Zuora vs Paddle: Choose the Right Billing Fit 2026

The choice is who sells the subscription

Zuora vs Paddle is not a beauty contest between two invoice screens. It is a choice about who is the seller on the customer’s bill. Zuora Billing is software your company runs while you stay the merchant, connect payment gateways, and own tax. Paddle Billing is a merchant-of-record service: Paddle sells the digital product, takes the payment, and says it remits indirect tax. A finance lead who picks on brand familiarity will mis-model both the cash fee and the control they give up. Neither name was retired. Zuora was taken private and kept shipping Billing, Revenue, CPQ, and Collect. Paddle added Billing in August 2023 and still documents Classic as a platform you port away from, not a dead brand.

A merchant of record is the legal seller that collects the customer’s money and carries the indirect-tax duty on that sale. TL;DR: choose Zuora when your company must remain the seller and the hard problem is rating a messy catalog, and choose Paddle when you want one party to run checkout, subscriptions, and tax, while treating both software prices as quote-based until a contract says otherwise.

Silver Lake and GIC closed the purchase on February 14, 2025 at $10.00 a share, the entry lists 1,618 employees for 2024 and revenue of $396 million for 2023, and the same entry names Billing, Revenue, CPQ, and Collect as the primary applications, a Take-private value: $1.7 billion according to Wikipedia (2025). The product was not discontinued in that deal. Tien Tzuo is still listed as CEO. What changed is the owner, not the job the software does.

Key Takeaways

• Zuora stays your merchant account and your tax file. Paddle becomes the seller and says it files the indirect tax.

• Zuora’s public software price is quote-based. Paddle’s pricing page, opened for this guide, shows no percentage, states no monthly fee, and sends products under $10 to a bespoke quote.

• Zuora’s own billing page claims $96+ billion processed a year, 50 pricing models, and 40+ gateway connectors. The entitlements grid includes only 2, 5, or 10 gateways depending on Core, Pro, or Elite.

• Paddle Billing launched in August 2023. Paddle says it has moved over 1 million Classic subscriptions, in a 4-step port that is still described as early access.

• A review rollup puts Zuora at a 78% satisfaction rating across 660 reviews, with 3.9 listed from 305 reviews. That is a mixed signal, not a reason to skip the seller question.

• An orchestration layer above either tool is a review queue with a human release. It is not a second billing engine and it is not a measured saving.

Who this is for

This guide is for a SaaS finance lead, or the controller who sits with them, at a firm still deciding who should be the seller of the subscriptions: the company, choosing a system that will rate, collect, and explain those charges, or a merchant of record that sells in the company's place. You are close enough to a decision that a feature slogan is useless. You need to know who the customer is buying from, which catalog shapes you can express, what is actually printed on a pricing page, and what your team must still build before the general ledger will trust the file.

Red flags: you need one product to be both the legal seller and a full revenue-recognition application; you sell mostly physical goods that ship in a box; you refuse to talk to sales until a public page prints a software fee in dollars.

If the first red flag is you, stop the head-to-head. Zuora’s encyclopedia entry lists Revenue as a separate application from Billing, and Paddle’s pricing page is a checkout, tax, and subscription offer, not a revenue subledger. If the second red flag is you, Paddle’s docs are about digital products and subscriptions, and forcing that model onto freight is the wrong project. If the third is you, both vendors will waste your week: the pages opened for this guide do not print a Zuora software fee or a Paddle percentage.

How we evaluated these tools

The weights below are buying priorities for this finance decision, not a lab score. No tenant was opened, and nothing here is a test result. Each weight is the share of the decision that should move if that topic fails. Two people can swap a few points and still land in the same place, because the seller question is large on purpose. A tool that loses on legal seller does not get rescued by a nicer settings screen.

Legal seller is 22% because it decides the invoice name, the tax registration, and whether card data sits in your gateway account. Charge-model depth is 18% because a SaaS catalog that mixes seats, usage, commits, and services will outgrow a simple recurring price. Payment control, ledger evidence, implementation, and pre-quote price clarity are 15% each. Together those four are 60%, so a vendor that is only “easier” still has to clear them. Public price clarity is weighted even though both offers are quote-based, because a finance lead should see that fact before the call, not after procurement has started.

What each name still means

Zuora is the same Billing product family a public-company buyer would have evaluated, now under private owners. The billing product page lists 50 pricing models, 40+ gateway connectors, 20+ payment methods, and usage-based monetization at 46% of Zuora Billing customers, alongside Annual volume: $96+ billion according to Zuora (2026). That page also says the company was named a Leader in the 2026 Gartner Magic Quadrant for Recurring Billing Applications and a Leader in the 2025 Forrester Wave for Recurring Billing Solutions. Treat those lines as the vendor’s citations of analyst reports, not as a score from this guide. The same page describes a native tax engine and names Sovos and Avalara among pre-built connectors. You remain the seller.

Paddle, on the pages opened here, is still Paddle. The current platform described for new work is Paddle Billing. Classic remains in the migration docs because older accounts still have it. Paddle’s pricing page does not show the percentage many older write-ups repeat. What it does show is an all-inclusive story with no monthly fee, custom pricing for high volume, and a bespoke quote if the product sells for under $10. Reading a remembered rate into that page would be a guess, so this guide does not.

Limits you can read on a page

Included counts and marketing totals are different numbers, and mixing them is how a business case goes wrong. The 40+ gateway connectors on the billing product page are a catalog. The contract grid is smaller. Included payment gateways are 2, 5, and 10 across Core, Pro, and Elite, tax connectors are 1, 1, and 2, analytics datasets are 10, 75, and 100, custom fields per object are 50, 75, and 100, developer sandboxes are 1, 1, and 2, and university seats are 2, 3, and 4, with Audit retention: 7 years according to Zuora (2025). Those entitlements start February 1, 2025. Older contracts can differ, which the page says in plain language. Blank cells on that grid are not treated here as a yes. If a row such as fraud tooling has no count, ask the quote to say included or not.

Paddle does not publish a matching three-tier grid. The comparison below uses only contrasts that the opened pages support. Where a count is missing, the cell says so.

A proposed configuration from US Tech Automations starts when the billing system emits a completed charge or drops a scheduled invoice file. The action is to compare that file with the contract schedule, hold any row with a mismatched amount, tax, or missing customer id, and write the hold to a review queue. The output is a morning exception list plus a clean file that does not post until a person releases it. Prerequisites are an API key or a nightly export, and a written map from product to ledger account. This is a configuration sketch, not a live deployment and not a measured result.

Pricing checked on the vendor pages

Pricing checked October 13, 2026. Zuora does not print a dollar fee for Core, Pro, or Elite on a pricing page this guide could open. The names exist on the entitlements grid as package labels. The price is quote-based. Paddle’s pricing page also withholds a percentage. It states there is no monthly fee, no added cost for managing payments, extra payment methods, subscription migration, or payment-related buyer support, no lock-in period, and a bespoke quote for products under $10, according to Paddle (2026). High-volume and low-price work is described as custom. Implementation help from solutions architects is listed as an additional service, not as a number.

The $0 cell means the page text “no monthly fee,” not a negotiated rate. Do not add a percentage in the model until Paddle writes one into a proposal. Do not invent a Zuora platform fee from a blog. Total cost still includes people. Zuora’s included gateway count, tax connector count, and sandbox count change the services you buy around the license. Paddle’s bespoke band under $10 changes the cost if your average sale is small, even though the headline story is all-inclusive. Ask both vendors for the export you will use at close, and price the human review of that export separately from the license.

Zuora for a catalog that will not sit still

Zuora fits a SaaS company that must stay the merchant, already has or will hire a billing owner, and sells a mix of recurring, usage, one-time, and services charges. The product page’s 50 pricing models and the 46% usage-monetization line are the vendor’s evidence that hybrid rating is the point of the product, not a side feature. Multi-year ramps, account hierarchies, and a quote path into Salesforce are described on that same page. Across 660 reviews the rollup shows 3.9 from 305 reviews on the same page and Satisfaction rating: 78% according to SelectHub (2026). A 78% satisfaction rating means a large minority is unhappy. Read that as a warning about complexity, not as a verdict on tax.

The limits are structural. You own tax registrations, or you pay a tax vendor to help, and the entitlements grid includes one tax connector on Core and Pro and two on Elite. You own the gateway relationship. A Core quote that includes 2 gateways does not become 40+ gateways because the marketing page lists a catalog. Fraud tools, payment retry, and many other rows are named without a rendered yes in the grid this guide opened, so they belong on the question list. Revenue recognition lives in the separately named Revenue application. Buying Billing does not, by itself, prove you bought the close process. There is no public software price, so a total-cost model that uses a borrowed monthly number is fiction.

Implementation is a catalog project. You will define products, rate plans, and charges, map bill-to country and state so tax can calculate, and decide which subscriptions invoice alone. The grid includes 1 developer sandbox on Core and Pro and 2 on Elite, plus 2, 3, or 4 university seats. That is training capacity, not a statement that go-live takes a set number of weeks. Nothing opened here publishes an implementation duration. Plan on a human review of the first several bill runs, a chart-of-accounts map, and a named owner for amendments. The evidence to attach to the business case is the billing product page and the entitlements page cited above, plus the contract that finally prices the tier.

Paddle for a seller who will hold the tax

Paddle fits a software company that wants checkout, recurring billing, fraud handling, and indirect tax in one merchant-of-record seat, and that can accept Paddle as the seller on the receipt. The pricing page says Paddle collects the tax, tracks deadlines, and files returns. The Billing docs describe prices in many currencies, a hosted or embedded checkout, trials, proration, and invoices as part of the newer platform. You are buying an operated selling path, not a database you host logic in. That is the right trade when the finance team is small and the sales footprint crosses many tax jurisdictions.

Paddle Billing launched in August 2023, the subscription port is described as 4 steps and still early access in the dashboard, the docs cite over 180 improvements since 2023, and Subscriptions migrated: over 1 million according to Paddle (2026). Classic is not a renamed ghost. If your account predates Billing, you reintegrate and then port. Historic behavior does not move itself. The pricing page’s promise of no added migration fee for subscriptions does not remove the engineering work of pointing your app at the new API and webhooks.

The limits follow from the role. The percentage is not on the page opened on October 13, 2026, so the take rate is quote-based. Products under $10 are called out for bespoke pricing, which matters if you sell low-priced add-ons. Invoicing and advisory services are presented as additional, not as the default line. You do not pick from a catalog of 40+ gateways, because Paddle is the merchant. Leaving later is a data-migration conversation: the page says there is no lock-in and that data can move, and it does not spell out whether card credentials move. Ask that in writing. Paddle is not described on these pages as the revenue-recognition subledger a public SaaS company may still need beside it.

Implementation can be lighter than a Zuora catalog build and still not be a weekend. You need a product and price catalog in Paddle, a webhook endpoint, a decision on overlay versus inline checkout, and a person who will read failed payments. Solutions-architect help is a separate service on the pricing page. The 4-step port is the vendor’s path for Classic subscriptions, not a claim that every history table comes with them. Keep a human check on the first payout reconciliation against your own revenue schedule before you trust the net deposit as the book number.

Illustrative math for one amendment month

Here is illustrative arithmetic for a desk, not a case study, not a vendor price, and not a measured outcome. Suppose the team touches 400 subscription amendments in a month and a clerk spends 12 minutes re-keying each one, which is 80 hours. At an assumed loaded cost of $65 an hour, that re-keying is $5,200. If a match on the Paddle event transaction.completed clears the 85% of events (340) that carry a subscription_id and an amount that agrees to the contract, the other 15% (60 events) still take 12 minutes each, which is 12 hours and $780. The difference is 68 hours and $4,420 of clerk time, before anyone debates a quote-based vendor fee.

The $65 rate is a planning assumption so the hours have a dollar figure. Replace it with your own loaded cost. The 400, the 12 minutes, and the 85% clear rate are scenario inputs, not statistics from either vendor. The only vendor-specific fact inside the paragraph is the event name, taken from Paddle’s webhook docs: transaction.completed fires when a transaction is fully paid and processed, and the payload can include subscription_id.

US Tech Automations, in a second proposed path aimed at failed renewals, would listen for transaction.past_due, open a retry record, and stop after a fixed number of attempts so a collections owner sees the account before any customer email is sent. The output is a reviewed retry log. The path does not charge a card by itself. It needs the billing vendor’s event feed and a named human owner, and it remains a sketch rather than a current customer result.

Building the handoff yourself

The real alternative to a configured handoff is to stitch the same steps in Zapier, Make, or n8n, or to have an engineer post journals from a script. Those tools can keep run histories, retry a failed call, branch when a payload is incomplete, and store an audit trail when you turn those features on. They do not decide your idempotency key, your access model, or who wakes up when the same event arrives twice. That design work stays with you, and it stays with you after the person who built the scenario leaves.

A proposed US Tech Automations design can place the review stop and the exception file in one configurable path, but only after the export or webhook exists and the ledger map is written. A person still approves any row the match cannot clear. The difference is where the review rule lives and who maintains it, not a promise that exceptions disappear. If your volume is the 60-event remainder in the illustration above, a carefully owned scenario in Make or n8n may be enough. If several systems emit charges, the maintenance of those scenarios becomes the finance control problem.

Words that change the bill

• Merchant of record: the legal seller on the transaction. Paddle takes this role. Zuora does not.

• Quote-based: no dollar software price was on the vendor page opened for this guide. Do not fill the blank from memory.

• Rate plan and charge: Zuora’s catalog language for how a product is priced. A product can carry more than one rate plan.

• Usage rating: a charge driven by consumption, which Zuora documents as a core billing shape and which 46% of its billing customers are said to use.

• Proration: the partial-period charge or credit when a subscription changes mid-cycle. Paddle’s Billing docs describe automatic proration.

• transaction.completed: a Paddle webhook that means the transaction is paid and processed. It is an event name, not a fee.

• Audit trail: on Zuora’s grid, billing audit retention is 7 years on Core, Pro, and Elite. Confirm your contract matches that grid.

• Payout: the net amount Paddle sends you after it has sold as merchant. It is not the same number as gross bookings.

Checklist before you request quotes

Use this list in the order written. Skipping to price repeats the mistake this comparison is meant to stop.

1. Write one sentence on who must be the seller. If the answer is “our company,” Paddle fails that test. If the answer is “not us,” Zuora fails it.

2. List the charge shapes you will sell in the next year: recurring, usage, commit, one-time, ramp, professional services. Strike any vendor that cannot show that shape in current docs.

3. For Zuora, ask the quote to state the tier, the included gateway count (2, 5, or 10), the tax connector count, the sandbox count, and whether Revenue is in or out.

4. For Paddle, ask the quote to state the percentage, the fixed fee if any, the treatment of sales under $10, and whether card credentials move if you leave. Do not assume a rate the pricing page does not show.

5. Name the ledger file, who reviews it, and what happens to a mismatched row. If the answer is “the tool just posts,” you do not have a control.

6. Price clerk time with your own hourly cost. The illustration above used $65, 400 amendments, and 12 minutes only to show the arithmetic.

7. Other buyer notes on this site, if you are assembling a wider memo, include a SaaS tools shortlist, a narrower SaaS tool note, and a month-end pacing recipe.

Mistakes on this shortlist

The first mistake is comparing a gateway fee with a merchant-of-record offer as if they were the same line on the income statement. Paddle’s fee, once quoted, is meant to bundle payment, tax work, and billing. Zuora’s quote is software. You still pay gateways and tax compliance beside it. A lower software invoice can be a higher total, and the reverse is also true. Model both columns or do not model.

The second mistake is reading 40+ gateways as the number in a Core contract. The entitlements page says 2 included on Core, 5 on Pro, and 10 on Elite, with more available at extra cost. The third mistake is treating Paddle Classic and Paddle Billing as one API. Billing launched in August 2023, and the vendor’s own migrate page says you reintegrate and then port subscriptions. The fourth is pasting a percentage into a spreadsheet because a secondary site still prints one. The page opened on October 13, 2026 did not. The fifth is buying either product as a substitute for a revenue-recognition process. Zuora lists Revenue as its own application. Paddle does not claim that job on the pages used here.

Questions before you sign

Is Zuora still the public company on the New York exchange?

No. The take-private closed on February 14, 2025 at $10.00 a share, a $1.7 billion deal, and the stock was delisted. Billing, Revenue, CPQ, and Collect are still the applications named for the company. You are buying software from a private firm owned by Silver Lake and GIC, not from a discontinued product.

Did Paddle shut Classic and rename it Billing?

No. Paddle Billing launched in August 2023 as a new platform with its own API and webhooks. Classic customers reintegrate and port subscriptions. Paddle says over 1 million Classic subscriptions have moved, in a 4-step dashboard port that the migrate page still marks as early access. If you are already on Classic, budget a project, not a logo swap.

Is there a public price for either product?

Zuora is quote-based. Paddle’s pricing page does not show a percentage either. It does state no monthly fee, no lock-in, no added fee for migrating subscriptions, and bespoke pricing under $10. Put “quote-based” in the model for both software rates until a proposal arrives. Pricing checked October 13, 2026.

Who files sales tax and VAT?

Paddle says it collects the right rate and files the returns, because it is the seller. Zuora leaves that duty with you. Its billing page describes a native tax engine and connectors, and the entitlements grid includes 1 tax connector on Core and Pro and 2 on Elite, plus 1 e-invoicing country on each tier before you pay for more. “Tax supported” is not the same sentence as “tax filed by the vendor.”

Can one company run Zuora and Paddle together?

Only with your eyes open. A plausible split is Paddle as the seller for self-serve software checkout, and Zuora as the rating engine for a complex enterprise catalog where you remain the merchant. You then have two systems of record, two cash stories, and a harder reconciliation. Most teams should pick the seller model first and live with one billing core. A second tool is justified when the motions truly differ, not when two departments each prefer a logo.

When is an extra workflow layer the wrong buy?

US Tech Automations is the wrong layer when the billing vendor already delivers a monthly file the bookkeeper trusts, when the only gap is the vendor’s own dunning screen, or when one accountant can reconcile in the ledger without another system in the middle. In those cases the simpler tool you already pay for should win. Add a review path when several charge sources must be matched and a person has to release the journal. Do not add one to decorate a clean export.

The handoff after you pick

Pick Zuora if you must remain the seller, you can staff a catalog, and your quote spells out gateways, tax connectors, sandboxes, and whether Revenue is included. Pick Paddle if you want a merchant of record, your goods are digital, you can accept a quote for the rate, and you have asked what happens to data if you leave. Walk away from both if you need merchant-of-record tax and a full revenue application in a single published price. The hours in the amendment illustration, 80 versus 12, are a way to value the review work that remains after either signature.

Once that seller decision is made, the leftover risk is the file that hits the ledger. You can see how US Tech Automations configures this as a proposed review step on top of the vendor you pick. The export and a human release are still required, and no figure in that design is a promise from a live book of customers.