Merchant of Record (MoR) vs. Direct Merchant Account (Stripe): Unit Economics at $1M–$10M ARR
For digital software and B2B SaaS companies, payment infrastructure design directly dictates gross margin efficiency, global tax exposure, and engineering overhead.
The two primary architectures available are:
1. Merchant of Record (MoR) (Paddle, Lemon Squeezy, FastSpring)
2. Direct Payment Service Provider (PSP) (Stripe, Adyen, Braintree)
---
> [!FOUNDER]
> Executive Insight from Enow A. Jovial (Founder & CEO, IBRAVRA)
> "At $200k ARR, using an MoR is a no-brainer—the engineering cost of managing EU OSS VAT, Canadian GST/HST, and 45 US state sales tax filings exceeds the 2% fee premium. But at $5M ARR, that 200 basis point spread represents $100,000 in lost annual EBITDA. Scale on an MoR; optimize onto direct interchange-plus rails once your annual tax compliance software cost is less than 20% of your gateway savings."
---
1. Blended Take-Rate Comparison Matrix
| Financial & Operational Dimension | Merchant of Record (Paddle / Lemon Squeezy) | Direct Gateway (Stripe Custom Interchange-Plus) |
| :--- | :--- | :--- |
| Headline Processing Fee | 5.0% + $0.50 per transaction | 2.9% + $0.30 (Standard) / 2.2% + $0.10 (Tiered IC+) |
| Sales Tax / VAT Liability | Assumed 100% by MoR | Assumed 100% by Founder's Entity |
| Cross-Border FX Currency Drag | 2.0% – 3.0% spread on conversions | 1.0% on automated multi-currency settlement |
| Chargeback Dispute Overhead | Handled by MoR compliance team | Managed internally via Stripe Radar ($15 fee/dispute) |
| Billing Integration Time | 2 – 5 Days (Drop-in checkout) | 2 – 4 Weeks (Custom webhooks, tax, and portal) |
| Customer Legal Contract | End user contracts directly with MoR | End user contracts directly with SaaS company |
---
2. Mathematical Inflection Point Model ($1M to $10M ARR)
To calculate the exact economic spread $S$ between an MoR and Direct Stripe at annual revenue $R$ with average contract value $ ext{ACV}$:
$
ext{Transaction Count } N = rac{R}{ ext{ACV}}
$
Cost of Merchant of Record:
$
C_{ ext{MoR}} = (0.05 cdot R) + (0.50 cdot N)
$
Cost of Direct Stripe + Automated Tax Compliance Stack:
$
C_{ ext{Stripe}} = (0.024 cdot R) + (0.15 cdot N) + ext{Tax Software Cost} + ext{CPA Filing Fees}
$
Annual Cost at $5,000,000 ARR ($ ext{ACV} = $100 ext{/month} = $1,200 ext{/year}$, $N = 4,166$ accounts):
$C_{ ext{MoR}} = (0.05 imes $5,000,000) + ($0.50 imes 4,166 imes 12) = $250,000 + $24,996 = $274,996$
$C_{ ext{Stripe}} = (0.024 imes $5,000,000) + ($0.15 imes 50,000) + $18,000 = $120,000 + $7,500 + $18,000 = $145,500$
$ ext{Net Savings} = $274,996 - $145,500 = mathbf{$129,496 ext{ / year}}$
---
3. Sales Tax Nexus Exposure Breakdown
```
+--------------------------------+
| Annual Global SaaS Revenue |
+---------------+----------------+
|
+-----------------------+-----------------------+
| |
Under $1.5M ARR Over $2.0M ARR
| |
+-----------v---------------------+ +-------------v-------------------+
| Merchant of Record Architecture | | Direct Stripe + Tax Compliance |
| • 0 State Tax Filings | | • Stripe Tax / TaxJar Engine |
| • 0 EU OSS VAT Registrations | | • Quarterly State Tax Filings |
| • Instant Global Checkout | | • Saves 200+ bps in Gross Margin|
+---------------------------------+ +---------------------------------+
```
---
4. Worked Economics at $1M and $10M ARR — and the True Break-Even
Section 2 works the $5M case. The same model, with the same assumptions ($100 per month ACV, $1,200 per year, $18,000 tax stack), gives the two ends of the range. Annual transactions equal accounts multiplied by 12 monthly charges.
Worked Example A: $1,000,000 ARR
Accounts: $1,000,000 / $1,200 = 833$, and transactions: $833 imes 12 = 9{,}996$.
$
C_{ ext{MoR}} = (0.05 imes $1{,}000{,}000) + ($0.50 imes 9{,}996) = $50{,}000 + $4{,}998 = $54{,}998
$
$
C_{ ext{Stripe}} = (0.024 imes $1{,}000{,}000) + ($0.15 imes 9{,}996) + $18{,}000 = $24{,}000 + $1{,}499.40 + $18{,}000 = $43{,}499.40
$
$
ext{Net Savings} = $54{,}998 - $43{,}499.40 = mathbf{$11{,}498.60 ext{ / year}}
$
At $1M the saving is real but thin — and it assumes the $18,000 stack figure covers everything, which it does not once engineering time and filing labour are counted. That is why the FAQ's practical inflection range starts higher than the pure model suggests.
Worked Example B: $10,000,000 ARR
Accounts: $10,000,000 / $1,200 = 8{,}333$, and transactions: $8{,}333 imes 12 = 99{,}996$.
$
C_{ ext{MoR}} = (0.05 imes $10{,}000{,}000) + ($0.50 imes 99{,}996) = $500{,}000 + $49{,}998 = $549{,}998
$
$
C_{ ext{Stripe}} = (0.024 imes $10{,}000{,}000) + ($0.15 imes 99{,}996) + $18{,}000 = $240{,}000 + $14{,}999.40 + $18{,}000 = $272{,}999.40
$
$
ext{Net Savings} = $549{,}998 - $272{,}999.40 = mathbf{$276{,}998.60 ext{ / year}}
$
At $10M the spread is no longer a rounding error: it is more than a quarter of a million dollars a year, every year, for the same payment volume.
Worked Example C: The break-even the model actually implies
Annual transactions equal $R / 100$ (because $R / $1{,}200$ accounts times 12 months equals $R / 100$). Substituting:
$
C_{ ext{MoR}} = 0.05R + 0.50 imes rac{R}{100} = 0.055R
$
$
C_{ ext{Stripe}} = 0.024R + 0.15 imes rac{R}{100} + $18{,}000 = 0.0255R + $18{,}000
$
Setting them equal:
$
0.055R - 0.0255R = $18{,}000 quad Rightarrow quad R = rac{$18{,}000}{0.0295} approx mathbf{$610{,}169 ext{ ARR}}
$
Under this section's $18,000-stack assumption, the arithmetic break-even is about $610k ARR. The FAQ's $1.5M–$3M practical range is not a contradiction: it prices in the costs the model leaves out — engineering weeks, registration filings, CPA review, and audit risk. Read the two numbers together: $610k is where the fee math turns, $1.5M–$3M is where the full-cost case turns. Migrating at $700k on fee math alone, without budgeting the rest, is how teams discover the second number the hard way.
Worked Example D: FX drag and chargebacks, the lines everyone forgets
Assume 40% of $5M ARR is cross-border volume, which is $2,000,000. At the matrix's midpoint MoR spread of 2.5% against 1.0% direct:
$
ext{MoR FX drag} = $2{,}000{,}000 imes 0.025 = $50{,}000 quad ; quad ext{Direct FX drag} = $2{,}000{,}000 imes 0.01 = $20{,}000
$
$
ext{FX saving} = $50{,}000 - $20{,}000 = mathbf{$30{,}000 ext{ / year}}
$
Chargebacks, at an illustrative 1% dispute rate on the section 2 volume of about 50,000 transactions and the matrix's $15 fee per dispute:
$
500 ext{ disputes} imes $15 = $7{,}500 ext{ / year in dispute fees alone}
$
That $7,500 excludes lost product, representment labour, and the Visa VROL threshold risk in the migration checklist. Fee tables compare headline rates; the real bill includes FX drag and dispute operations, and both favour direct rails as volume grows.
> [!KEY TAKEAWAY]
> The fee math turns at about $610k ARR on an $18,000 tax stack, the full-cost case turns at $1.5M–$3M, and at $10M the annual spread is $276,998.60. Price the migration on all four lines — headline rate, per-transaction fees, FX drag, and dispute operations — not on the headline rate alone.
---
5. Decision Table: Stay on the MoR or Migrate
| ARR band / profile | Recommended path | Deciding factor |
| :--- | :--- | :--- |
| Under $610k ARR | Stay on the MoR | Fee math has not turned; compliance labour dwarfs the spread |
| $610k–$1.5M ARR | Stay, but budget the migration | Model break-even passed, full-cost case not yet; use the time to register nexus states |
| $1.5M–$3M ARR | Migrate, cohort by cohort | Full-cost saving of roughly $87,000–$129,496 a year funds the tax stack several times over |
| $3M–$10M ARR | Migrate, and treat delay as a budget line | Every year of delay costs six figures at $5M and $276,998.60 at $10M on this model |
| High cross-border share (40%+ of volume) | Migrate earlier than the band suggests | FX drag alone adds $30,000 a year per $2M of cross-border volume on the stated spreads |
| High chargeback or regulated profile | Stay longer, or migrate with Radar rules pre-built | Dispute operations and threshold risk punish a rushed migration more than fees reward it |
---
6. Operator Boundaries and Risk Layer
Who should not migrate yet. If nexus registrations are not filed, if the tax engine is not calculating correctly on a shadow run, or if payment-token portability is not confirmed in writing, the migration date is not real. A cheaper gateway that collects tax incorrectly is not cheaper; it is a liability with a lower headline rate.
Who should not stay on the MoR. At $5M+ ARR with stable, domestic-heavy volume and clean tax posture, staying is no longer a compliance decision — it is a $129,496-a-year convenience fee. Name it as such in the budget so the choice stays deliberate rather than inherited.
> [!WARNING]
> Never start collecting on direct rails before nexus registrations are live in every state where the business already exceeds the $100,000 or 200-transaction thresholds. Collecting without registering does not pause the liability — it accumulates it, with interest and penalties that vary by state, while the MoR would have absorbed it entirely. Registration first, migration second, no exceptions.
> [!TIP]
> Migrate in cohorts, not in a single cutover: route all new customers to direct rails first, then move existing subscribers in monthly cohorts, keeping the MoR live until the last cohort's renewal has cleared. A dual-stack quarter costs a little double-running; a failed single cutover costs involuntary churn across the whole base at once.
---