seriously if you sell globally. For indie SaaS without an accountant, Paddle and LemonSqueezy's 5% fee is the price of freedom.

26. Ten Antipatterns

Common mistakes you will see in billing systems.

Determining paid status from a paid_at column. Misses refunds, chargebacks, claw-backs. Always read subscription.status.
Doing heavy work synchronously in webhook handlers. Stripe's 5-second timeout triggers retry storms. Queue work async.
Processing webhooks without idempotency keys. Charges customers twice for the same event.
Recognizing amount_paid as revenue. ASC 606 violation. Allocate over the service period.
Adding tax after the fact. Sticker price vs charged price diverge, triggering refund storms. EU requires VAT-inclusive display.
Hand-rolling proration. Stripe and Chargebee already do it; rewriting it produces bugs.
Ignoring time zones. UTC billing dates collide with user local time.
No dunning policy. No card-expiry notice or retry cadence — you leak 5–10% of revenue.
No customer portal. Customers cannot update cards or cancel, escalating to chargeback disputes.
Not piping billing data to the ERP. Your accountant cries at quarter-end close.

27. References

— End of B2B SaaS Billing & Metering 2026.